UvA-DARE (Digital Academic Repository) Cognitive patterns ... · Resumen (Spanish summary) 153 Appendices 161 Appendix A Subtest scaled score range (subtest scatter) 166 Table A.1

UvA-DARE is a service provided by the library of the University of Amsterdam (http://dare.uva.nl)

UvA-DARE (Digital Academic Repository)

Cognitive patterns in paediatric epilepsy: Intra-individual variability, cognitive patternsand patterns of cognitive change in children with epilepsy on the Wechsler IntelligenceScales for Childrenvan Iterson, L.

Link to publication

Citation for published version (APA):van Iterson, L. (2015). Cognitive patterns in paediatric epilepsy: Intra-individual variability, cognitive patterns andpatterns of cognitive change in children with epilepsy on the Wechsler Intelligence Scales for Children

General rightsIt is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s),other than for strictly personal, individual use, unless the work is under an open content license (like Creative Commons).

Disclaimer/Complaints regulationsIf you believe that digital publication of certain material infringes any of your rights or (privacy) interests, please let the Library know, statingyour reasons. In case of a legitimate complaint, the Library will make the material inaccessible and/or remove it from the website. Please Askthe Library: http://uba.uva.nl/en/contact, or a letter to: Library of the University of Amsterdam, Secretariat, Singel 425, 1012 WP Amsterdam,The Netherlands. You will be contacted as soon as possible.

Download date: 02 Nov 2018

http://dare.uva.nl/personal/pure/en/publications/cognitive-patterns-in-paediatric-epilepsy-intraindividual-variability-cognitive-patterns-and-patterns-of-cognitive-change-in-children-with-epilepsy-on-the-wechsler-intelligence-scales-for-children(478c3654-a53f-4226-b931-376a321a8798).html

Loretta van Iterson

Cognitive Patterns in Paediatric Epilepsy

Intra-individual variability, cognitive patterns and patterns of cognitive change

in children with epilepsy on the Wechsler Intelligence Scales for Children

© Loretta van Iterson, 2015 Proefschrift Neuropsychologie (Engels) Cover photo. Sierras malagueñas, Parque Nacional Sierra de la Tejeda, Almijara y Alhama. Spain, 2012, by Loretta van Iterson. ISBN/EAN 978-90-801507-0-6 NUR: 773 Print Gildeprint, Enschede. No financial support was received for this project.

COGNITIVE PATTERNS IN PAEDIATRIC EPILEPSY

Intra-individual variability, cognitive patterns and patterns of cognitive change

in children with epilepsy on the Wechsler Intelligence Scales for Children

ACADEMISCH PROEFSCHRIFT

ter verkrijging van de graad van doctor

aan de Universiteit van Amsterdam

op gezag van de Rector Magnificus

prof. dr. D.C. van den Boom

ten overstaan van een door het College voor Promoties ingestelde commissie,

in het openbaar te verdedigen in de Agnietenkapel

op woensdag 7 oktober 2015, te 14.00 uur

door Loretta van Iterson

geboren te ‘s Gravenhage

Promotiecommissie:

Promotor: Prof. dr. P.F. de Jong Universiteit van Amsterdam

Copromotores: Prof. dr. A.S. Kaufman Yale School of Medicine

Dr. B.H.J. Zijlstra Universiteit van Amsterdam

Overige leden: Prof. dr. H.M. Geurts Universiteit van Amsterdam

Prof. dr. M.W. van der Molen Universiteit van Amsterdam

Prof. dr. W.C.M. Resing Universiteit Leiden

Prof. dr. P. Ghesquière Katholieke Universiteit Leuven

Dr. G. Thoonen Radboud Universiteit Nijmegen

Faculteit der Maatschappij- en Gedragswetenschappen

Contents

Introduction 7 Chapter 2. Intra-individual Subtest Variability on the Dutch Wechsler Intelligence Scales for Children–Revised (WISC-RNL) for children with Learning Disabilities, Psychiatric Disorders, and Epilepsy 23

Chapter 3. Differential effect of lesion side on intra-individual variability in children with focal lateralized epilepsy 39

Chapter 4. Establishing Reliable Cognitive Change in Children with Epilepsy: The Procedures and Results for a Sample with Epilepsy 47

Chapter 5. Duration of epilepsy and cognitive development in children: A longitudinal study 61

Chapter 6. Paediatric epilepsy and comorbid reading, math and autism spectrum disorders: impact of epilepsy on the cognitive patterns 81 Discussion 109

Summary (English summary) 133

Samenvatting (Dutch summary) 143

Resumen (Spanish summary) 153

Appendices 161

Appendix A Subtest scaled score range (subtest scatter) 166

Table A.1. Mixed referred and non-referred samples. Characteristics of

the samples. 167

Table A.2. Base rate table. Subtest scatter on 5 subtests of the verbal scale

and on 5 subtests of the performance scale. 168

Table A.3. Base rate table. Subtest scatter on 10 subtests of the full scale. 169

Appendix B. Base rate tables: Verbal – Performance Discrepancies 170

Table B.1. Characteristics of the samples. 171

Table B.2. Base rate tables. VIQ > PIQ. 172

Table B.3. Base rate tables. VIQ < PIQ. 173

Appendix C. Base rate tables for the discrepancies between factor index

scores.Verbal Comprehension (VCI), Perceptual Organization (POI) and

Processing Speed (PSI): (VCI – POI , VCI – PSI, POI – PSI). 174

Table C.1: Characteristics of the samples. 175

Table C.2. Base rate tables. VCI – POI discrepancy. 176

Table C.3. Base rate tables. VCI – PSI discrepancy. 177

Table C.4. Base rate tables. POI – PSI discrepancy. 178

Appendix D. From Test 1 to Test 2. Cognitive Change After Serial Testing. 179

Table D.1. Characteristics of the samples. 180

Table D.2. Base rate table. Cognitive gains and cognitive losses on the

verbal scale. 181


performance scale. 182


full scale. 183

Table D.5. Base rate table. Cognitive Change on the Factor Index scores.

Characteristics of the sample. 184

Table D.6. Base rate table. Cognitive gains and losses on VCI, POI, and PSI. 185

Appendix E. Isolated epilepsy, isolated developmental disorders and

comorbidities in epilepsy: ROC images for Chapter 6. 186

Table E. Rate of children showing VIQ – PIQ discrepancies of 15 or more

points. 188

References 189

Additional publications and posters (related and unrelated topics) 202

Acknowledgements 205

About the author 211

Introduction

INTRODUCTION

9

Introduction

There is increasing evidence that when epilepsy occurs in childhood, it will have a major

impact on the course of the child’s life in terms of cognition, learning, behaviour and,

ultimately, psychosocial outcome (Camfield & Camfield, 2007; Hermann, Jones, Jackson,

& Seidenberg, 2012). Cognitive problems have been found to be one of the major factors

accounting for psychosocial outcome (Camfield & Camfield, 2007). Cognitive

development in children with epilepsy will be the topic of the present work.

Epilepsy is a heterogeneous disorder and shows diversity in terms of age at onset,

presentation, severity, response to treatment, duration, accompanying comorbidities, and

cognitive course. Given this diversity, it is not surprising that when it strikes, in terms of

cognitive outcomes, epilepsy will be considered a relatively benign disorder in some

children, and a relatively complicated disorder in other children. The present work will

relate to children with “not uncomplicated” epilepsy, that is, children referred to a tertiary

epilepsy centre or its affiliated epilepsy school for neuropsychological assessment

because of concerns about their cognitive development. The principal topics of interest

will be intra-individual variability within test scores, patterns of cognitive abilities in

children tested for the first time, and patterns of change over time of children who have

been tested more than once. The studies will compare the patterns with those of children

with other developmental disorders as well as with those of children with double

diagnoses of epilepsy and a learning or behavioural disorder, that is, of children with a

comorbid disorder.

To better understand the scope of the field of epilepsy and cognition, the following

pages will start with a definition of epilepsy and its occurrence in children. The main

topic, cognitive patterns displayed by the verbal and performance (nonverbal) scales of

the Wechsler Intelligence Scales for Children (WISC series) will be discussed in

association with epilepsy variables known to affect cognition in epilepsy: age at onset,

duration, severity, brain lesions and comorbidities. At the end of the chapter, attention

will turn to the research questions investigated in the present work.


10

Epilepsy in Children

Definition, Prevalence, Incidence and Classification. The International League

Against Epilepsy (ILAE) defines epilepsy as a neurological disorder characterized by an

“enduring disposition of the brain to generate epileptic seizures” (Fisher et al., 2005, p

470) and epileptic seizures as “transient occurring signs or symptoms due to abnormal

excessive or synchronous neuronal activity in the brain” (Fisher et al., 2005, p 470). That

is, clinically, epilepsy is characterized by seizures, and its electrophysiological

counterpart is abnormal electrical activity seen on the EEG.

Epilepsy may have its onset at any age, but children are particularly affected. The

proportion of children affected by epilepsy (prevalence) is almost one percent (Russ,

Larson, & Halfon, 2012; Sillanpää, 1992). The probability of occurrence of epilepsy

(incidence) is ~64 per 100.000 children per year of age. Incidence is highest in the first

year of life, with ~102 per 100.000 children. In the second to fourth year, incidence is ~65

per year; incidence declines to ~25 for the teenage years (Wirrell, Grossardt, Wong-

Kisiel, & Nickels, 2011).

The classification of seizures, epilepsies and epilepsy syndromes (I.L.A.E., 1981,

1989) has formed the basis of diagnosis, treatment and research on epilepsy during the

past decades. The classification system has been further revised and updated (Berg &

Scheffer, 2011; Engel, 2006) The classification system for both seizures and epilepsies,

used over the years and largely maintained in the newest classification, is based on

various levels of classification:

One level concerns seizure type. Focal seizures, also called partial seizures or

localization-related seizures, are limited to specific areas of the brain (like frontal lobe

seizures) and are mostly limited to a hemispheric side of seizure onset (right versus left

hemisphere). Further distinctions in the description of focal seizures relate to the

lateralization (left versus right hemisphere onset) and to the topographical localization

(frontal, temporal, parietal, central, occipital and combinations of these). Generalized

seizures involve both hemispheres. They may start at one hemisphere and instantly spread

to the other side.

A second level is aetiology. In idiopathic epilepsy no underlying cause other than

a possible genetic predisposition has been found. In symptomatic epilepsy, an underlying

cause, as an MRI-abnormality or another known aetiology is identified. A presumed, but

not identified cause, “probably symptomatic”, has been called cryptogenic, while the term

unknown is now being suggested for this group of seizures. Furthermore, there are

INTRODUCTION

11

epilepsies and epileptic syndromes where it is undetermined or uncertain whether the

seizures are focal or generalized, like the epileptic encephalopathies. Generally,

idiopathic epilepsies are associated with better cognitive outcomes than those of unknown

origin or symptomatic epilepsies (Nolan et al., 2003). For the term idiopathic epilepsy,

the notion of genetic causation is being used. For epilepsies with identified genetic

causes, however, cognitive outcome may show large variation (Olson, Poduri, & Pearl,

2014).

The Impact of Epilepsy on Cognition

One of the major areas of concern in paediatric epilepsy is cognitive development.

Children with epilepsy score an average 11 points lower on IQ tests than children without

seizures and about four points lower than their siblings without seizures (Ellenberg, Hirtz,

& Nelson, 1986). Half of the children have developmental delays, in contrast to 3% of the

general population (Russ et al., 2012); about 26% present with IQs below 80 (Berg et al.,

2008a), whereas IQ scores that low would be expected in only 9.2% of the general

population. Neurobehavioral comorbidities like learning disorders and behavioural

disorders have been frequently reported as well (Russ et al., 2012), and are becoming an

area of increased interest (Lin, Mula, & Hermann, 2012).

Verbal and nonverbal abilities. Verbal and nonverbal – hence performance –

abilities have historically been the core cognitive abilities within the broad domain of

intellectual functioning. Myriad studies have shown that verbal and performance abilities

are compromised in epilepsy (Aldenkamp, Alpherts, De Bruine-Seeder, & Dekker, 1990;

Bjornaes, Stabell, Henriksen, & Loyning, 2001; Braakman et al., 2012; Gülgönen,

Demirbilek, Korkmaz, Dervent, & Townes, 2000; Lopes et al., 2013; Northcott et al.,

2007; O'Leary, Burns, & Borden, 2006; Overvliet et al., 2011; Vermeulen, Kortstee,

Alpherts, & Aldenkamp, 1994).

Verbal and performance abilities are sampled in a standardized manner in the core

scales of the WISC series and encompass a wide range of cognitive skills (van Haasen et

al., 1986; Wechsler, 1992, 2004, 2005). The present work will focus on these core

abilities, the verbal and performance abilities, in children with epilepsy and will use the

Dutch Wechsler Intelligence Scales to measure them. The newest Wechsler test in The

Netherlands is the WISC-III (Wechsler, 2005), which, like the WISC-R (van Haasen et al.,

1986) still adheres to the traditional dichotomy of verbal and performance scales. The

verbal IQ scales (VIQ) and verbal comprehension indexes (VCI) are broad measures of


12

verbal abilities and include various subtests, which assess understanding of questions,

general knowledge, verbal reasoning and abstraction, and knowledge of the meaning of

words. The performance IQ scale (PIQ) and the perceptual organization or perceptual

reasoning index (POI/PRI) also are broad measures of nonverbal abilities, and include

several subtests which assess visual spatial reasoning and depend also on constructional

abilities, motor dexterity and speed. The scales measuring verbal and performance

abilities have been part of the Wechsler intelligence scales for over 70 years. They

remained largely unchanged for decades in the early WISC versions (up to WISC-III) but

they have been modified in the latest versions (WISC-IV, and the WISC-V, which was

published in the US in 2014) to rely less on motor skills and speed (Baron, 2005;

Flanagan & Kaufman, 2009). The newest versions of the Wechsler tests show a growing

tendency towards including more pure measures of neuropsychological functions (like

speed of processing and working memory) as factor indexes. However, even now that

current versions emphasize separate Indexes rather than separate IQs, the verbal and

nonverbal indexes (Verbal Comprehension Index and Perceptual Reasoning Index in 4th

editions) have a special status relative to other Indexes; they alone are combined to

constitute the General Ability Index, an alternate to Full-Scale IQ (FS-IQ, Flanagan &

Kaufman, 2009; Wechsler, 2004; Weiss, Saklofske, Schwartz, Prifitera, & Courville,

2006). The intelligence scales, and particularly the verbal scales, of both the original and

the newer versions, are found to be associated with learning and school achievement

(Glutting, Watkins, Konold, & McDermott, 2006; van Haasen et al., 1986).

Verbal and performance abilities, as measured by the intelligence scales for

children are frequently reported in studies on epilepsy, either as a topic of interest in the

study or as standard background information on the samples. Although Wechsler scales

are widely used, published studies on the WISC are not always readily comparable. This

is partly due to the changes in the newer test versions, but also because some researchers

have not used all 10 verbal and performance subtests (or 13, to include the indexes), but

rather have administered short forms. These short forms have ranged from one subtest

only (Oostrom, van Teeseling, Smeets-Schouten, Peters, & Jennekens-Schinkel, 2005) to

eight subtests (Bailet & Turk, 2000; Berg et al., 2008a, 2008b; Gülgönen et al., 2000).

Existing epilepsy studies usually contrast children´s scores on the verbal and

nonverbal scales to a reference sample like a sample of healthy control children

(Braakman et al., 2012; Gülgönen et al., 2000; Northcott et al., 2007; O'Leary et al.,

2006). Alternatively, the scores on a scale on one sample with epilepsy may be contrasted

INTRODUCTION

13

to another sample with epilepsy, as left versus right hemisphere onset seizures, or frontal

versus temporal epilepsy (Campiglia et al., 2014; Lopes et al., 2013; Miranda & Smith,

2001). Studies on variability in test profiles within an individual, that is, studies on

cognitive patterns, however, are relatively scarce.

Variability in test profiles. Some authors have suggested that there is a need for

more fine-grained studies on cognitive development in epilepsy (Hermann et al., 2012).

One way of making finer distinctions is studying cognitive profiles or cognitive patterns.

Cognitive patterns focus on relative strengths and weaknesses in a cognitive profile and

are therefore measures of variability within a test profile. While data on verbal and

performance scales are often reported in studies in epilepsy in children, relatively fewer

studies have been done on the cognitive patterns, and not much is known about the

epilepsy variables which may influence such patterns.

One of the measures of variability is the intra-individual subtest variability or

subtest “scatter”. Subtest scatter relates to the difference between the highest and the

lowest subscale within the verbal, the performance or the full scale. Although studies on

adults with normal intelligence have associated elevated scatter with brain lesions (Ryan,

Tree, Morris, & Gontkovsky, 2006), intra-individual subtest variability has not yet

received attention in the literature on epilepsy. This may be because the importance of

intra-individual subtest variability has been the subject of debate. Some researchers

advocate against the use of intra-individual subtest variability measures (Watkins &

Glutting, 2000; Watkins, Glutting, & Youngstrom, 2005), while at the same time, the

information is becoming standard in the test manuals or “companions” to test manuals of

English speaking countries (Kaufman, 1979; Wechsler, 2004) in order to be used by

clinicians.

It is still unknown, whether there is increased intra-individual subtest variability in

children with epilepsy, and how this relates to scatter in other clinical samples. In

Chapters 2 and 3, intra-individual subtest variability will be studied in large samples of

referred children, both with and without epilepsy. The aim is to investigate whether

relatively large amounts of scatter can be interpreted as a characteristic of clinically

referred samples, and, as such, as some kind of “pathology”. Also, which scale, if any, is

particularly prone to show elevated scatter in a particular neurodevelopmental disorder. In

addition, intra-individual subtest variability will be studied in relation to epilepsy

variables, particularly lateralization of epilepsy and presence of brain lesions on

neuroimaging.


14

A second measure of variability within a test profile is the difference displayed by

a child on the verbal and performance scales, the VIQ – PIQ discrepancy. The

discrepancy between the two scales provides information about whether one of the scales

is more vulnerable to seizures than the other. This is important because remediation can

be targeted to address the weaknesses in the profile, either directly, or indirectly by taking

advantage of the strengths to compensate for the weaknesses. It has been understood that,

in children, the left and right hemisphere functions are not mirrored in the verbal and

performance scales of the Wechsler scales (Miranda & Smith, 2001). Also, it has been

found that after epilepsy surgery, the performance abilities show a slightly better recovery

than the verbal abilities, regardless of side of surgery (Westerveld et al., 2000). In

addition, some evidence has been presented for a deleterious impact of daytime seizures

especially on the performance scale, and for night time seizure activity especially on the

verbal scale (Overvliet et al., 2011). Overall, however, the pattern displayed by the verbal

and performance scale has seldom been the direct focus of studies. It is possible,

however, to assess the pattern indirectly through the WISC scores provided as descriptive

background information. As will be discussed in one of the next chapters (Chapter 6),

when assessed indirectly, the literature remains inconclusive as to whether epilepsy shows

its impact on one of the scales (VIQ or PIQ) differentially, and which variables are

associated with lowered verbal or performance abilities. This question, the differential

impact of epilepsy (and epilepsy variables) on verbal and performance abilities, will be

the main topic in the present work.

This differential impact can be studied in association with epilepsy variables, like

lateralization, seizure type, or medication, but it is also important to relate the possible

differential impact to time-related aspects, like the onset of the seizure condition early or

later in life, and a shorter or longer duration of the seizure condition. It should be

explored, for example, whether cognitive patterns change over time with increased

duration of the seizure condition. In addition to the epilepsy variables, the differential

impact of epilepsy on cognition could also be explored in the context of other

neurodevelopmental disorders in the child, that is, in children who have behavioural or

learning disorders in addition to epilepsy. These two topics – time-related aspects in

epilepsy and comorbidities in epilepsy – will be discussed in the following paragraphs.

Cognitive change over time. Epilepsy may often be considered a long term, and

sometimes even life-long condition, with remission and relapses (Geerts et al., 2010;

Koepp, Thomas, Wandschneider, Berkovic, & Schmidt, 2014; Schmidt & Sillanpää,

INTRODUCTION

15

2012). A Dutch 15-year follow-up study on childhood onset epilepsy using questionnaires

after two, five and fifteen years, indicated that 49% of the children were seizure free

within two years and remained in remission at all measurement points. An additional 29%

became seizure free after 2-5 years. As many as 12% showed a varying course of

remissions followed by relapses and 10% were never seizure free longer than three

months (Geerts et al., 2010). Mean duration of the seizure condition was six years,

ranging from 0 to 21 years (Geerts et al., 2010). Similar results were reported by Schmidt

and Sillanpää (2012). Relapse of seizures may occur even after a seizure-free period as

long as seven years (Berg, Testa, & Levy, 2011). Anti-epileptic drugs (AEDs) are the

treatment of choice for epilepsy and although the majority of children achieve seizure

freedom with medication, a considerable number of children are difficult to treat and

continue to have seizures for a prolonged period of time. This raises the question of the

long-term impact of epilepsy on the cognitive development of a child.

Variables like age at onset and aetiology of the epilepsy have been associated with

increased severity of the epilepsy and worse cognitive outcome, although for none of

these variables results can be considered conclusive.

An onset of epilepsy (AOE) early in life will show worse cognitive outcome than

epilepsy which starts later in childhood (Berg et al., 2008a; Berg, Zelko, Levy, & Testa,

2012; Cormack et al., 2007). Berg et al. (2012) pointed out, however, that the bad

outcomes in early-life epilepsy are mediated by the response to anti-epileptic medication:

87% of the infants who developed epilepsy in the first year of life and did not respond

favourably to medication showed an IQ below 80 when reassessed eight years later. In

contrast, in a sample of children with older age of onset (7 years), none had an IQ below

80 later on, even if the children showed resistance to antiepileptic medication.

Epilepsy syndromes with known aetiology are associated with poor cognitive

outcome. Scales on syndrome severity (Dunn, Buelow, Austin, Shinnar, & Perkins, 2004)

rate these epilepsies as most severe or “most complicated”; fortunately they are relatively

infrequent epilepsies (Covanis, 2012). The most complicated epilepsies, like the epileptic

encephalopathies are, by definition, associated with cognitive arrest and cognitive

deterioration (Covanis, 2012; van Rijckevorsel, 2006). Underlying brain pathology (Berg

et al., 2012), generalized symptomatic epilepsies (Nolan et al., 2003) and failure to

respond to medication have been associated with worse outcome (Berg et al., 2012).

However, lowered IQ can be associated with any kind of epilepsy. Lowered IQ has been

found in children with epilepsies of “moderate severity”, like temporal lobe epilepsy


16

(Miranda & Smith, 2001; Westerveld et al., 2000), frontal lobe seizures (Braakman et al.,

2012), and cryptogenic focal epilepsy (Van Mil et al., 2010). In addition, epilepsies which

have been considered of “low severity” in a medical sense for a long time may still be

accompanied by cognitive problems. Such is the case for benign epilepsy with centro

temporal spikes (BECTS) which is often accompanied by language and reading problems

(T. Clarke et al., 2007; Northcott et al., 2007; Overvliet et al., 2011). Childhood absence

epilepsy (CAE), also considered a relatively mild type of epilepsy, is frequently

associated with impaired attention, even if there is satisfactory response to medication

(Masur et al., 2013).

After reviewing longitudinal studies on cognitive change in epilepsy, various authors

conclude that cognitive change in children with epilepsy manifests itself as cognitive

decline, sometimes referred to as “cognitive progression” (Dodrill, 2004; Seidenberg,

Pulsipher, & Hermann, 2007). The same authors conclude that cognitive progression and

its relationship to epilepsy variables is still insufficiently understood and that more studies

are needed (Dodrill, 2004; Seidenberg et al., 2007). More studies, including several

longitudinal studies, have appeared later on, but they have paid insufficient attention to

the broad spectrum of epilepsies. Rather, various studies focussed on relatively

“uncomplicated” epilepsies. While there is no formal definition of “uncomplicated

epilepsies”, the term will be used to designate non-referred children, children with a short

duration of epilepsy and children with an onset of epilepsy later in childhood. Studies on

children with uncomplicated epilepsies have reported a normal cognitive level (Berg,

Hesdorffer, & Zelko, 2011; Piccinelli et al., 2010), as well as a normal development over

time with only minor differences from healthy controls (Ellenberg et al., 1986; Hermann

et al., 2008; Jones, Siddarth, Gurbani, Shields, & Caplan, 2010). In a follow-up study on

cognitive outcome after the epilepsy had remitted, residual cognitive effects limited to

difficulties in processing speed were found (Berg et al., 2008b). Again, the study was

based on uncomplicated epilepsy, with a short duration of seizures (5 years of seizure

freedom at 8-year follow-up). The children were largely off medication and they were

selected for having FS-IQs above 80 at the last measurement and coming from families

where the siblings also had normal IQs (Berg et al., 2008b). A relatively uncomplicated

cognitive course, with no large differences from controls (Hermann et al., 2008; Oostrom

et al., 2005), has also been reported in longitudinal studies that selected children with an

onset of epilepsy relatively late in childhood (with a mean age at onset of almost nine

INTRODUCTION

17

years, as in Oostrom et al.), or in early adolescence (almost 12 years as in Hermann et al.,

2008). These studies on uncomplicated epilepsies are of great value because they aim at

isolating the impact of seizures themselves on cognitive development in children. The

results are encouraging and indicate that children’s development shows resilience in the

light of a single adverse event, epileptic seizures.

Many seizure conditions, and especially those of children attending specialized

centres, are not “uncomplicated”. Children with “not uncomplicated” epilepsies are

referred to psychological evaluation or special services because concerns have risen; they

often have not shown satisfactory response to the first medication and continued to have

seizures over time.

Whereas complicated epilepsy appears as a clinically more urgent topic of study,

and inclusion of a wide spectrum of epilepsies is encouraged (Nolan et al., 2003), much

has still to be unravelled when it comes to understanding cognitive change over time in

referred children (Hermann et al., 2012).

The present work will study changes of cognitive abilities – cognitive decline –

over time in children who were followed longitudinally. Again, the focus will be the

pattern of verbal and performance abilities and the changes that may occur during the

course of the epilepsy differentially on the verbal and performance abilities. Cognitive

decline of the verbal and performance abilities will be studied longitudinally in relation to

epilepsy variables thought to be associated with more complicated epilepsies like early

age at onset and longer duration of epilepsy. In addition, rates of children showing

reliable cognitive change will be established to know what proportion of children is likely

to show a clinically meaningful change in IQ. Reliable cognitive change refers to

cognitive changes seen between serial measurement points which surpass a specified cut-

off value and are therefore considered clinically meaningful changes (Chelune, Naugle,

Lüders, Sedlak, & Awad, 1993).

Comorbidities. Strictly speaking, when two conditions co-occur in an individual,

they are understood as being comorbid conditions (Angold, Costello, & Erkanli, 1999).

Alternatively, a narrower and more common definition proposes that when epilepsy and

other developmental disorders co-occur at rates higher than expected, these

developmental disorders are referred to as comorbidities in epilepsy (Brooks-Kayal et al.,

2013; Lin et al., 2012). Comorbidities add to the burden of epilepsy (Berg, Caplan, &

Hesdorffer, 2011) and impact on the quality of life. Therefore, it is not surprising that the


18

plea to study comorbidities in epilepsy is sounding increasingly louder (Asato, Caplan, &

Hermann, 2014; Helmstaedter et al., 2014).

Psychiatric and behavioural comorbidities. Epilepsy in children is associated with

an increased risk of psychiatric and behavioural comorbidities. Comorbidities described

in epilepsy are attention deficit disorders, conduct disorders, anxiety, and depression

(Hermann et al., 2008; C. J. Reilly, 2011; Russ et al., 2012). In addition, one of the

comorbidities of major concern in epilepsy is autism spectrum disorders (ASD). Rates of

ASD in epilepsy range from 15 to 32% (Berg & Plioplys, 2012; D. F. Clarke et al., 2005;

Russ et al., 2012). ASD comorbid with epilepsy is considered a major challenge in terms

of treatment (Tuchman, Alessandri, & Cuccaro, 2010), and ASD and epilepsy are often

seen in combination with cognitive impairment (Berg & Plioplys, 2012).

School achievement difficulties. School problems are common in children with

epilepsy, including children with epilepsy who have IQs in the average range (Austin,

Huberty, Huster, & Dunn, 1999; C. Reilly & Neville, 2011). Russ et al. (2012) presented

epidemiological data showing that up to 75% of children with epilepsy needed special

(individualized) educational services, 52% of children with seizures had learning

problems, and 32% repeated a grade (Russ et al., 2012). Fastenau et al. (2008) found

academic underachievement (learning problems discrepant with IQ) in ~50% of the

children. About 40 to 60% of the children showed low achievement in at least one area

(reading, arithmetic, writing to dictation and writing samples), wherein writing was the

most frequently affected area.

Achievement difficulties have been reported to antedate the emergence of the

epileptic condition proper, as reflected in grade repetition and use of special services in as

many as ~23% of the children who later developed epilepsy (Berg, Hesdorffer, et al.,

2011; Schouten, Oostrom, Jennekens-Schinkel, & Peters, 2001). Therefore, learning may

be understood as a cognitive domain which is vulnerable to the epilepsy which is about to

emerge, but learning disorders may also be comorbidities which manifest themselves

already before the onset of epilepsy.

Different models have been described to understand the co-occurrence of epilepsy

and comorbidities. Pal (2011), for example, suggests that in some cases the comorbidity

can be understood as being caused by the seizure condition, while in other cases the

epilepsy and the comorbidity appear as distinct conditions, with shared or independent

causal factors. In addition, Pal suggests that in yet other cases, the presence of epilepsy

INTRODUCTION

19

may modify the comorbid disorder, for example worsening it (to illustrate, a language

disorder may appear worse if accompanied by seizures).

Thus, while it is known that behavioural and learning disorders are frequent

comorbidities in epilepsy, a question which remains largely open is whether these

comorbidities present in a similar fashion, that is, with similar cognitive patterns, as in

their isolated condition without epilepsy. Conversely, not much is known on the impact

the comorbidity exerts on the cognitive pattern of the child with epilepsy. If epilepsy and

comorbidity would influence cognitive patterns, this could have clinical implications in

terms of diagnosis and remediation. These questions will be addressed in in Chapter 6.

Present Study

While cognitive development is a well-researched topic in epilepsy in children,

there have been relatively fewer studies on cognitive patterns in epilepsy. It is still largely

unknown whether there is increased intra-individual subtest variability in epilepsy and

whether the verbal and performance abilities are affected differentially. If so, what

epilepsy variables are related to such patterns? Also, whereas epilepsy is a long-term

condition, it remains largely unknown whether patterns remain stable over time or show

change over time. While comorbidities are frequent in epilepsy, little is known about the

impact of epileptic comorbidities on cognitive patterns. These will be the topics of the

present work.

The studies are based on large numbers of referred children who have been

surveyed in a tertiary epilepsy clinic or who have been receiving special educational

services over a prolonged period of time. The children were “selected” children in the

sense that they were all referred to tertiary settings for clinical neuropsychological

evaluation or educational services. They can be considered as “unselected” in the sense

that they encompassed a wide range of epilepsies in terms of age of onset, type and

severity of epilepsy, duration of seizures, aetiology and cognitive level.

In the end, research in neuropsychology should be of utility for the clinical setting.

It should be “consumer friendly” (Chelune, 2002). Where relevant, the different chapters

will aim at providing cut-off values and data on classificatory statistics. As suggested by

various authors (Watkins et al., 2005; Woods, Weinborn, & Lovejoy, 2003) data on

sensitivity and specificity, or Receiver Operating Characteristics (ROC) analysis, will be

added.


20

Participants. The present study will focus on children between the in ages 4 to 16

years with a diagnosis of epilepsy who were clinically referred to a tertiary epilepsy

centre, or to a school providing services for children with epilepsy associated with the

centre. They were referred for special school services within a normal or special school

setting – with a partial overlap between the children known to the epilepsy centre and

those known to the school providing special education services. The data were

observational and were collected over a protracted period of time (2002 – 2014), but

archival data were searched as well on children tested earlier.

The children can be considered representative for children with epilepsy referred

for psychological assessment in The Netherlands, and as such, for children with “not

uncomplicated” epilepsy. In this sense, they are likely to belong to the 75% of the

children found in an epidemiological study (Russ et al., 2012) who need some kind of

specialized assistance in school. In the Netherlands, there are two epilepsy centres with

their corresponding school services which provide tertiary epilepsy care nationwide: one

centre and school for the northern half of the country, and one centre and school for the

south. The two centres may be considered largely equivalent. The two schools work

together in a nationwide school service centre on education and epilepsy. These school

services provided for by the nationwide centre are not restricted to the children known to

the tertiary epilepsy clinic or the pupils from the epilepsy school. Rather, they are offered

to all children with epilepsy, regardless of their school setting, provided that an

independent committee has entitled them for these services (Pijl & Pijl, 1998; Resing et

al., 2002). The children of the present studies came from the center in the north at its

associated school.

Research questions.

The main research questions addressed will be:

(1) Is intra-individual subtest variability elevated in epilepsy? It will be studied

whether increased intra-individual subtest variability is elevated in clinical

samples with and without epilepsy (Chapter 2); and whether intra-individual

subtest variability in epilepsy is associated with seizure variables (seizure

lateralization and presence of lesions on neuroimaging) (Chapter 3). The

hypothesis tested is that children with developmental disorders show elevated

intra-individual subtest variability; that is, that elevated intra-individual subtest

variability can be seen as a sign of pathology.

INTRODUCTION

21

(2) Do children with epilepsy have elevated rates of Reliable Cognitive Change? The

hypothesis tested in Chapter 4 will be that children with epilepsy are more liable

to show reliable cognitive change than other referred children, and that this

change presents predominantly as cognitive decline.

(3) What is the pattern of cognitive change seen in longitudinal studies in epilepsy

over time? In children tested two or three times, time-related variables like age at

onset and duration of epilepsy, and other variables associated with severity of

epilepsy, will be tested in relation to the change (i.e., decline) which can be seen

on the verbal and performance scales (Chapter 5). The hypothesis will be that in a

heterogeneous sample of relatively complicated epilepsies, cognitive decline can

be seen and that this decline affects the verbal and performance scales differently.

Based on the literature, it is hypothesized that epilepsy variables suggestive of

higher severity are likely to be associated with greater cognitive decline over time.

The study will also take into account participation in special education and lower

parental education; both variables are expected to be associated with lower IQs.

(4) What is the impact of comorbidities in epilepsy on cognitive patterns? Based on

children with epilepsy, with and without comorbid developmental disorders, and

on children without epilepsy but with other developmental disorders, it will be

studied whether the patterns from “isolated” conditions (a single diagnosis) differ

from those with comorbid conditions (a double diagnosis of epilepsy and another

disorder). The developmental disorders studied will be learning disorders (reading

disorders, math disorders) and of autism spectrum disorders (Chapter 6). The

hypothesis tested will be that the co-occurrence of epilepsy and a developmental

disorder will affect the cognitive pattern, leading to patterns different from those

of isolated conditions.

CHAPTER 2

Intra-individual Subtest Variability on the Dutch Wechsler Intelligence

Scales for Children–Revised (WISC-RNL) for children with Learning

Disabilities, Psychiatric Disorders, and Epilepsy

Loretta van Iterson

Alan S. Kaufman


24

Abstract

It is common practice to look at disparities among subtest scores (“scatter”) on an

intelligence test to establish if a score is deviant. However, it remains unclear whether

subtest scatter reflects primarily normal variation within individuals or is clinically

meaningful. The present study explored this issue based on data from 467 children with

developmental disabilities tested on the Dutch WISC-RNL. Of these children, 132 had

learning disabilities, 178 had psychiatric disorders, and 157 had epilepsy. Subtest scatter

was defined as scaled-score range (highest minus lowest scaled score). When contrasted

with “normal scatter,” the overall sample revealed higher ranges on the performance scale

and full scale, although effect sizes were small. Analysis of the data for the three separate

clinical samples revealed unusual scatter only for the sample of children with psychiatric

disorders. When comparing the clinical samples, scaled-score range was larger for the

sample of children with psychiatric disorders than for those with epilepsy. Two distinct

subsamples revealed elevated ranges with moderate effect sizes: children with autistic

spectrum disorders and children with left hemisphere seizures. These results suggest that

elevated subtest scaled-score range might characterize specific clinical samples rather

than denoting an overall sign of pathology.

van Iterson, L., & Kaufman, A. S. (2009). Intra-individual subtest variability on the Dutch Wechsler Intelligence Scales for Children-Revised (WISC-RNL) for children with learning disabilities, psychiatric disorders, and epilepsy. Psychological Reports, 105(3 Pt 2), 995-1008.

INTRA-INDIVIDUAL SUBTEST VARIABILITY

25

In child neuropsychology, the clinician frequently looks for strengths and weaknesses in

the cognitive profile, often operationalized as a positive or negative difference of 1 or 2

standard deviations, in order to make a diagnosis of a developmental disorder (Sattler,

2001). This approach is based on the assumption that a child’s profile should be uniform,

and that undue inter-subtest or intra-individual variability (scatter) can be interpreted as

pointing toward a specific strength or deficit. Two common indexes of scatter are subtest

scaled-score range (the simple difference between the highest and the lowest score in a

profile) and univariate scatter (the number of subtests deviating 1 SD from an individual’s

own mean). Kaufman (1976, 1979) showed that large intra-individual variability on these

indexes, far from being unusual, was seen frequently in the standardization sample of the

WISC-R. Later, Silverstein (1987) demonstrated that the empirically-derived moments

(mean and SD) from Kaufman's data were a function of the psychometric qualities of the

test and could be estimated from the average intercorrelations among the subtests

comprising the scales. Both subtest scaled-score range and univariate scatter make use

only of the extreme values in a profile. As a more sensitive measure of intra-individual

variability, the Profile Variability Index was proposed which, like a standard deviation,

uses information derived from all subtests (Matarazzo, Daniel, Prifitera, & Herman, 1988;

McLean, Reynolds, & Kaufman, 1990). Interestingly, subtest scaled-score range was

shown to correlate highly with Profile Variability Index (Boone, 1993; Matarazzo et al.,

1988).

The question of whether elevated intra-individual variability is a sign of

pathology, or only a reflection of the psychometric properties of the test, remains

unsettled. Some researchers have provided evidence for elevated variability in pathology

(Mayes, Calhoun, & Crowell, 1998; Ryan, Tree, Morris, & Gontkovsky, 2006; Zimet,

Goodman Zimet, Farley, Shapiro Adler, & Zimmerman, 1994), while others ardently

advocate against any use of measures based on inter-subtest variability (Watkins &

Glutting, 2000; Watkins, Glutting, & Youngstrom, 2005). In the studies by Watkins and

his colleagues, inter-subtest variability did not have a significant incremental validity in

predicting academic achievement over and above full-scale IQ in samples of exceptional

children, mainly children with learning disabilities. The authors argue that inter-subtest

variability is of no use as a diagnostic indicator, and its use can be considered

“prescientific” (Watkins et al., 2005, p.263).

In spite of this controversy, recent Wechsler test manuals have incorporated

subtest scaled-score range in the form of base rate tables (e.g., for the WISC-III,


26

Wechsler, 1992; and for the WISC-IV, Wechsler, 2004b). For the Dutch, the adult test

version includes subtest scaled-score range (Wechsler, 2004a), while the children’s

versions do not (van Haasen et al., 1986; Wechsler, 2005).

In order to be a sign of pathology, the intra-individual variability should be

significantly different in clinical samples when compared to the standardization sample.

Significant scatter should not only be interpreted as reliable scatter – that is, genuine and

not the effect of measurement error – but also as uncommon, in the sense that the

magnitude of occurrence within the normal population is small, e.g., 5% (Crawford &

Allan, 1996).

Because the Wechsler scales are well-standardized and well-researched, they are

used, in the present paper, as a model to further evaluate whether intra-individual

variability is a clinically meaningful measure of pathology. This study focused on subtest

scaled-score range because it is practical and easily computed by clinicians and is

included in most recent test manuals. Furthermore, it correlates highly with Profile

Variability Index. Univariate scatter, on the other hand, was not included because its

distribution is skewed, preventing parametrical analyses. The data refer to the Dutch

WISC-R (for this purpose, WISC-RNL; van Haasen et al., 1986) and were collected up to

2007. This version was in use in the Netherlands for a prolonged time, thus allowing the

collection of larger samples. For the newer test version, it will take some time before

sufficiently large samples are accrued, but underlying notions about subtest scatter can be

understood independent of test version. The expected mean values of subtest scaled-score

range and the cut-off values for uncommonly large ranges for the WISC-RNL were

estimated according to Silverstein (1987; 1989), aided by Owen’s (1962) range statistics.

These estimates draw on the averaged intercorrelations between the subtests, which came

from the technical manual of the WISC-RNL (de Bruyn, Vandersteene, & van Haasen, 1986,

p. 139, from N = 1961 children).

Based on WISC-RNL data on 467 children from three clinical samples, the aims of

this study were (1) to study whether subtest scaled-score range in children with

developmental disabilities shows differences compared to expected (normal) values; (2)

to study whether there are differences among the clinical samples, and, if so, (3) to

explore whether specific subsamples account for these differences; and (4), to report rates

of individuals with uncommonly large subtest scaled-score ranges found in clinical

samples.


27

Table 2.1.Samples and Diagnoses

Sample N % N % Learning Disabilities 132 32.3

Psychiatric Disorders 178 38.1 Autism Spectrum Disorders 58 32.6 Conduct Disorders / Oppositional Defiant Disorders 36 20.2 Attachment Disorders 28 15.7 Attention Deficit and Hyperactivity Disorders 27 15.2 Tic Disorders 19 10.7 Depression 14 7.9 Other 66 37.1

Epilepsy 157 33.6

Seizure Type: Focal Onset / Localisation Related Seizures 87 55.4

Left Hemisphere 33 Right Hemisphere 24 Bilateral / Multifocal 32

Generalized Seizures 32 20.4 Uncertain whether Focal or Generalized 21 13.4 Unknown 17 10.8 Anti-epileptic Drug:

0 9 5.7 1 59 37.6

>1 63 40.1 n a 26 16.6

MRI positive data 29 18.5

Total 467 100

Methods

Participants

Participants were N = 467 children, aged six to 16 years, with FS-IQs > 75. Overall, 353

(76%) were male. The children were entitled to benefit from distinct special school

services in The Netherlands, according to national regulations (Resing et al., 2002). These

regulations describe the criteria for placement in different settings specialized in,

respectively, (specific) learning disabilities, childhood psychiatric disorders, or childhood

epilepsy. Generally, information from four sources is weighted by an independent

committee. These sources are the family of the child, the present school, a psychologist


28

who did the assessment (including the intelligence testing), and an educational,

psychiatric or medical specialist. For learning disabilities, specified criteria in terms of

academic failure must be met; for psychiatric disorders, a DSM-IV diagnosis from a

psychiatrist or qualified mental health psychologist is required; for epilepsy, a diagnosis

from a neurologist is required. In all cases, the difficulties caused by the diagnosed

condition must exceed the competencies of the regular school. Normal intellectual

abilities were a further criterion for the schools of the first two samples, but not the third

(epilepsy). As indicated, in this study FS-IQ was set to be above 75 for all. Co-morbidity

is a common phenomenon in childhood developmental disabilities and the samples are

diagnostically heterogeneous; the primary diagnosis as reflected by special school

placement was the criterion for inclusion in a sample. Diagnostic group membership –

type of special school – was the independent variable in this study. Demographic data are

presented in Table 2.1 and data on the Wechsler scales are shown in Table 2.2.

The first sample consisted of N = 132 children with (specific) learning disabilities

and the second sample consisted of N= 178 children with psychiatric disorders. The latter

group included children with neurodevelopmental disorders as well as children with

behavioral and emotional disorders related to major life events (e.g. traumas). The main

diagnoses of this sample were autism spectrum disorders (ASD), conduct disorders or

oppositional defiant disorders, reactive attachment disorders, attention deficit and

hyperactivity disorders, tic disorders, and depression. The subsamples are listed in Table

2.1. The percentages add up to over 100% due to psychiatric co-morbidity.

The third sample consisted of N = 157 children with seizure disorders. Mean age

at epilepsy onset was 5.6 years (ranging from the first day of life to age 15 years with SD

= 3.2). Mean duration of epilepsy was 4.0 years (SD = 3.2). Seizure type classification,

side of epilepsy onset, and information on medication and neuroimaging are presented in

Table 2.1.

Analyses

For each participant, subtest scaled-score range was calculated for five verbal, five

performance, and ten full-scale subtests. Verbal, performance, and full-scale subtest

scaled-score range was the dependent variable in the study. Mean scores are given in

Table 2.3; z-converted means are depicted in Figure 2.1. Levene’s testing for homogeneity

of variances showed no significant values for ANOVA or ANCOVA. ANOVA revealed

differences in mean age (age was higher in learning disabilities compared to epilepsy) and

mean PIQ (but not VIQ or FS-IQ), indicating higher PIQ in the samples of children with


29

learning disabilities and psychiatric disorders compared to the sample of children with

epilepsy. Also, chi-square showed that boys and girls were unevenly distributed among

the samples; significant differences were found, indicating that the rate of boys was

higher in the sample with psychiatric disorders than the sample with epilepsy. Three

separate ANCOVA’s were undertaken (for verbal, performance and full-scale subtest

scaled-score range), controlling for the pre-existing differences in PIQ, age, and sex. With

multiple, one-sided, one-sample t-tests, the observed clinical values were compared to the

estimated expected values (A. B. Silverstein, 1987), and effect sizes were calculated

accordingly (Cohen, 1988, p. 45). Overall, alpha was set at .05 and Bonferroni corrections

were used to control for family-wise errors. With chi-square, rates of children with

uncommonly high subtest scaled-score range (verbal scale: ≥ 8 points; performance scale

≥10, and full scale ≥11), expected in ~5% of the normal population (A.B. Silverstein,

1989), were compared to this value. As this value was seen as an approximation only,

alpha was set to .001. Rates of uncommonly high subtest scaled-score range were also

compared between the clinical samples.

Results

Verbal Scale

Comparison of means. Table 2.3 presents the expected mean subtest scaled-score

range for the verbal scale (mean = 4.7, SD = 1.7) and the observed valued for the distinct

samples, together with the results of the one sample t-tests, and Figure 2.1 depicts the z-

converted values of subtest scaled-score range. No differences were found between the

mean expected values and the observed values of the total sample or any of the distinct

clinical samples (Table 2.3). No differences were found between the means of the clinical

samples (ANCOVA: F(2,462) = 0.138, p = .871, n.s).

Rates of uncommonly large ranges. Large ranges (≥ 8 points) were found,

respectively, in 8.3%, 12.9% and 15.3% of the children with learning disabilities,

psychiatric disorders, and epilepsy. Compared to the expected rates, these values reached

significance for psychiatric disorders (Χ2 = 23.51, p < .001) and epilepsy (Χ2 = 34.97, p <

.001). Chi-square revealed no difference in the distributions of uncommonly large ranges

between the clinical samples. There was an almost twofold rate (likelihood ratio 1.8, 95%

Confidence Interval [CI]: 0.93 – 3.6) of children with epilepsy versus children with

learning disabilities.

CO

GN

ITIV

E PA

TTER

NS

IN P

AED

IATR

IC E

PILE

PSY

Tabl

e 2.

2.M

ean

Age

, Mea

n W

ISC

-RN

L IQs a

nd S

ex fo

r Thr

ee C

linic

al S

ampl

es a

nd T

wo

Subs

ampl

es (A

utis

m S

pect

rum

D

isor

ders

and

Lef

t Hem

isph

ere

Ons

et S

eizu

res)

A

ge (y

rs)

V

IQ

PI

Q

FS

-IQ

Boy

s M

SD

M

SD

M

SD

M

SD

N

N

%

&

rang

e

& ra

nge

&

rang

e

& ra

nge

Lear

ning

Dis

abili

ties

132

91

68

.9

12

.8a 1

.3

93

.3

10.8

97.3

12

.3

94

.6

10.8

7.

6 to

15.

6 72

to 1

19

68 to

125

77

to 1

24

Psyc

hiat

ric D

isor

ders

17

8 15

9 89

.3a

10.9

2.

7 93

.8

11.5

95

.7

13.5

93

.9

10.7

6.

0 to

16.

7 70

to 1

32

61 to

130

76

to 1

27

Aut

ism

Spe

ctru

m D

isor

ders

58

56

96

.6

9.9

2.3

96.2

12

.2

96.6

14

.2

95.6

11

.2

6.2

to 1

5.1

70 to

132

61

to 1

24

76 to

127

Ep

ileps

y 15

7 10

3 65

.6

9.7

2.7

95.3

12

.1

91.0

b 11

.9

92.5

10

.7

6.2

to 1

6.7

71 to

134

66

to 1

35

76 to

125

Le

ft H

emis

pher

e Se

izur

es

33

26

78.8

9.

5 2.

5 96

.6

11.1

89

.1

11.7

92

.3

8.7

6.3

to 1

6.1

71 to

120

66

to 1

20

78 to

108

To

tal

467

353

75.6

11

.1

2.7

94.2

11

.5

94.6

12

.9

93.6

10

.8

6.

0 to

16.

7

70 to

134

61 to

135

76 to

127

Not

e. a si

gnifi

cant

ly h

ighe

r tha

n ch

ildre

n w

ith e

pile

psy.

b . S

igni

fican

tly lo

wer

than

chi

ldre

n w

ith le

arni

ng d

isab

ilitie

s and

psy

chia

tric

diso

rder

s

30


31

Performance Scale

Comparison of means. The estimated expected mean subtest scaled-score range

was 5.8 (SD = 2.4). The total sample and the children with psychiatric disorders differed

from the expected value (Table 2.3). Significant differences were also suggested among

the clinical samples (ANCOVA: F(2,461) = 3.024, p = .050, partial η2 = .01). However,

pair-wise comparisons between the clinical samples did not yield significant results.

Rates of uncommonly large ranges. Large ranges (≥ 10 points) were found in

7.6%, 12.9%, and 7.0% of the children with, respectively, learning disabilities,

psychiatric disorders, and epilepsy. Compared to expected values, these percentages were

elevated for psychiatric disorders only (Χ2 = 23.51, p < .001). Chi-square revealed no

significantly different rates among the samples. Likelihood ratios were 1.7 (95% CI: 0.8

– 3.5) and 1.8 (95% CI: 0.93 – 3.7) for children with psychiatric disorders versus,

respectively, children with learning disabilities and children with epilepsy.

Full Scale

Comparison of means. The expected mean subtest scaled score range was 7.3 (SD

= 2.1). Values differing significantly from expected were found for the total sample and

the sample with psychiatric disorders (Table 2.3). Significant differences were found

between clinical samples (ANCOVA: F(2,462) = 4.130; p = .017, partial η2 = .02);

specifically, subtest scaled-score range was higher in children with psychiatric disorders

than in children with epilepsy.

Rates of uncommonly large ranges. Larges ranges (≥ 11 points) were found,

respectively, in 6.8%, 12.9%, and 10.8 % of the children with learning disabilities,

psychiatric disorders and epilepsy. Compared to expected values, significant differences

were found for psychiatric disorders (Χ2 = 23.51, p < .001) and epilepsy (Χ2 = 11.23, p =

.001). Again, differences of the large ranges between samples showed a trend that did not

reach statistical significance. Notably, however, there was almost a twofold rate

(likelihood ratio 1.9, 95% CI 0.9 - 4.0) for children with psychiatric disorders compared

to children with learning disabilities.

Subsamples

Although it is beyond the scope of this paper to enter into detail on all subsamples,

two subsamples were identified as showing conspicuously elevated subtest scaled-score

ranges relative to expected values: (a) from the sample with psychiatric disorders, the


32

subsample with autistic spectrum disorders (ASD, N = 58) was identified; and (b) from

the epilepsy sample, children with left focal onset seizures (LH; N = 33) were identified.

Data on these subsamples are also included in Table 2.2, Table 2.3, and Figure 2.1.

Table 2.3 shows that the ASD sample had a significantly larger subtest scaled-

score range than the expected values on the verbal scale, performance scale and full

scale, all with moderate effect sizes. When the ASD sample was compared to the other

children with diagnoses of psychiatric disorders (N = 120) in the psychiatric sample,

mean subtest scaled-score range was elevated for the ASD sample on the verbal scale

(t(1,94.7) = 2.49, p = .014, ES = 0.5) and the full scale (t(1,176) = 2.44, p = .016, ES =

0.4). Uncommonly large ranges were found for the verbal scale in 20.7% of these

children, for the performance scale in 15.5%, and for the full scale in 17.2%. The

percentages were significantly elevated when compared to expected values for the verbal

scale (Χ2 = 30.06, p < .001), performance scale (Χ2 = 13.51, p < .001), and full scale (Χ2 =

18.3, p < .001). Classificatory statistics revealed that when the ASD group was contrasted

to the others children with psychiatric disorders, uncommonly large ranges had

classificatory utility for the verbal scale: sensitivity was 21%, specificity was 91%,

Positive Predictive Power (PPP) was 52%, Negative Predictive Power (NPP) was 70%,

and likelihood ratio was 2.26 (95% CI 1.06 - 4.81). These values indicated that when a

child with psychiatric disorders is found to have a subtest scaled-score range of 8 or more

points on the verbal scale, it will more likely belong to the group with autistic spectrum

disorders.

Within the sample of children with psychiatric disorders, none of the other

subsamples showed elevated subtest scaled-score range consistently on all scales.

However, two subsamples of children with neurocognitive developmental disorders

showed elevations on one scale—specifically, the subsample with conduct disorders had

substantial scatter on the performance scale and the subsample with tic disorders had

elevated scatter on the verbal scale. These data merely suggest hypotheses for future

study, but are not presented here because many children had multiple diagnoses and the

sample sizes were too small to permit meaningful analyses.


33

Figure 2.1. Mean z-Converted Uncorrected Subtest Scaled-score Range Values and SEM-

bars for Verbal (black), Performance (white) and Full Scales (patterned), for Three

Clinical Samples and Two Subsamples

Table 2.3 shows that mean scaled-score range was higher than expected in the sample of

children with left hemisphere seizures. Significantly elevated values were seen for the

verbal scale (small effect size) and the full scale (moderate effect size). Such elevations

were not seen in the other epilepsy subsamples; planned comparisons indicated that

values were significantly different compared to the right hemisphere seizure-group for the

verbal scale (t(1,137) = 2.05, p = .042, ES = 0.3). Large range was found for the verbal

scale in 21.2% of these LH children, for the performance scale in 6.1%, and for the full

scale in 15.2%. The percentages were significant when compared to expected values for

the verbal scale only (Χ2 = 18.26, p < .001). Classificatory statistics revealed that when

the left hemisphere seizure-group was contrasted to the right hemisphere seizure-group,

there was a clear trend to find more children with uncommonly large ranges in the verbal

scale in the left hemisphere seizure-group. The valued failed to reach significance due to

lack of statistical power: sensitivity was 21%, specificity was 96%, PPP was 88%, NPP

was 47%, and likelihood ratio was 6.19 (95% CI 0.67 - 38.7).

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

Total LearningDisabilities

PsychiatricDisorders

AutismSpectrumDisorders

Epilepsy LeftHemisphere

Seizures

z-sc

ore

Verbal range

Performance range

Full Scale range


34

Discussion

The assertion that intra-individual variability is elevated in pathology, taken for granted

by some researchers and opposed by others, was the subject of analysis in this study,

which focused on subtest scaled-score range for clinical samples of children with learning

disabilities, psychiatric disorders, and epilepsy. Analyses were conducted at three levels.

At the broadest level, 467 children from three categories of developmental

disabilities (learning disabilities, psychiatric disorders, and epilepsy) were compared to

the expected (“normal”) values. Significant elevations were found in the performance and

full scales – with conspicuously small effect sizes. This finding suggested that the study

was profiting from the effects of a relatively large sample size and also that possible

meaningful information was being masked by focusing on the heterogeneous total group.

At the second level of analysis, each of the clinical samples was compared to the expected

values and to each other. The sample with psychiatric disorders showed significantly

more than normal intra-individual variability on both the performance and full scales.

Also, the sample with psychiatric disorders showed more variability than the sample with

epilepsy on the full scale. Effect sizes were larger than for the total clinical sample, but

were still small. At the third and most specific level of analysis, two homogeneous

subsamples were subjected to further scrutiny. It appeared that the sample with ASD

(within psychiatric disorders) and the sample with focal LH seizures (within epilepsy)

showed elevated scatter, compared both to the expected values and to the other children

in their respective clinical original samples. For ASD, this was true for all three scales

(moderate effect sizes). For left hemisphere epilepsy, this was true for the verbal scale

(small effect size) and the full scale (moderate effect size).

When evaluating the percents of children with uncommonly large ranges, children

with (specific) learning disabilities did not display any unusual elevations relative to

groups of normal children. However, rates were increased in children with psychiatric

disorders on all scales, specifically for the ADS subsample. Rates were also increased for

the sample of children with epilepsy on the verbal and full scales, more clearly so in the

subsample of LH seizures.

INTR

A-I

ND

IVID

UA

L SU

BTE

ST V

AR

IAB

ILIT

Y

Tabl

e 2.

3.Su

btes

t Sca

led-

scor

e R

ange

: Mea

ns, S

Ds,

t-val

ues,

Prob

abili

ties a

nd E

ffec

t Siz

es fo

r Thr

ee S

ampl

es a

nd T

wo

Subs

ampl

es

cont

rast

ed to

Exp

ecte

d V

alue

s.

Ver

bal S

cale

Pe

rfor

man

ce S

cale

Fu

ll Sc

ale

Sam

ple

n M

SD

ta

p ES

M

SD

ta

p ES

M

SD

ta

p ES

Ex

pect

ed V

alue

b 4.

7 1.

7 5.

8 2.

1 7.

3 1.

9 Le

arni

ng D

isab

ilitie

s 13

2 4.

8 2.

0 0.

52

ns

- 5.

8 2.

3 0.

37

ns -

7.

4 2.

1 0.

28

ns

-

Psyc

hiat

ric D

isor

ders

17

8 5.

0 2.

0 1.

84

ns

- 6.

5 2.

6 3.

77

<.00

1 0.

3 8.

1 2.

2 4.

83

<.00

1 0.

4 A

utis

m S

pect

rum

D

isor

ders

58

5.

6 2.

3 2.

85

.003

0.5

6.

8 2.

8 2.

74

.004

0.5

8.

7 2.

3 4.

48

<.00

1 0.

7

Epile

psy

157

5.0

2.0

1.76

ns

-

6.0

2.4

1.10

ns

-

7.7

2.3

1.81

ns

-

Left

Hem

isph

ere

Seiz

ures

33

5.

3 2.

2 1.

64

.050

0.4

6.

5 2.

4 1.

63

ns -

8.

4 2.

0 3.

00

.003

0.6

Tota

l

46

7

4.9

2.0

2.45

ns

-

6.

1 2.

4 3.

28

.001

0.2

7.8

2.2

4.19

<.

001

0.2

Not

e. a O

ne sa

mpl

e t-t

ests

for c

ompa

rison

s aga

inst

est

imat

ed (e

xpec

ted)

val

ue. O

ne ta

iled-

test

, d.f.

= N

-1 in

all

case

s b E

stim

ated

exp

ecte

d va

lues

acc

ordi

ng to

Silv

erst

ein

(198

7)

35


36

The fact that the performance scale and not the verbal scale yielded the significant

differences in the primary samples of this study is consistent with a diverse body of

neuropsychological literature that has shown Wechsler’s performance subtests to be more

sensitive to brain injury and brain dysfunction than Verbal subtests (Kaufman &

Lichtenberger, 2006, chapters 8 and 9). Nonetheless, the present study suggests that

elevated subtest scaled-score range can also be seen on the verbal scale in specific

samples.

For children with (specific) learning disabilities, no elevations were found on any

measure. These results are in line of earlier studies (Flanagan & Kaufman, 2009; Watkins

et al., 2005). For children with ASD, elevated intra-individual variability has been

reported earlier (Joseph, Tager-Flusberg, & Lord, 2002). For children with epilepsy, to

the authors’ knowledge, no such studies have been reported, but the results are in line

with the large VIQ > PIQ discrepancies reported for children with unilateral focal onset

epilepsy (van Iterson & Augustijn, 2006) regardless of side of seizure onset. The

elevations in subtest scaled-score range on the verbal scale in left hemisphere epilepsy

may be the result of plasticity in the developing brain (Vicari et al., 2000).

The results found for children with ASD and children with epilepsy are interesting

in the light of recent research on the commonalities underlying both conditions and the

findings of high rates of subclinical EEG abnormalities in children with ASDs even in the

absence of manifest clinical seizures (Spence & Schneider, 2009).

Effect sizes increased when the selected samples were more homogeneous,

suggesting that specific samples of children with developmental disabilities may show

elevated intra-individual variability while others may not, or may even show decreased

variability. Thus, studies of subtest scaled-score range and their interpretation should take

into account type of pathology.

Scaled-score range was not found suitable for classification purposes between the

large samples; 95% confidence intervals for likelihood ratios were non-significant,

though some trends could be found. This is not surprising as scatter is a non-specific

measure which does not provide an answer as to where the variability is coming from, or

if it follows some specific pattern.

Classificatory statistics applied on selected samples indicated that an uncommonly

large range in the verbal scale was more likely to belong to a child with ASD and not a

child with “another diagnosis” within psychiatric disorders. Also the data suggest that

uncommonly large variability on the verbal scale may be characteristic of children with


37

left, but not right, hemisphere onset seizures; likelihood ratio was not significant,

probably due to small sample sizes. In line with the results of this study, and pertinent to

the discussion of the interpretability of scores beyond the summed scores of a scale

(Flanagan & Kaufman, 2009; Watkins et al., 2005), it is worth noting Saling’s (2009)

perspective. Based on the results of research within a highly specific area of research in

neuropathology – epilepsy surgery – Saling (2009), advocates against the use of scales of

summed scores in neuropsychological assessment of memory functions and argues in

favour of task specific measurement.

Intra-individual variability was studied with the WISC-RNL-version – which has

now been replaced by the WISC-IIINL, and by the WISC-IVUS/UK in English speaking

countries. The study of intra-individual variability is not confined to a specific version of

the Wechsler scales, but is applicable to any of Wechsler's scales and, in principle, to

subtest profiles yielded by different batteries as well. Flanagan and Kaufman (2009)

discuss the issue of inter-subtest variability within WISC-IV Factor Indexes. The

appreciation of a true difference between subtest scores depends on knowledge of the

relationship among the measures (i.e., the intercorrelations of the subtests) as well as the

frequency of occurrence of differences in the population studied.

CHAPTER 3

Differential effect of lesion side on intra-individual variability in children

with focal lateralized epilepsy

Loretta van Iterson


40

Abstract

A differential impact of hemispheric side (left versus right) on cognitive measures,

specifically verbal and performance IQ, has been described previously for both focal

onset seizures and lateralized brain lesions. This study revealed a differential effect on

intraindividual variability, measured as subtest scaled-score range, on the Dutch WISC-R

and WISC-III, in children with epilepsy. The presence of documented brain lesion was

associated with elevated variability on the verbal scale for the left hemisphere seizure

group and with decreased variability on the verbal and full scales for the right hemisphere

seizure group.

van Iterson, L. (2010). Differential effect of lesion side on intra-individual variability in children with focal lateralized epilepsy. Psychological Reports, 107(1), 113-119,

VARIABILITY IN LESIONAL FOCAL ONSET SEIZURES

41

Introduction

In childhood epilepsy, brain lesions affect cognitive outcomes differentially, depending

on age at time of lesion (Satz, Strauss, Wada, & Orsini, 1988), lesion type (Klein, Levin,

Duchowny, & Llabre, 2000; Mayes, Calhoun, & Crowell, 1998), the likelihood that

reorganization of brain functions has occurred (Blackburn et al., 2007; Liégeois et al.,

2004; Loring et al., 1999), and the actual presence of seizures (Deonna & Roulet-Perez,

2005). Effects of various measures, like degree of diffuseness versus circumscription of

brain lesion (Klein et al., 2000) and probability of showing IQ gains following epilepsy

surgery (Westerveld et al., 2000) have been shown to interact with lateralization (left

versus right hemisphere).

The literature cited studied elevation of IQ scales or the relationship between the

scales. Studies in epilepsy that focus on intra-individual subtest variability (subtest

scatter) are sparse. A recent study on subtest scatter (defined as scaled-score range – i.e.,

the difference between the person’s highest and lowest scaled score on a scale) was

conducted on large samples of children with Learning Disabilities, Psychiatric Disorders

and Epilepsy (van Iterson & Kaufman, 2009). The results were interesting. In some

samples of children with developmental disabilities (particularly the children with

Psychiatric Disorders) increased intra-individual subtest variability was a marker of

pathology, while in others (Learning Disabilities) it was not. Specifically, children with

Autism Spectrum Disorders and children with left hemisphere seizures were more likely

to show elevated variability. This elevation was not seen in children with right

hemisphere seizures, suggesting a differential impact of seizure lateralization on subtest

scaled-score range. These results raise the question whether this difference is related only

to hemispheric side of seizure onset or whether there is a role for brain lesion as well.

This study focuses on the impact of side of focal onset seizures on children’s subtest

scatter in the Dutch WISC-R and WISC-III in the light of the presence or absence of

documented brain lesion.


42

Table3.1 Demographic data, epilepsy variables and results on Wechsler Scales for the four

samples

Hemispheric Side of Seizure Onset Left Right

Sample No lesion MRI lesion No lesion MRI lesion

N 34 22 19 15 Boys N (%) 22 64.7 16 72.7 10 52.6 8 53.3 Righthanded N (%) 29 85.3 17 77.3 15 78.9 14 93.3 Age (yr.) M (SD) 9.4 2.5 10.2 3.1 8.7 2.5 10.6 2.5

range 6.3 to 16.5 6.8 to 15.8 6.3 to 14.0 7.4 to 15.8 WISC version WISC-RNL N (%)

21 61.8 12 54.5 9 47.4 12 80 WISC-IIINL N (%)

13 38.2 10 45.5 10 52.6 3 20 Verbal IQ M (SD) 94.1 10.5 95.7 10.7 99.4 10.5 95.9 10.5

range 71 to 116 76 to 120 82 to 118 72 to 111 Performance IQ M (SD) 89.9 12.6 91.8 10.4 87.9 18.3 89.0 9.6

range 66 to 120 71 to 121 58 to 118 78 to 118 Full Scale IQ M (SD) 91.2 9.6 93.0 8.8 92.8 12.0 91.6 9.6

range 75 to 108 78 to 108 77 to 120 75 to 116 Age at onset of epilepsy M (SD) 5.2 3.1 6.6 3.7 5.6 2.8 5.3 4.5

range 0.3 to 14.8 0.5 to 16.2 0.1 to 10.3 0.3 to 14.5 Interval to Test (yrs) M (SD) 3.9 2.5 3.6 3.4 3.1 2.7 5.4 5.3

range 0.5 to 9.8 -3.2 to 12.3 0.4 to 10.0 -2.8 to 13.3 Antiepileptic Drugs M (SD) 1.6 0.6 1.7 0.8 1.6 0.8 1.1 0.9

range 1 to 3 1 to 4 0 to 3 0 to 3 Antiepileptic Drugs Tried M (SD) 2.2 1.4 2.6 1.5 2.3 1.6 1.5 1.1

range 1 to 7 1 to 5 1 to 7 0 to 4 Type of Lesion

Cortical dysplasia N (%) 5 22.7 5 33.3 Mesial Temporal Sclerosis N (%) 3 13.6 1 6.7 Tumor N (%) 4 18.2 2 13.3 Other N (%) 10 45.5 7 46.7

Surgery N (%) 6 27.3 4 26.7

Note. No statistically significant differences between the four samples are seen on any variable after ANOVA or chi-square testing.

Methods

Participants.

The sample consisted of 90 children with localization-related or focal onset

epilepsy (Engel, 2006), classified into four samples according to lateralization of seizure

onset (based on EEG and seizure semiology) and presence or absence of positive findings

on MRI (positive findings indicative of lesion versus no documented lesion). The

epilepsy data were available from neurological files and were collected independently

from the psychological data. The children were between six and 16 years of age (mean =

9.6 years, SD = 2.7); mean age at epilepsy onset was 5.7 years (SD = 3.5) and mean


43

duration of epilepsy was 3.9 years (SD = 3.4). Except for five children, the participants

were on anti-epileptic drugs and 42 were on more than one anti-epileptic drug. Only two

children were candidates for epilepsy surgery; in eight more, brain surgery had occurred

earlier in life for various reasons, e.g., tumors or arteriovenal bleeding; in two, testing

antedated seizure onset. Children were selected as having full scale IQs of 75 or higher.

Table 3.1 presents the demographic data as well as epilepsy variables and Wechsler IQ.

Fifty four children were tested with the Dutch WISC-R (van Haasen et al., 1986)

and 36 with the WISC-IIINL (Wechsler, 2005). The two test versions can be considered to

measure the same construct (Wechsler, 1992, 2005) as correlations between the scales are

high, most clearly so for the verbal and full scales (.90 for the verbal scale, .81 for the

performance scale and .90 for the full scale, Wechsler, 2005, p. 70). Values of expected

subtest scatter were virtually identical for both test versions, when these were calculated

following Silverstein (1987; see also van Iterson & Kaufman, 2009). For the WISC-RNL

(WISC-IIINL in brackets) the mean value for scatter among 5 verbal scale subtests was 4.71,

SD = 1.8 (mean = 4.73, SD = 1.8); for scatter among 5 performance scale subtests, mean

= 5.77; SD = 2.1 (mean = 5.71, SD = 2.1); and among 10 full scale subtests, mean = 7.34,

SD = 1.9 (mean = 7.36, SD = 1.9).

Analyses

For each child and for each scale, subtest scaled-score range was converted into a

z-score. The results of multiple one-sample two-sided t-tests for the four groups, corrected

for family-wise errors, are presented in Figure 3.1.

Results

Analysis of variance and chi-square analyses revealed no differences in means or

proportions between the four samples on any demographic variable, any of the WISC-

variables or any epilepsy variable (Table 3.1). Verbal – Performance discrepancy

favoured the verbal scale in all samples; paired sample t-tests revealed significant

differences for the two right hemisphere samples only.

ANOVA showed significant differences on subtest scatter for (a) the verbal scale (F(3,86)

= 2.923, p = .038, eta squared (η2) = .09), indicating more variability in the MRI lesion

for the left than the right hemisphere group; and (b) the full scale (F(3,86) = 5.136, p =

.003, η2 = .15), indicating less variability in the MRI lesion than no lesion group (Figure

3.1).


44

Figure 3.1. Subtest scaled-score range for the Verbal

(black), Performance (white) and Full Scale (striped):

means and SE.

a = one-sample t-testing for difference from zero after Bonferroni correction, p < .05 b = within MRI lesion groups: different from Verbal scaled-score range in left hemisphere, p < .01 c = within MRI lesion groups: different from Full Scale scaled-score range in left hemisphere, p < .05 d = within right hemisphere: different from Full Scale scaled-score range in group without MRI-lesion group, p < .001

Subtest scatter for each scale was entered as the dependent variable and seizure side and

MRI findings as the independent variables in a MANOVA. This analysis yielded

significant results for both main effects as well as their interaction. Significant values

were seen for seizure onset side for the verbal scale (F(3,86) = 4.067, p = .047, partial eta

squared (η2) = .05) indicating more variability in the left hemisphere seizure group. For

the MRI findings, the effect was seen on the full scale (F(3,86) = 11.724, p = .001, partial

η2 = .12); variability was lower in children with MRI lesions. There was a significant

seizure side by MRI interaction for the verbal scale (F(3,86) = 6.017, p = .016, partial η2

= .07), indicating that in the presence of a lesion variability was elevated in the left and

Subtest scaled score-range

a, b

a, c, d

-1,2

-1

-0,8

-0,6

-0,4

-0,2

0

0,2

0,4

0,6

0,8

1

1,2

no MRI lesion n = 34

MRI lesion n = 22

no MRI lesion n = 19

MRI lesion n = 15

Left Hemisphere Right Hemisphere

z-sc

ores


45

lowered in the right hemisphere seizure group. Also, a side by lesion interaction was seen

on the full scale (F(3,86) = 6.228, p = .014, partial η2 = .07); the right hemisphere group

without lesions showed elevated variability, and the group with lesions showed decreased

variability.

Although it is not meaningful to break down the samples further according to test

version given the small samples sizes, when test version was included as a predictor

however, an interaction with MRI lesion for the performance scale was seen affecting

variability on the right hemisphere sample (higher in the children tested with the WISC-

IIINL).

Moderate test accuracy was revealed by the areas under the curve (AUC) when

receiver operating characteristics (ROC) were calculated in order to study data on

individuals. Within the right hemisphere group, the procedure indicated that lower values

on the full scale would more likely belong to children with a lesion (AUC = 0.807, p =

.002, 95% Confidence Interval (CI) = 0.66 to 0.96, a cut-off value of one standard

deviation below the mean (-0.97SD) had a sensitivity of 40% and a specificity of 89%).

Within the lesion groups, a higher value in the verbal scale would more likely belong to a

child in the left hemisphere seizure group (AUC = 0.829, p = .001, 95% CI = 0.70 to

0.96; for a cut-off of +1 SD, sensitivity was 37% and specificity was 100%). A lower

value on the full scale would more likely belong to the right hemisphere group (AUC =

0.739, p = .015, 95% CI = 0.58 to 0.90; for a cut-off of -0.97 SD, sensitivity was 40% and

specificity was 91%).

Discussion

The present study confirms the finding that variability is elevated in left but not right

hemisphere onset seizures (van Iterson & Kaufman, 2009), and goes beyond it indicating

that a differential effect can be seen which depends on the presence or absence of a brain

lesion. In the presence of a lesion, variability was found to be elevated in the left

hemisphere group in the verbal Scale, and lowered in the right hemisphere group on the

verbal and full scales.

The results add to the existing body of literature on the differential impact of

lateralized brain pathology on cognitive measures like level of performance on the verbal

and performance scales and the difference between the two scales (Loring et al., 1999;

Westerveld et al., 2000) extending the differential impact to inter-subtest variability. The


46

findings also suggest that brain pathology, besides leading to elevated variability –

possibly reflecting poorer “central coherence” (Noens & van Berckelaer-Onnes, 2005) –

may also manifest itself as lowered, or “less than normal” variability (A.S. Kaufman,

1979). These differences can possibly be understood as being mediated by functional

reorganization of the brain (Blackburn et al., 2007).

Significantly elevated variability was not seen on the performance scale. There are

several possible explanations. First, the overall lower performance than verbal scores (van

Iterson & Augustijn, 2006) may have been reflected in lower subtest scatter (Matarazzo,

Daniel, Prifitera, & Herman, 1988). Second, data on the participants were collected over a

prolonged period of time which can affect diagnostic decisions – that is to say, advances

in neuroimaging techniques enable detection of lesions which earlier remained undetected

and may increase the number of children classified as having a lesion. Furthermore,

changes in intelligence test version may potentially affect results on one scale or subtest

more than the other (Flynn, 2007; A. S. Kaufman, 2010). As stated, IQs yielded by the

two versions of WISC used in this study correlated substantially. However, the

performance IQs did not correlate as highly as the verbal IQs (.81 vs. .90), suggesting that

the performance scale, in particular, may measure slightly different constructs on the

Dutch WISC-R and WISC-IIINL. All of these factors may have contributed to the somewhat

inconsistent results seen on the performance scale between the earlier and more recently

included cases.

Elevated variability in cognitive function in children has implications for

educational practice, as children may surprise and startle psychologists, parents and

educators when showing high levels of proficiency on one area and low on another area,

even though the two areas seem quite similar in the cognitive abilities they assess. Thus,

knowledge of this variability is the first step to understanding and guiding these children

when planning their interventions.

CHAPTER 4

Establishing Reliable Cognitive Change in Children with Epilepsy:

The Procedures and Results for a Sample with Epilepsy

Loretta van Iterson

Paul B. Augustijn

Peter F. de Jong

Aryan van der Leij


48

Abstract

The goal of this study was to investigate reliable cognitive change in epilepsy by

developing computational procedures to determine Reliable Change Index scores (RCIs)

for the Dutch Wechsler Intelligence Scales for Children. First, RCIs were calculated

based on stability coefficients from a reference sample. Then, these RCIs were applied to

a sample of 73 children with refractory epilepsy who were tested twice with the WISC

after a mean interval of 2.3 years. Results indicated that children with refractory epilepsy

are at risk for cognitive decline over time: 26.0 percent of the children showed reliable

losses on verbal IQ and 16.4 percent on the full-scale IQ (expected rate = 5%). Declines

on performance IQ were within expected limits.

van Iterson, L., Augustijn, P. B., de Jong, P. F., & van der Leij, A. (2013). Establishing reliable cognitive change in children with epilepsy: The procedures and results for a sample with epilepsy. Journal of Psychoeducational Assessment, 31(5), 448-458.

RELIABLE COGNITIVE CHANGE

49

Introduction

It remains uncertain whether cognitive declines occur over time during the course of

epilepsy (Devinsky & Tarulli, 2002). Two updates on longitudinal studies of children

with epilepsy indicate that evidence of cognitive decline in epilepsy continues to be

sparse and inconclusive (Dodrill, 2004; Seidenberg, Pulsipher, & Hermann, 2007).

Numerous studies, in fact, report non-significant changes over time in groups of children

with epilepsy (Aldenkamp, Alpherts, De Bruine-Seeder, & Dekker, 1990; Bjornaes,

Stabell, Henriksen, & Loyning, 2001; Bourgeois, Prensky, Palkes, Talent, & Busch,

1983; Jones, Siddarth, Gurbani, Shields, & Caplan, 2010; Oostrom, van Teeseling,

Smeets-Schouten, Peters, & Jennekens-Schinkel, 2005; Rodin, Schmaltz, & Twitty,

1986). These findings of no significant decline, though encouraging, are generally

associated with non-referred samples of children with relatively uncomplicated epilepsies

and high rates of seizure remission (Jones et al., 2010; Oostrom et al., 2005). However,

severity of epilepsy syndrome – particularly epileptic encephalopathies and epilepsies

with a known (symptomatic) or suspected (cryptogenic) cause – has been associated with

(a) low IQ (Bulteau et al., 2000; Nolan et al., 2003) and (b) increasing deviation from the

developmental curve (Berg et al., 2004). Furthermore, within an epilepsy syndrome,

variability among individuals has been reported to be high (Reijs et al., 2006).

To gain insight into the heterogeneity within a clinical sample, and to increase the

clinical meaningfulness of research data beyond the study of group means, studies should

also provide information on individuals (Chelune, 2002; Martin et al., 2002). Studies of

children with epilepsy that do report rates of individuals with cognitive change are rare,

both in samples without (Rodin et al., 1986) or with epilepsy surgery (Westerveld et al.,

2000).

These occasional studies of cognitive change in children have rarely used the

psychometric properties of the test to determine reliable cognitive change.

In adult literature, approaches to quantify reliable cognitive change have used

normal or clinically stable samples as reference groups. From these references samples,

reliable change has been computed applying regression-based procedures (Cysique et al.,

2011; Martin et al., 2002) or, alternatively, the Reliable Change Index (RCI; G.J.

Chelune, Naugle, Lüders, Sedlak, & Awad, 1993; Woods et al., 2006). These RCI

formulas tend to perform as well as more complex regression formulas (Heaton et al.,

2001) and to be more “user friendly” (G. J. Chelune, 2002, p.424).The RCI is a cut-off


50

score and ±RCI yields a confidence interval (typically 90%). Descriptions of procedures

to establish RCI appear in the literature (G.J. Chelune et al., 1993; Maassen, Bossema, &

Brand, 2009). When retesting is conducted, “change scores” that fall within the

confidence interval are interpreted as “common” or non-meaningful differences; on the

contrary, change scores outside the interval are deemed uncommon or “reliable cognitive

changes.” RCI formulas are based on the standard deviations of the mean scores at first or

“baseline” testing (Chelune et al., 1993) or at both baseline and retesting (G.J. Chelune et

al., 1993; Maassen, Bossema, & Brand, 2009). The formulas are also based on the

stability coefficients. Where practice effects are known to occur, the RCI cut-off score is

adjusted accordingly (Chelune et al., 1993; Woods et al., 2006).

Woods et al. (2006) underscored the importance of examining the validity of this

RCI approach in clinical samples expected to present cognitive change, whether

improvement or decline. Although epilepsy is considered an ongoing neurological

condition that can potentially lead to cognitive change, these kinds of validation studies

barely exist in the epilepsy research on children. Thus, reliable cognitive change in

children with epilepsy was the focus of the present study.

Statistical properties, Test Familiarity and Test Version

Changes in IQs at retesting vary depending on familiarity with the test, length of the

interval between measurements(Kaufman, 1994), IQ at baseline (Schittekatte, 2005), and

test version used (Canivez & Watkins, 1998). A period of 9 to 12 months is often

considered sufficient to counter practice effects in referred children (Canivez & Watkins,

1998); however, some evidence exists that in children with epilepsy, the impact of

practice effects may level off within six months (Neyens, Aldenkamp, & Meinardi, 1999).

When the same test is administered at both measurements, long-term studies on referred

children without epilepsy – retested after a 2½ to 3-year interval – generally show stable

scores, with differences that are either not significant or not clinically meaningful

(Canivez & Watkins, 1998; Pesch & Ponsioen, 2004). These test-retest studies are based

on data from the WISC, the WISC-R, the Dutch WISC-R (to be called WISC-RNL), or the WISC-

III. Stability coefficients over time have been found to be high for the three IQ Scales that

comprise the various versions of the WISC, verbal scale (VIQ), performance scale (PIQ)

and full scale (FS-IQ) (Canivez & Watkins, 1998; Schittekatte, 2005). For the Index

scores, Canivez and Watkins (1998) also report high stability coefficients; for the Dutch


51

Wechsler tests, however, coefficients were based on small numbers of children

(Schittekatte, 2005) and are of limited utility.

Purpose

In order to establish the rate of children with epilepsy who show reliable cognitive

change, the present study focused on the three IQ Scales of the Dutch Wechsler

Intelligence Test for Children in a sample of referred children with refractory epilepsy

who were tested twice. The following research questions were addressed:

(1) Given the presence of two sets Wechsler IQs obtained at Time 1 (T1) and Time 2

(T2), what magnitude of T1–T2 change is sufficient to denote a significant

change, that is, a “reliable cognitive change”?

(2) How often do these reliable changes occur in children with epilepsy, and is this

percentage larger than expected?

The principal hypothesis is that referred children with epilepsy show decline over time in

cognitive function (Seidenberg et al., 2007). It is hypothesized that the study of

individuals will uncover elevated rates of children who demonstrate reliable cognitive

decline during the course of epilepsy relative to a clinical control reference sample

without epilepsy.

The study included two phases: (1) testing the utility of reliable change formulas

and establishing RCIs (90% confidence intervals) based on Dutch Wechsler test-retest

data from a reference sample; and (2) applying the RCIs to establish the proportion of

children with epilepsy who show reliable cognitive changes at retesting. Phase 1 is

described in the Method section with Phase 2 presented in Results.

Methods

Participants

The sample comprised 73 Dutch children who met the following criteria: (a) they had a

diagnosis of epilepsy, (b) they were tested (T1) with either the WISC-RNL (van Haasen et

al., 1986) or WISC-IIINL (Wechsler, 2005) (c) they were retested (T2) after an interval of 12

months or more on the same test version that was used at T1, and (d) no epilepsy surgery

had occurred in between measurements. The children presented at a Dutch tertiary centre

for epilepsy or at a special school providing services for children with epilepsy. Reasons


52

for testing were concerns expressed by parents, schools, or neurologists about the child’s

cognitive development. Wechsler testing generally occurred as part of a comprehensive

neuropsychological evaluation and took place at the start of the evaluation or after a

break. Commonly, results of the assessment were used in applications for special

financial and educational services for the child. Thus, this study dealt with a selected

sample of referred children who were tested with the same WISC on both occasions, either

the WISC-RNL or the WISC-IIINL. No exclusionary criteria were applied in terms of epilepsy

type or IQ at T1.

The sample was selected from 420 children with epilepsy who had completed a

WISC. Of these, 290 children had been tested once. From the 130 children who had been

tested twice, 57 were not eligible because: (a) retesting had taken place within twelve

months (n = 23), (b) epilepsy surgery was conducted (n = 2), or (c) different versions of

the Dutch WISC were administered at T1 and T2 (n = 32). The selected sample (n = 73)

and the sample of children who were not eligible for this study (n = 347) were overall

similar. ANOVAS or chi-square tests using an alpha level of .01 to control for familywise

error rates, showed that the samples did not differ in sex, handedness, VIQ, PIQ or FS-IQ,

age at onset of epilepsy, duration up to first testing, number of anti-epileptic drugs (AED)

used, seizure type, severity of epilepsy syndrome, or rates of children with documented

MRI-abnormalities. The children of the sample of interest, however, were tested for the

first time at a younger age (selected sample: mean = 9.1, SD = 2.2; not selected sample:

mean = 10.2, SD = 2.8). The difference in age was significant, F (1,419) = 9.84, p = .002,

r = .15.

Table 4.1. Epilepsy variables at time 1 (T1) and time 2 (T2)

Mean SD Range

Age at onset of epilepsy (yrs) 5.4 3.0 0.1 to 13.2 Age at T1 (yrs) 9.1 2.2 6.0 to 15.9 Age at T2 (yrs) 11.4 2.3 7.8 to 16.9 Duration epilepsy to T1 (yrs) 3,7 3.0 0.2 to 13.4 Duration epilepsy to T2 (yrs) 6.0 3.1 1.6 to 15.8 Interval T1 T2 (yrs) 2.3 1.2 1.0 to 7.5 Anti-epileptic drugs used at T2 (n = 65) 1.3 0.8 0 to 4 Anti-epileptic drugs tried at T2 (n = 65) 2.3 1.3 0 to 5 Epilepsy syndrome severity, min = 1, max = 10 (n = 70) 6.0 1.5 2 to 8


53

Test Versions

Comparison of the 41 children tested twice with the WISC-RNL with the 32 children

tested twice with the WISC-IIINL revealed overall equality. ANOVA and chi-square (with

alpha set at .01) revealed no significant differences between the subsamples on any IQ

Scale at T1 or T2; on T1–T2 change in IQ; on any demographic variable, namely sex,

handedness, age at T1 or age at T2; or on any epilepsy variable (age at onset, duration of

epilepsy up to T1 or T2, T1–T2 time interval, number of AEDs used, seizure type,

epilepsy syndrome severity, seizure status at T2, MRI status).

Table 4.2.Demographic and epilepsy variables

N %

Sex Boys 38 52.1 Girls 35 47.9

Handedness (n = 63) Right-handed 54 84.1 Left-handed 9 15.9

Seizure type and side Focal 51 69.9

Left hemisphere 24 32.9 Right hemisphere 12 16.4 Bilateral or multifocal 15 20.5

Generalized 6 8.2 Uncertain 11 15.1 Unknown 5 6.8

MRI+ (positive findings) 21 28.8 Seizure status

Active epilepsy 41 56.2 Inactive epilepsy 16 21.9 Uncertain 7 9.6 Unknown 9 12.3

Epilepsy Variables

The characteristics of the sample are presented in Table 4.1 and Table 4.2. Data collected

on epilepsy relate to age at onset of epilepsy; side of seizure onset (left hemisphere onset,

right hemisphere onset, or bilateral seizures); type of seizures (focal versus generalized);

presence of documented brain lesion on neuroimaging (MRI+); and epilepsy syndrome

severity, classified according to the 10-point Syndrome Severity Scale for Children with


54

Epilepsy (ESSS-C, Dunn, Buelow, Austin, Shinnar, & Perkins, 2004). Seizure status was

divided in active or inactive epilepsy; inactive epilepsy was defined as seizure freedom at

T2 of 12 months or longer.

Clinical Control Reference Sample

The use of a referral or clinical control sample rather than a sample of normal controls in

clinical studies has been recommended (Cysique et al., 2011; Woods et al., 2006).

Following this line, to calculate RCI values for the Dutch Wechsler tests, the referral

sample from the test manual (Wechsler, 2005), based on the study of Schittekatte (2005)

was used as a clinical control reference sample.

Table 4.3. Reliable Change Indexes (RCI) based on coefficients of stability RT1T2

American WISC-III Dutch WISCs Observed Expected according to Based on Schittekatte Wechsler Scale Canivez Chelune Maassen RT1T2 RCI RCI RCI RT1T2 RCI

VIQ .87 13 13 13 .87 14 PIQ .87 14 - 15 14 15 .81 18 FS-IQ .91 11 11 11 .88 14

Schittekatte’s study appears as a valuable reference for the present purpose. It is

based on a large sample of 353 children tested with the Dutch Wechsler tests with a mean

interval between testings of 3 years and a mean FS-IQ in the Low Average range (82) at

T1. The sample consisted of children referred for cognitive and learning problems. One

major advantage of the clinical control reference sample is that it does not primarily relate

to children with ongoing neurological conditions possibly associated with cognitive loss.

The previous study differs from the present study in two main aspects – in Schittekatte’s

study (a) a change in test version had taken place from WISC-RNL at T1 to WISC-IIINL at T2,

and (b) the T1–T2 time interval was longer (mean difference of 7.9 months) and had a

wider range (range 0.2 to 8.7 years). Both differences, changes in test version and longer

time intervals, lead to lower stability scores (Sattler, 2001). Therefore, the reference

sample is likely to yield conservative RCI estimates and, therefore, less likely to lead to

type I errors.

Thus, a neurologically “uncomplicated” clinical control reference sample provides

the standard for evaluation of the neurologically compromised sample. If elevated rates of


55

decline are found, as hypothesized, they are more likely to be related to the ongoing

epileptic condition itself and less to other factors like familiarity with the test or non-

specific effects of remediation programs.

Coefficients of stability: comparison between the reference sample and the sample

with epilepsy

For the reference sample, Schittekatte (2005) reports coefficients of T1–T2

stability of .87, .81, and .88 respectively for the verbal, performance and full scales (see

Table 4.3). For the sample with epilepsy, these values were .75, .76, and .77. The

coefficients for the sample with epilepsy were significantly lower than the values found

by Schittekatte for the verbal and full scales, but not the performance scale (using

Fisher’s z for independent samples and alpha set at p < .01 to control for multiple

comparisons).

Establishing and testing Reliable Change Index (RCI) formulas

Similar to Schittekatte (2005), Canivez and Watkins (1998) collected a large

sample of referred children tested twice with the WISC-III. The sample included 667

American children tested for special education eligibility. The authors calculated

coefficients of stability for the VIQ, PIQ and FS-IQ. Notably, the authors also constructed

an empirically-derived base-rate table that presented frequencies of changes in IQ

(Canivez & Watkins, 1998, p.289). These data provided an excellent starting point to test

RCI formulas.

Chelune et al. (1993, pp 45-46) published a RCIs formula based on the

coefficients of stability and the standard deviation from the first test administration.

Maassen et al. (2009, p.340, formula 2) expanded the formula to include the standard

deviation of both the first and second testings. To compare the formulas from the two

studies, they were applied to the coefficients of stability from Canivez and Watkins

(1998, p.287). As can be seen in Table 4.3, both formulas provided accurate estimates:

they yielded RCI values within 1 point of those found empirically by the authors (Canivez

& Watkins, 1998, p.289). In the present study, the formula from Maassen et al. was

applied as follows:

SEDIFF = SDT12 SDT 2

2 1 RT1T 2


56

where SDT12 and SDT2

2 are the squared standard deviations at T1 and T2, respectively, and

RT1T2 is the stability coefficient between measurements at T1 and T2. The value for the

90% confidence interval: ± (SEDIFF * 1.64).

Cut-off values for the RCIs

After applying the formula to Schittekatte’s coefficients of stability, the

established RCI cut-off values were 14 for VIQ, 18 for PIQ, and 14 for FS-IQ (rounding

off to the nearest integer). The last two columns of Table 4.3 present the stability

coefficients and the RCIs used to establish reliable change in the present study. Chelune

et al. (1993) suggested that the differences between T1 and T2 be adjusted with the T1–

T2 changes found in the reference sample to control for practice effects. Given the

intertest interval of at least 12 months for the present sample, no practice effects were

expected.

Analyses

Temporal stability (mean change) was calculated with paired-samples t tests for the IQ

scales. Two-tailed tests were applied, alpha was set to .01 to correct for multiple

comparisons, and r was used as a measure of effect size; r >= .5 was interpreted as a large

effect size (Field, 2005). The percentage of children with epilepsy outside the 90%

reliable-change interval was established. For each scale, a single chi-square statistic was

used to test whether the observed rates for no reliable change (±RCI), gains (>= +RCI),

and losses (<= –RCI) differed from: the expected rates of 90% (“normal gain or loss”),

5% (“reliable gain”), and 5% (“reliable loss”). Again, alpha was set at .01.

Results

Changes over Time on Wechsler Scales

As presented in Table 4.4, mean decline between T1 and T2 on the verbal scale

was 7.2 IQ points; a paired-sample t test indicated this was significant (p < .001, effect

size r = .5). The performance scale showed a non-significant decline of 0.6 IQ points. The

full scale showed a significant decline of 4.4 IQ points (p = .001, r = .4). Changes ranged

from –52 (loss of 52 IQ points) to +21 (gain of 21 points) on the verbal scale, –38 to +29

on the performance scale and –48 to +17 on the full scale.


57

Reliable Cognitive Change in Epilepsy

Table 4.4 presents the rates of children showing reliable gains and losses, as well

as the results of the chi-square analyses for the three IQ scales. The table shows that the

percent of children who displayed a reliable gain on any IQ scale is close to the expected

5%, while the rate of children with significant loss is elevated in relation to the expected

5% on the verbal and full scales. On the verbal scale, 53 children (72.6%) presented with

T1–T2 differences within the reliable change interval; one child (1.4%) showed a reliable

gain; and a substantial 19 children (26.0%) showed a reliable loss. The proportions

differed significantly from the expected values (Χ2 = 68.93, p < .001). The values found

for the performance scale – gain in four children (5.5%) and loss in four children (5.5%)

– were not different from those expected (Χ2 = 0.07, p > .01). On the full scale, two

children (2.7%) showed a gain and 12 (16.4%) a loss; values were significant (Χ2 =

20.53, p < .001).

Discussion

The aim of this study was to establish reliable cognitive change in children with epilepsy

tested twice with the WISC-RNL/WISC-IIINL, applying a 90% RCI. The present study differed

from earlier research on cognitive change in child epilepsy (Aldenkamp, Alpherts, De

Bruine-Seeder, & Dekker, 1990; Westerveld et al., 2000) as it considered the

psychometric properties of IQ changes on the Wechsler Scales to predetermine RCIs. It

was shown that observed changes could be predicted with great accuracy, providing

support for the usefulness of these formulas to determine cognitive change and

confirming that empirical data on change can be estimated from psychometric data

(Chelune et al., 1993; Maassen et al., 2009). Lower coefficients of stability found in the

sample with epilepsy compared to the reference sample suggested more variability in IQ

scores in children with epilepsy from T1 to T2. Indeed, compared to the expected values

of 5%, elevated rates of reliable loss at T2 were seen in epilepsy on the verbal scale

(26.0%) and on the full scale (16.4%). Rates for gains did not exceed expected values on

any IQ-scale. The present data support the hypothesis that the seizure condition is

associated with an elevated risk for cognitive decline (Dodrill, 2004; Seidenberg,

Pulsipher, & Hermann, 2007). Notably, this decline is only seen on the verbal and full

scales and not on the performance scale.

C

OG

NIT

IVE

PATT

ERN

S IN

PA

EDIA

TRIC

EPI

LEPS

Y

Tabl

e 4.

4. W

echs

ler I

Q sc

ores

at T

1 an

d T2

and

cha

nge

over

tim

e (

T2-T

1) w

ith p

aire

d-sa

mpl

es t

test

and

per

cent

ages

of c

hild

ren

show

ing

relia

ble

gain

s, no

relia

ble

chan

ge a

nd re

liabl

e lo

sses

IQ

at T

1

IQ a

t T2

Δ

T1 -T

2

Rel

iabl

e ch

ange

Mea

n SD

Mea

n SD

Mea

n SD

t

p ES

( r )

Gai

n (5

%)

n.s.

(90%

) Lo

ss

(5%

) Χ

2 p

VIQ

90.8

9 15

.5

83

.68

16.3

7.21

11

.3

5.43

5 <

.001

.5

1.4

72.6

26

.0

68.9

93 <

.001

PI

Q

87

.45

16.3

86.8

6 17

.8

0.

59

11.9

0.

423

n.s.

-

5.5

89.0

5.

5 0.

075

n.s.

FS-IQ

88.1

0 15

.9

83

.73

17.0

4.37

11

.1

3.35

8 .0

01

.4

2.

7 80

.8

16.4

20

.531

<.0

01

58


59

The differential changes for the verbal and performance scales may be partially

related to the higher initial VIQ than PIQ. It is unlikely, however, that they were related

to regression to the mean phenomenon (Heaton et al., 2001), which more often affects

scores in the extremes, because the present sample had low average IQs at first testing. A

VIQ > PIQ difference has been reported regardless of seizure onset side (van Iterson &

Augustijn, 2006). A somewhat more favourable course of the performance scale than the

verbal scale at retesting has also been described in samples without epilepsy (Kaufman,

1994; Schittekatte, 2005), as well as in children who have undergone epilepsy surgery

(Westerveld et al., 2000), and is possibly due to the decreased “novelty” of the

performance tasks at T2 (Canivez & Watkins, 1998).

Epilepsy variables

Seidenberg et al. (2007) state that there is a dearth of studies on cognitive decline,

particularly studies that include epilepsy variables. The breakdown of the present sample

according to epilepsy variables yielded small subsample sizes, and did not permit

conducting meaningful analyses. Earlier longitudinal studies on children with epilepsy

suggested that patterns of change were independent of seizure laterality (Westerveld et

al., 2000), type of epilepsy and antiepileptic drugs (Oostrom, van Teeseling, Smeets-

Schouten, Peters, & Jennekens-Schinkel, 2005). Mixed results have been reported on

cognitive development over time in relation to persistence of seizures (Bjornaes, Stabell,

Henriksen, & Loyning, 2001; Jones, Siddarth, Gurbani, Shields, & Caplan, 2010;

Oostrom et al., 2005).

Limitations of the Study

The sample in the present study was available from tertiary epilepsy settings, which, by

nature, deal with more difficult to treat epilepsies and therefore possibly with children

with worse prognosis. The lower age at T1 of the children included in the sample,

compared to those not selected, also points to the inclusion of “worse” epilepsies (Bulteau

et al., 2000) and may limit the generalizability of the results.

Future Directions

Future studies with larger samples may allow insight into the possible impact of epilepsy

variables on cognitive decline. Alternatively, more specific epilepsy subsamples (for


60

example, specific epilepsy syndromes) could be studied with similar methods. Also, given

that epilepsy surfaces at various ages and may last for a prolonged period of time, it is

important to obtain data on reliable cognitive when different intelligence tests are used at

T1 and T2 versions (for example, a childhood version of Wechsler’s scales followed by

an adult version).

Clinical Implications

The results of the present study contribute to the literature on the cognitive course of

epilepsy in children and should be of value for clinicians and researchers. The procedure

described for the Dutch Wechsler tests can readily be applied to other languages and

cultures, provided that coefficients of stability and standard deviations are available from

a reference sample. Clinicians may want to apply the reliable change cut-off values when

retesting a child. For the American WISC-III, and for the Dutch Wechsler tests, the data

presented in Table 4.3 would yield appropriate estimates. After a change in test version

(WISC-RNL to WISC-IIINL), adjustment with the differences reported by Schittekatte (2005) is

pertinent; applying the more stringent criteria of 19 (VIQ), 18 (PIQ) and 17 (FS-IQ)

points for reliable loss is recommended to account for the predictable Flynn effect. The

present study used a statistically sound methodology to help addressing the question, “Is

this child presenting a reliable cognitive change at retesting?” The applicability of the

procedure goes beyond children with epilepsy to other neurodevelopmental disabilities

potentially associated with cognitive decline.

CHAPTER 5

Duration of epilepsy and cognitive development in children:

A longitudinal study

Loretta van Iterson

Bonne H.J. Zijlstra

Aryan van der Leij

Peter F. de Jong


62

Abstract

Objective: To study the pattern of cognitive development in relation to duration of

epilepsy.

Methods: Participants were 113 children with epilepsy referred because of concerns about

their cognitive development and tested at least twice at tertiary epilepsy settings. Verbal,

performance and full-scale IQ were measured with Wechsler Intelligence Scales. Various

epilepsy and demographic variables were included. Change over time was modelled with

multilevel analysis for longitudinal data with variable measurement occasion.

Results : The verbal and full scales could be fitted best as a downward progressing

function. Earlier in time, decline was likely to be largest; later in time, decline followed a

continuous, dwindling course. A similar trend was seen for the performance scale.

Initially, verbal IQ was higher than performance IQ but this discrepancy decreased over

time. Later onset of epilepsy was associated with an attenuated decline of the verbal

Scale. None of the other epilepsy variables were related to the course of cognitive

development. Higher parental education was associated with higher IQ, but was not

protective against decline.

Conclusions: verbal IQ, though initially spared, drops. The performance IQ, which may

have shown its vulnerability earlier in the course of the epilepsy, shows overall smaller

changes. It is suggested that seizures impact synergistically on an affected brain, which

leads to progressive cognitive decline. Earlier onset of epilepsy is associated with

relatively higher VIQ, larger VIQ > PIQ discrepancies and more decline.

van Iterson, L., Zijlstra, B. J., Augustijn, P. B., van der Leij, A., & de Jong, P. F. (2014). Duration of epilepsy and cognitive development in children: a longitudinal study. Neuropsychology, 28(2), 212-221.

DURATION OF EPILEPSY AND COGNITIVE DEVELOPMENT

63

Introduction

There is little doubt that epilepsy in childhood has an adverse impact on a child’s life

(Hermann, Jones, Jackson, & Seidenberg, 2012). Outcome is varied in terms of seizure

control (Geerts et al., 2010) and cognitive development (Berg et al., 2008). Follow-up

studies indicate that 50 to 60% of patients with epilepsy have a favourable course and

achieve seizure freedom after use of anti-epileptic drugs (AED) (Geerts et al., 2010;

Schmidt & Sillanpää, 2012). Neuropsychological studies on “uncomplicated epilepsies”

have shown a close to normal cognitive development over time in children with epilepsy

(Hermann, Seidenberg, & Jones, 2008; Jones, Siddarth, Gurbani, Shields, & Caplan,

2010). These studies relate to children with epilepsy who are not referred for

psychological assessment, who show seizure amelioration with or without medication;

and who do not have co-occurring problems like brain lesions or attention problems.

Generally, children attend regular classes, although school problems have been reported

in about half of the children with epilepsy (Reilly & Neville, 2011).

In a considerable proportion of children (~30%), however, epilepsy is not

uncomplicated, in terms of seizure control, cognitive development, or both. After 15 years

of follow-up, ~10% of the children with epilepsy never had been seizure free longer than

3 months, and an additional ~13% showed a varying course of remissions followed by

relapses (Geerts et al., 2010). Cognitive impairment has often been described in epilepsy

in children (Ellenberg, Hirtz, & Nelson, 1986; Nolan et al., 2003). In a community based

study it was shown that, 10 years after seizure onset, ~26% of children with epilepsy had

an estimated IQ below 80 (Berg et al., 2008).

This raises a number of questions: What is the developmental course of cognitive

functioning over time? Can evidence be found for cognitive decline? Is cognitive decline

associated with age at onset? Can epilepsy and demographic variables be identified which

affect cognitive development? Which area of cognitive development – verbal or

nonverbal – is likely to be affected most? Do the verbal and nonverbal domains follow

similar trajectories?

There is a body of research regarding epilepsy factors that contribute to the

severity of cognitive impairment. Apart from persistence of seizures (Bailet & Turk,

2000; Berg, Zelko, Levy, & Testa, 2012), epilepsy syndrome is recognized as an

important factor. Generalized symptomatic epilepsies are associated with low IQs (Berg

et al., 2008; Bulteau et al., 2000; Nolan et al., 2003). Localization related epilepsies and


64

idiopathic epilepsies are associated with relatively better outcome (Bulteau et al., 2000;

Nolan et al., 2003; Northcott et al., 2007).

Early age at onset of epilepsy (AOE) has been related to worse outcome (Cormack

et al., 2007), especially when the seizures remain active (Berg et al., 2012). Also, greater

number AEDs (Bulteau et al., 2000; Nolan et al., 2003; Smith, Elliott, & Lach, 2002),

have been associated with more cognitive impairment. Demographic factors like parental

educational level have been acknowledged as being strongly associated with children’s

IQ in children without epilepsy (Lange, Froimowitz, Bigler, Lainhart, & Brain

Development Cooperative, 2010), as well as in children with epilepsy (Mitchell, Scheier,

& Baker, 1994).

In cross-sectional studies, cognitive problems have been described both for the

verbal and nonverbal (performance) domains. Studies on specific syndromes have

reported lowered verbal IQ (Overvliet et al., 2011); studies on mixed samples have

reported lowered nonverbal IQ (Høie et al., 2005; O'Leary, Burns, & Borden, 2006;

Smith et al., 2002). Longitudinal studies focussing on the relation of the verbal and

nonverbal (i.e. performance) domains are scarce. Changes have been reported to be small

and similar for both the verbal and performance IQ (Aldenkamp, Alpherts, De Bruine-

Seeder, & Dekker, 1990; Bjornaes, Stabell, Henriksen, & Loyning, 2001). Studies on

children who underwent epilepsy surgery report increases on the performance scale,

regardless of hemispheric side of surgery (Skirrow et al., 2011).

The pattern of cognitive impairment and cognitive change over time in children

with epilepsy is still insufficiently understood. Models have been proposed describing

cognitive decline as either gradually progressive (“linear”) or, as a stepwise (“cascadic”)

decline. A cascadic decline is described as marked in the early stages of epilepsy and

plateauing thereafter (Devinsky & Tarulli, 2002; Meinardi, Aldenkamp, & Nunes, 1992;

Seidenberg, Pulsipher, & Hermann, 2007). There is still a dearth of studies to substantiate

these models on cognitive decline over time and there is still a need for a finer

characterization of the course of development in children with epilepsy (Hermann, Jones,

Jackson, & Seidenberg, 2012). This study examined the developmental trajectory of

cognitive decline in children with epilepsy. More specifically, the course of epilepsy –

without intervening epilepsy surgery – over time was considered, based on cognitive data

on children at a Dutch tertiary epilepsy center. The children were tested two or three

successive times with the Wechsler Intelligence Scales.


65

Various methodological problems need to be considered. These problems stem

from the heterogeneity of epilepsies in terms AOE, IQ and course (Camfield, Camfield,

Gordon, Smith, & Dooley, 1993; Schmidt & Sillanpää, 2012). Inexorably, this means a

large variability in time elapsed between epilepsy onset and time of first, second, or even

third neuropsychological testing. As children grow older, they are likely to make a

transition from one Wechsler test to another, implying variability in test versions used.

Multilevel modelling (Snijders & Bosker, 1999), a special statistical technique, was

applied to account for these differences.

Methods

Participants

From 452 Dutch children with epilepsy who had completed a Wechsler test, 113

were selected as they met the criteria for inclusion. Children were selected if they (1)

were 4 to 15 years of age at first testing (T1), (2) had been tested at least two times with

age-appropriate child Wechsler tests with an intertest interval of one year or longer, and

(3) had not had intervening epilepsy surgery in between testing. The children presented

either at a Dutch tertiary epilepsy centre or at a special school affiliated with the centre,

which provided educational support for children with epilepsy. Reasons for T1 were

concerns about the cognitive development of the child. Commonly results of the

assessment were also used in applications for special financial and educational services to

support the child within a regular or a special school. Reasons for second and third testing

(T2, T3) were requests for follow-up as the epilepsy evolved and retesting was required

for continuation of the educational support. No exclusionary criteria were set for type of

epilepsy or initial IQ. Overall, there were 249 Wechsler test measurements: 113 at T1,

113 at T2, and 23 at T3.

The sample tested 3 times (n = 23) did not differ from the sample tested 2 times (n = 90)

on any IQ scale at T1 or T2, at any epilepsy variable (AOE, duration up to T1 or T2,

epilepsy type or syndrome severity), or demographic variable (sex, handedness, parental

education).

The 339 children not selected for the study had been tested only once (n = 303);

had been tested with an age-inappropriate test (n = 3); had been retested within a year (n

= 22); had been retested with another Wechsler Scale (WAIS/WAIS-IIINL n = 5, other n = 1);

had undergone surgery (n = 2) or had missing values on an IQ scale (n = 2); other reasons


66

(n = 1). Comparison of selected and non-selected children (after Bonferroni adjustment)

showed similarities in terms of IQ at T1, seizure type and syndrome severity or AOE.

However, duration up to T1 and age at T1 was higher in the non-selected sample (n = 19

had been tested with the WAIS/WAIS-IIINL).

Wechsler test versions

The study concerns the scaled scores of the Wechsler Intelligence Scales in the

Netherlands, here designated as WPPSI-RNL (Vander Steene & Bos, 1997), WISC-RNL (van

Haasen et al., 1986) and WISC-IIINL (Wechsler, 2005) allowing test changes between T1

and T2 or T3.

Other measures

Epilepsy variables. The epilepsy variables were available from neurological or

neuropsychological reports and relate to information as documented at last testing. AOE,

seizure type, onset side and topographical localization, presence of MRI lesion, and

number of AEDs tried in the course of the epilepsy were included. Seizure status was

scored as active or inactive (seizure freedom of at least one year), uncertain or unknown.

The Syndrome Severity Scale for Children with Epilepsy (ESSS-C; Dunn, Buelow,

Austin, Shinnar, & Perkins, 2004) was used to measure epilepsy syndrome severity. This

10-point scale encompasses various epilepsy variables such as seizure type, aetiology and

AOE within a single scale. The present sample included syndrome severity scores ranging

from 2 to 9: idiopathic localization related epilepsy (benign epilepsy with centrotemporal

spikes [BECTS], n = 4, 3.5%, score 2); localization related symptomatic epilepsy (by

virtue of aetiology, n = 19, 16.8%, score 7; by virtue of localization n = 36, 31.9%, score

5; cryptogenic n = 5, 4.4%, score 6.55); idiopathic generalized epilepsies (childhood

absence epilepsy [CAE], n = 2, 1.8%, score 3; other n = 10, 8.8%, score 5); symptomatic

generalized epilepsies (n = 6, 5.3%, score 8-9); epilepsy syndromes (epilepsy with

continuous spikes and waves during slow sleep [CSWS] and atypical BECTS, n = 14,

12.4%, score 8; other (n = 10, 8.8%, scores 2-6); unknown (n = 7, 6.2%).

Demographic variables. The educational status of the child was dichotomized as

regular education (with special facilities) or special school placement. For parental

education, the highest educational level completed (CBS, 2007) was averaged across

parents. Sex and handedness were also included.


67

Analyses

The trajectories over time of verbal, performance, and full-scale IQs and VIQ – PIQ

discrepancy were estimated with a multilevel model for longitudinal data with variable

measurements occasions (Snijders & Bosker, 1999). An advantage of this model is that it

allows any number of repeated observations for each subject, without restrictions on the

temporal spacing between the measurements. Time was taken to be the duration since the

first epileptic seizure. Test versions were modelled with separate dummy variables for the

WPPSI-RNL and the WISC-IIINL, taking the WISC-RNL as reference. The predictors time and

test version could change over repeated observations, whereas the demographic and

epilepsy variables were constant for each child. For these variables the effect on

individual differences in IQ level (regardless of time), as well as the effect on individual

differences in the rate of change in IQ over time were estimated. Individual differences

unaccounted for by the predictors were modelled with a random intercept. The standard

deviation of the random intercept indicates the amount of residual differences in IQ level.

A random slope for residual differences in the rate of change in IQ over time was not

included because these models could not be estimated (the models were not identified).

The selection of an appropriate model to fit the trajectories of the IQ scores over

time was done in three steps, applying a statistical significance level of .01. Results of the

first step led to the Base Model, results of the last step to the Final Model.

In the first step, an adequate model for the change over time was sought, entering

duration of epilepsy as well as Wechsler test version. Curvilinear functions (quadratic,

logarithmic and square root) of duration of epilepsy in months (and months plus one for

the logarithmic transformation) were added to the linear model to check for a significant

increase in model fit. To find the most parsimonious model, a reverse strategy was also

applied by dropping the linear component whenever this did not significantly decrease the

model fit. Models with the same number of parameters were compared on Bayes factors

(Kass & Raftery, 1995), approximated from the Schwarz criterion, to assess the evidence

in favour of the best fitting model. This model was called the Base Model.

In the second step, time and Wechsler test version were maintained and AOE (in

months) and the demographic variables were included in the model: special education,

parental education, and the dummy variables handedness (left-handedness was coded 1),

and sex (boy was coded 1). For these predictors, the effect on the individual differences in

IQ level were always included in the model. The effect on individual differences in the


68

rate of change in IQ over time were only included whenever they reached statistical

significance. For AOE and parental education, the overall means were set to zero.

Table 5.1. Characteristics of the sample

N mean SD range Full sample 113 Age at onset of epilepsy (AOE) 113 4.8 3.0 0.1 to 13.2 Age at T1 113 8.4 2.3 4.7 to 15.0 Age at T2 113 11.2 2.7 5.8 to 16.9 Age at T3 23 12.9 2.7 6.9 to 16.8 Duration epilepsy to T1 113 3.5 2.6 0.2 to 12.2 Duration epilepsy to T2 113 6.3 3.1 1.6 to 15.8 Duration epilepsy to T3 23 8.6 3.5 3.3 to 16.2 AEDs tried 102 2.5 1,7 0 to 12 Epilepsy syndrome severity 106 5.9 1.6 2 to 8 Parental education 105 4.4 0.9 2.7 to 6.0 N n % Seizure type 113 Generalized seizures 15 13.3 Focal (all focal) 75 66.4 Focal: LH / RH 33 / 17 29.2 / 15.0 Bilateral of mutifocal 25 22.1 Uncertain 16 14.2 Unknown 7 6.2 MRI - / MRI + 113 82 / 31 72.6 / 27.4 Boys / girls 113 61 / 52 54 / 46 Education: regular / special 112 61 / 51 54.5 / 45.5 Test versions Test version at T1 113 WPPSI-RNL 20 17.7 WISC-RNL 63 55.8 WISC-IIINL 30 26.5 Test version at T2 113 WPPSI-RNL 3 2.7 WISC-RNL 46 40.7 WISC-IIINL 64 56.6 Test version at T3 23 20.4 WPPSI-RNL 1 0.9 WISC-RNL 5 4.4 WISC-IIINL 17 15.0

Note. T1 = test 1, T2 = test 2, T3 = test 3, AEDs tried = number of anti-epileptic drugs tried, LH = left hemisphere, RH = right hemisphere, MRI + = lesion found on neuroimaging, MRI– = no lesion found on neuroimaging or no neuroimaging available.


69

In the third step, the predictors from the first two steps were maintained and the epilepsy

variables were included in the model. The predictors entered were focal seizures,

generalized seizures, left hemisphere seizures, right hemisphere seizures, frontal seizures,

ESSS-C score of epilepsy syndrome severity, seizure freedom, number of AEDs tried,

MRI lesion (and its interaction with handedness). In this step, the effects of the predictors

on individual differences in IQ level and on individual differences in the rate of change in

IQ over time were included in the model only whenever they reached statistical

significance. However, for the latter effect (change over time) statistical significance had

to be established taking into account the (possibly non-significant) effect of the predictor

on individual differences in IQ level.

Table5.2 Wechsler Intelligence Scale data at different measurement points

Full sample Subsample Scale T1 T2 T1 T2 T3 Mean SD Mean SD Mean SD Mean SD Mean SD VIQ 89.3 15.4 81.5 15.9 85.7 11.8 73.1 13.0 71.4 11.9 PIQ 84.3 17.0 82.1 18.0 80.3 15.4 76.5 15.9 72.3 16.2 FS-IQ 85.5 16.0 79.8 16.9 81.5 12.8 73.7 11.7 69.3 13.0 VIQ - PIQ 5.0 14.1 -0.6 14.1 5.7 14.8 -0.3 18.4 -0.9 14.0

Note. The Full sample was based on n = 113. The subsample included the 23 children who had been adminstered the Wechsler three times.

Results

Table 5.1 shows the characteristics of the sample and Table 5.2 the unadjusted mean IQs.

Results for the longitudinal multilevel models are presented in Table 5.3 (for the verbal

scale, performance scale, and full scale) and Table 5.4 (VIQ – PIQ discrepancy) and

include the Base and Final Model. None of the epilepsy predictors could be added to the

final models in the third stage of model selection. Therefore, the Final Model comprises

the results of the second stage.

Figure 2.1 shows the approximated predicted outcomes according to the Base Model

for the verbal, performance and full-scale IQs for the middle 95 percent of the observed

durations of epilepsy (i.e. ranging from 8 to 146 months). The figure shows a strong

decline initially, leveling off with increasing duration of the epilepsy. Bayes factors for

the performance scale and full scale indicated positive to very strong evidence (Kass &,

C

OG

NIT

IVE

PATT

ERN

S IN

PA

EDIA

TRIC

EPI

LEPS

Y

Ta

ble

5.3.

Res

ults

of m

ultil

evel

ana

lysi

s for

the

Bas

e M

odel

(Mod

el) 1

and

the

Fina

l Mod

el.

Ver

bal S

cale

Pe

rfor

man

ce S

cale

Fu

ll Sc

ale

Estim

ate

S.E.

p-

valu

e 95

% C

I Es

timat

e S.

E.

p-va

lue

95%

CI

Estim

ate

S.E.

p-

valu

e 95

%C

I

Bas

e M

odel

Fixe

d ef

fect

s Inte

rcep

t 11

5.57

3.

86

<.00

1 (1

08.0

, 123

.2)

103.

16

4.50

<

.001

(9

4.3,

112

.0)

109.

35

3.98

<

.001

(1

01.5

,117

.2)

Tim

e: lo

garit

hmic

-7

.56

0.97

<.

001

(-9.

5, -5

.6)

-4.9

7 1.

14

< .0

01

(-7.

2, -2

.7)

-6.6

6 1.

00

< .0

01

(-8.

6, -4

.7)

WIS

C-II

INL

-3.0

4 1.

72

.078

(-

6.4

, 0.3

) -1

.93

2.01

.3

37

(- 5

.8, 2

.0)

-2.9

0 1.

77

.104

(-

6.4

, 0.6

)

WPP

SI-R

NL

-1.1

7 2.

51

.641

(-

6.1,

3.7

) -2

.99

2.98

.3

17

(-8.

9, 2

.9)

-1.7

5 2.

60

.502

(-

6.9,

3.4

)

Ran

dom

eff

ects

In

terc

ept S

D

13.2

9 13

.85

13

.54

R

esid

ual S

D

7.53

9.

19

7.

82

Dev

ianc

e 19

44.0

0

20

12.7

9

19

58.9

6

70

DU

RA

TIO

N O

F EP

ILEP

SY A

ND

CO

GN

ITIV

E D

EVEL

OPM

ENT

Ta

ble

5.3.

Res

ults

of m

ultil

evel

ana

lysi

s for

the

Bas

e M

odel

(Mod

el) 1

and

the

Fina

l Mod

el (c

ontin

ued)

.

V

erba

l Sca

le

Perf

orm

ance

Sca

le

Full

Scal

e

Estim

ate

S.E.

p-

valu

e 95

% C

I Es

timat

e S.

E.

p-va

lue

95%

CI

Estim

ate

S.E.

p-

valu

e 95

%C

I

Fina

l Mod

el

Fixe

d ef

fect

s Inte

rcep

t 12

4.71

5.

58

< .0

01

(113

.7, 1

35.7

) 98

.24

5.77

<

.001

(8

6.9,

109.

6)

106.

75

4.77

<

.001

(9

7.3,

116

.1)

Tim

e: lo

garit

hmic

-9

.34

1.30

<

.001

(-

11.9

, -6.

8)

-3.2

8 1.

37

.018

(-

6.0,

-0.6

) -5

.62

1.12

<

.001

(-

7.8,

-3.4

)

WIS

C-II

INL

-2.1

3 1.

66

.200

(-

5.4,

1.1

) -2

.20

2.16

.3

11

(-6.

4, 2

.1)

-2.4

9 1.

80

.170

(-

6.0,

1.1

)

WPP

SI-R

NL

-5.0

9 2.

65

.056

(-

10.3

, 0.1

) -1

.38

3.28

.6

74

(-7.

8, 5

.1)

-1.6

8 2.

69

.534

(-

7.0,

3.6

)

Age

at o

nset

-0

.33

0.12

.0

07

(-0.

6, -0

.9)

0.15

0.

05

.001

(0

.1, 0

.2)

0.08

0.

04

.059

(-

0.00

3, 0

.2)

Pare

ntal

edu

catio

n 6.

54

1.36

<

.001

(3

.8, 9

.2)

4.25

1.

63

.011

(1

.0, 7

.5)

5.79

1.

42

< .0

01

(3.0

, 8.6

)

Le

fthan

dedn

ess

-3.5

8 2.

95

.228

(-

9.4,

2.3

)

-4.7

9 3.

52

.178

(-

11.8

, 2.2

)

-4.2

4 3.

07

.170

(-

10.3

,1.9

)

Sex

(boy

) 5.

96

2.34

.0

12

(1.3

, 10.

6)

2.56

2.

80

.362

(-

3.0,

8.1

) 4.

46

2.44

.0

70

(-0.

4, 9

.3)

Spec

ial e

duca

tion

-9.1

8 2.

49

<.00

1 (-

14.1

, 4.2

) -6

.08

2.98

.0

44

(-12

.0, -

0.2)

-8

.56

2.60

.0

01

(-13

.7, -

3.4)

Ti

me:

loga

rithm

ic b

y A

OE

0.08

0.

03

.005

(0

.02,

0.1

)

R

ando

m e

ffec

ts

Inte

rcep

t SD

9.

66

11.2

7 10

.04

Res

idua

l SD

6.

87

9.52

7.

70

D

evia

nce

152

0.60

1

627.

54

155

4.47

Not

e. S

.E. =

Sta

ndar

d Er

ror.

SD =

stan

dard

dev

iatio

n

71


72

Table 5.4. Results of multilevel analysis for the VIQ – PIQ discrepancy.

VIQ - PIQ Estimate S.E. p-value 95%CI

Base Model

Fixed effects Intercept 9.06 2.63 .001 (3.9, 14.2)

Time: square root -0.93 0.33 .005 (-1.6, -0.3)

WISC-IIINL -0.63 1.80 .726 (-4.2, 2.9)

WPPSI-RNL 1.49 2.61 .570 (-3.7, 6.6) Random effects Intercept s.d. 11.84 Residual s.d. 8.11 Deviance 1944.38 Final Model Fixed effects Intercept 12.87 3.61 < .001 (5.7, 20.0)

Time: square root -1.53 0.40 < .001 (-2.3, -0.7)

WISC-IIINL 0.64 1.95 .745 (-3.2, 4.5)

WPPSI-RNL -1.64 2.82 .562 (-7.2, 3.9)

Age at onset -0.17 0.04 < .001 (-0.3, -0.1)

Parental education 1.85 1.52 .227 (-1.2, 4.9)

Lefthandedness 1.64 3.30 .620 (-4.9, 8.1)

Sex (boy) 2.99 2.61 .255 (-2.2, 8.2)

Special education -3.55 2.79 .206 (-9.1, 2.0) Random effects Intercept s.d. 10.85 Residual s.d. 8.01 Deviance 1576.42


73

Raftery, 1995) for a logarithmic decline compared to the square, linear and square root

decline. For the verbal scale there was nearly as much evidence for square root decline,

indicating that the decline leveled off to a slightly lesser degree, but there was strong

evidence against linear and square decline.

In the Final Model, results were overall similar for the verbal and full scales. The

magnitude of estimates of the effect of time was comparable between the models for VIQ

and FS-IQ. The positive effect for parental education in Table 5.3 suggested that higher

parental education was associated with higher VIQ and FS-IQ. No effect of parental

education on individual differences in change over time was found, suggesting that lower

parental education did not imply an increased risk of decline over time. Similarly, special

education was associated with lower VIQ and FS-IQ, while no significant effect for

change over time was seen. For the VIQ, an effect was seen for AOE, also in interaction

with time. The (negative) values on time and AOE indicated that a longer duration and a

later onset were associated with a lower VIQ. The (positive) value for the interaction of

AOE and time showed that the decline of the VIQ over time was somewhat less

pronounced for children with a higher AOE. No other epilepsy variable made a

significant contribution to the models.

For the performance scale, a similar albeit non-significant effect of time could be

found in the Final Model, compared to the Base Model. A positive effect could be

observed for AOE (Table 5.3), meaning that children with later AOE (above the mean of

the sample) were likely to have a slightly higher PIQ score. A positive effect was seen for

parental education also.

The intercept for VIQ – PIQ differed from zero (Table 5.4); the positive value

indicated that the difference favoured the verbal scale (VIQ > PIQ). Change over time

presented as a square root curve, although Bayes factors for the Base Model suggested

that there was almost as much evidence for a linear or logarithmic downward slope. The

negative value for time suggested that the VIQ – PIQ discrepancies decreased over time.

The negative value for age of onset implied that a younger AOE was associated with

larger VIQ > PIQ discrepancies.

For all models, the standard deviations of the random intercepts were larger than

the residual standard deviations, implying that the differences not accounted for by the

predictors in the models between the children were larger than the residuals around the

predicted individual trajectories. Therefore, the model estimates indicated there was less

uncertainty about the individual trajectories (and their shape) than about the IQ levels of


74

children (regardless of time). No significant effect of Wechsler test version was seen on

any scale. Large differences between individual trajectories can be found.

Figure 5.1. Approximated effect of time according to the Base Model for Verbal IQ, Performance IQ, and Full Scale IQ.

Discussion

This study provided evidence for progressive cognitive decline over time in clinically

referred children with epilepsy. Decline was largest in the early stages of epilepsy;

thereafter, decline continued at an increasingly slower pace. Also, a differential trajectory

for VIQ and PIQ was seen, which was not described earlier. The curve described was

logarithmic – not linear. On an individual bases, cascadic decline cannot be excluded.

Large individual variation was found.

Different trajectories for the Verbal and Performance IQ

Previous studies have described lower IQ for very early AOE (Bulteau et al.,

2000; Cormack et al., 2007). The present study points towards differential impact on the

various scales: earlier AOE was associated with a better (initial) VIQ and a more


75

pronounced decline – therefore a worse trajectory. Earlier AOE was associated with

particularly lowered PIQ and larger VIQ > PIQ discrepancies.

The VIQ > PIQ gap was seen early in the course of the epilepsy, and closed over

time. The results suggested that verbal IQ was “spared” initially and declined over time,

while performance IQ possibly showed its vulnerability early in the course of the epilepsy

and showed an attenuated decline later on. Future research including children tested

before epilepsy onset (for example, children with an increased genetic risk for developing

epilepsy), could be directed at elucidating whether the VIQ > PIQ discrepancy exists

already prior to the onset of epilepsy, or emerges together with the seizure condition.

As in the present study, some evidence for less decline (or more gains) at retesting

for PIQ rather than VIQ has been given in samples without intervening surgery

(Aldenkamp, Alpherts, De Bruine-Seeder, & Dekker, 1990), after epilepsy surgery

(Skirrow et al., 2011; Westerveld et al., 2000), and in the light of amelioration of seizures

(van Mil et al., 2010). Part of these effects may be interpreted in the light of studies of

children without epilepsy, where the performance scale has been shown to be more prone

to profit from practice effects or test familiarity (Canivez & Watkins, 1998). The impact

of test familiarity in the present study should be limited, given the interval of a year or

longer between testings.

Variables contributing to cognitive level and cognitive change

Epilepsy variables.

Similar to earlier studies on heterogeneous samples (Reijs et al., 2007; Strauss et

al., 1995), no epilepsy variable other than age of epilepsy onset and duration of epilepsy,

could be singled out as contributing significantly to cognitive level or to cognitive change

over time. This is particularly puzzling concerning variables like epilepsy syndrome

severity and underlying symptomatology. Several issues should be pointed out regarding

the variables studied. First, the results may challenge the utility of the syndrome severity

scale as used in this study. In fact, various authors have indicated that the best way to

determine epilepsy syndrome severity is still under debate and that syndrome severity

should be studied in combination with cognitive outcome (Dunn et al., 2004; Reijs et al.,

2006; Wirrell, Grossardt, So, & Nickels, 2011). Second, as Elger, Helmsteadter and

Kurthen (2004) pointed out, aetiology and AOE are difficult to disentangle, because

specific disorders peak at certain age groups (Wirrell, Grossardt, Wong-Kisiel, & Nickels,

2011). This means that the findings related to early AOE may be seen as valid for early


76

onset aetiologies. Third, some cases with MRI-negative findings are reclassified as

positive cases after MRI reevaluation (Funke, Moore, Orrison, & Lewine, 2011). Changes

in MRI-interpretation have implications for the reliability of the distinction between

symptomatic and nonsymptomatic aetiologies and consequent syndrome classification.

Fourth, seizure freedom may be temporary and may be followed by relapse (Schmidt &

Sillanpää, 2012). Fifth, AEDs can both impair and enhance cognitive functioning (Kwan

& Brodie, 2001). All these issues may be of particular relevance in long-term evaluations

of children and may aid in explaining why none of these variables had a statistically

significant contribution to the models.

Newer types of seizure classification and conceptualization are being proposed

(Berg et al., 2010). These classifications may potentially prove to be differentially

associated with cognitive outcome in epilepsy. Literature suggests that an underlying

cause leading to seizures – be it hereditary, structural, metabolic, or unknown (Berg et al.,

2010) – may affect the cognitive development of the child even before the epilepsy

surfaces (Schouten, Oostrom, Jennekens-Schinkel, & Peters, 2001), and may continue to

influence cognitive development for a prolonged period of time.

The present study proposes that the impact of the seizure condition on an already

affected brain (Hermann et al., 2006) is synergistic, leading to progressive decline in

cognitive function. The impact of selected epilepsy variables on this decline could not

easily be singled out. Evidence of this synergistic effect is also provided by studies on

unilateral brain lesions, showing that the added presence of seizures alters the course of

cognitive development, turning growth into decline (Ballantyne, Spilkin, Hesselink, &

Trauner, 2008). Further support comes from recent studies showing changes in brain

networks of children with focal epilepsy, which extend beyond the epileptic region and

was most prominent in children with lower IQ (Braakman et al., 2012). A study on adults

showed that changes in brain networks could be associated with cognitive decline

(Vaessen et al., 2012). The decisive factor leading to cognitive decline may not be the

presence or absence of seizures or a brain lesion alone, but their co-occurrence and

possible interaction.

Demographic variables, special education.

School problems are frequently seen in children with epilepsy (Fastenau, Shen,

Dunn, & Austin, 2008; Reilly & Neville, 2011). In the present study, special education

was associated with lower VIQ. No interaction with the duration of epilepsy was seen,

suggesting that special education per se is not associated with lowering of IQ over time.


77

In line with research showing that parental education is a predictor of cognitive

functioning (Mitchell, Scheier, & Baker, 1994), a significant effect was seen for parental

educational level. Again, no interaction was seen between parental education and change

over time, suggesting that lower parental education was no risk factor for decline.

Conversely, higher parental education was not “protective” against decline.

Test versions.

Epilepsies emerge at different ages and progress with remissions and relapses

(Camfield, Camfield, Gordon, Smith, & Dooley, 1993; Schmidt & Sillanpää, 2012). A

first step to approaching a wide spectrum of epilepsies longitudinally is the inclusion of

children at various ages – and therefore the inclusion of various test versions, and changes

of test versions over time. However, studies with more than one test version carry the risk

of contaminating results with non-equivalence of test versions and Flynn effects

(Bourgeois, Prensky, Palkes, Talent, & Busch, 1983; Flynn, 2007; Kaufman, 2010;

Neyens & Aldenkamp, 1996). A major advantage of the present study is that it modeled

different test versions explicitly, adjusting for their potential differential contributions.

Similarly, the present model allowed entering children regardless of the number of times

they were tested.

Mechanisms related to the initial drop in IQ and the posterior slowing of decline

The mechanisms leading to an initial drop – differential for the two IQ scales and more

pronounced in the younger child – and posterior stabilization of cognitive functions are

not completely understood. The large individual variability among individuals suggests

different mechanisms between individuals and between aetiological groups. An approach

to understanding the mechanisms may be undertaken from the perspective of the

interaction between brain, epilepsy and cognitive function; and psychometrics.

Before the onset of epilepsy, learning problems may already be seen

(Hermann, Jones, Jackson, & Seidenberg, 2012; Schouten, Oostrom, Jennekens-Schinkel,

& Peters, 2001) pointing towards ongoing latent changes in the brain (Hermann et

al., 2010). The period of epileptogenesis culminates in the disruption of the balance

between excitation and inhibition of the brain network (Jensen, 2011), and in the

emergence of seizures proper. Generalized, non-specific cognitive problems become

evident, affecting mainly attention, executive functions and visual-motor speed (Bhise,

Burack, & Mandelbaum, 2009; Fastenau et al., 2009; Hermann et al., 2006;

Hermann et al., 2012). These difficulties may give rise to the lowered PIQ and to the


78

concomitant VIQ > PIQ gap described in the present paper. Soon after the seizures

become apparent, abnormalities in brain organization are seen. Of particular interest in

relation to the low PIQ, are the white matter abnormalities and a disturbed pattern of

white matter growth observed in several studies (B. P. Hermann et al., 2010; Hutchinson

et al., 2010) which may hamper speed of information processing.

The more preserved VIQ at first testing may be giving a closer estimate of the

child’s original cognitive level. PIQ, with its lower initial score and more gradual decline,

may be giving a better indication of the vulnerable reaction of the brain already during

this process of epileptogenesis and emergence of the seizure condition. The younger

child, with its more immature brain, has a reduced seizure threshold and is particularly

vulnerable to disruption (Rakhade & Jensen, 2009) and more prone to show impaired

cognitive development (Cormack et al., 2007), a lowered PIQ and a worse trajectory of

VIQ.

With the emergence of seizures, the already ongoing process of abnormal

development exacerbates, leading to a cascade of changes, both in the brain and in

cognition (Jensen, 2011; Rakhade & Jensen, 2009). Cognitive decline becomes more

generalized and affects also the initially spared verbal IQ. Information that was already

acquired and consolidated (“wired”) may be preserved and account for the initial higher

level of VIQ. In order to maintain the original IQ, children must earn higher raw scores

when they grow older (Wechsler, 2005). An adverse impact of the epilepsy on novel

problem solving abilities and on the ability to acquire new information, may account for

the reduced rate of cognitive growth. Reduced cognitive growth is detected by the IQ test

as lower scores at retesting. Decline in verbal IQ shows a steeper downward curve and

becomes more evident over months or years. The present data suggest that this decline in

verbal IQ can possibly be understood as being largely non-specific (associated to the

failure of the brain to develop and to acquire new information in the same pace as before)

rather than specific (associated with the brain areas responsible for language and verbal

abilities). A non-specific effect is suggested by the lack of association between side of

seizure onset (right versus left hemisphere) with VIQ – PIQ difference or with the VIQ –

PIQ pattern (the closing of the VIQ – PIQ gap) over time. Also, the association between

low verbal IQ and the presence of epileptic activity during the night described in the

literature (Overvliet et al., 2011) may be interpreted as largely non-specific.

Over time, the brain itself may activate (inhibitory) mechanisms to deal with the

heightened excitation of the brain, possibly due to brain maturation (Rakhade & Jensen,


79

2009), leading to self-containment of the seizure condition. In addition, anti-epileptic

medications aid in seizure suppression and affect cognitive function (Geerts et al., 2010;

Kwan & Brodie, 2001). Many of the childhood epilepsies ameliorate after several years

(Geerts et al., 2010). These factors may all contribute to slowing down cognitive decline,

giving way to renewed development, although often at a lower level than the original

level. The role of reorganization of brain networks, and its implications for cognitive

development remains unclear. Alterations in brain networks in children with frontal lobe

epilepsy were seen more clearly in those with low intellectual ability but were not

associated with duration of epilepsy (Braakman et al., 2012). The timing of the initial

epileptic seizure, the effects of AEDs and changes in brain development may depend on

aetiology and may be positive in some children and negative in others, explaining the

difference between those who continue to decline and those who resume development.

Further research is needed to determine the factors. In some children with low IQs

already at baseline, reaching the floor of the test may occur; thereafter, decline can no

longer be quantified by the test.

Limitations and utility of the study

An important consideration is that this study used a clinical sample from a tertiary

epilepsy setting to study the cognitive course over time. The children had been referred

and repeatedly assessed because concerns about the neuropsychological functioning had

risen. The sample consisted of children who were more likely to have refractory epilepsy,

epilepsy with an unstable course (Geerts et al., 2010; Schmidt & Sillanpää, 2012),

epilepsy that changed into a more atypical and severe forms (Fejerman, Caraballo, &

Tenembaum, 2000), and relatively higher rates of children with moderate and high

epilepsy severity. Therefore, the results cannot be generalized to “uncomplicated”

epilepsy without developmental concerns. However, it should be borne in mind that

cognitive problems are not restricted to children with more severe epilepsies or epileptic

encephalopathies.

Just as uncomplicated cases may not be referred for testing, it would be expected

that children tested more often (one versus two, two versus three times) may be “worse”

cases. Preliminary analyses showed that – up to second testing – these differences were

not statistically significant.

Given the differential impact of duration of epilepsy on the VIQ and PIQ, (cross-

sectional) studies should always include data on duration of epilepsy.


80

Clinical implications

Vulnerability at group level implies that clinically significant decline is likely to

be found for a proportion of the children with epilepsy. An earlier study on reliable

cognitive change suggested that this proportion is indeed elevated (van Iterson, Augustijn,

de Jong, & van der Leij, 2013).

Lowering of IQ is associated with the failure of school children to progress in

school. Repeating grades, being set down to a lower type of education than they enrolled

at the start of secondary school, or both, were seen in a substantial 70% of a sample of 32

youngsters referred to a tertiary centre. A correlation of .41 was found between failure to

maintain school level and reliable cognitive decline (van Iterson, 2010). Even after IQs

cease to drop, the result is a significantly reduced school career – with its implications for

emotional adjustment and future perspectives.

Clinicians, parents and educators should stay alert to signs suggesting that a child

is deviating from its developmental curve, particularly early in the course of the epilepsy,

but also beyond the first years. This deviation may be observed as failure to grow in the

pace expected for the age, or worse, as loss of acquired cognitive abilities. The risk of

decline should also be considered in epilepsies deemed of lower severity.

The study advocates for the inclusion of cognitive outcome in measures of

severity. That is, a child with epilepsy who shows concomitant slowing or stagnation of

development or loss of cognitive functions, should be reclassified adding a constituent on

cognitive function in the descriptive diagnosis, regardless of whether the initial epilepsy

diagnosis per se is of low or moderate severity and whether diagnosis has changed after

re-evaluation. This would be in agreement with the revised guidelines of the ILAE stating

that: “encephalopathic effects [i.e. severe cognitive impairment] of epilepsy may occur in

association with any form of epilepsy” (Berg et al., 2010, p.682). This addition on

cognitive course should have implications for the educational needs of the child.

In like manner, the study advocates an early start of long-term remediation

interventions for all children who develop epilepsy. This educational remediation should

be an individually-tailored process, if necessary lasting beyond seizure remission. Also,

the results of the present study urge researchers to intensify the search for underlying

aetiologies and optimization of medical treatment.

CHAPTER 6

Paediatric Epilepsy and Comorbid Reading Disorders, Math Disorders or

Autism Spectrum Disorders: Impact of Epilepsy on Cognitive Patterns

Loretta van Iterson

Peter F. de Jong

Bonne J.H. Zijlstra


82

Abstract

Introduction: In paediatric epilepsy, comorbidities are reported to be frequent. The

present study focussed at the cognitive patterns of children with isolated epilepsy,

children with isolated neurodevelopmental disorders (reading disorders, math disorders,

autism spectrum disorders) and children with epilepsy and these neurodevelopmental

disorders as comorbidities.

Methods: based on two samples of referred children, one with epilepsy, reading disorders,

math disorders or ASD occurring in isolation (n = 117), and one with reading disorders,

math disorders and ASD occurring comorbid with epilepsy (n = 171), cognitive patterns

were compared. The patterns displayed by verbal and nonverbal abilities from the WISC

series were studied with repeated measures ANOVA. Thereafter, an exploratory 2*3*2

factorial analysis was done to study the independent contribution of type of comorbidity

and of presence or absence of epilepsy on the VIQ – PIQ pattern.

Results: In isolated epilepsy, a VIQ > PIQ pattern was found, not seen in the other

disorders. When epilepsy and another disorder co-occurred, patterns were altered. They

resembled partly the pattern seen in isolated epilepsy and partly the pattern seen in the

isolated neurodevelopmental disorder. In comorbid reading disorders, the VIQ > PIQ was

mitigated; in comorbid math disorders, it was exacerbated. In comorbid ASD, no clear

pattern emerged. In the presence of epilepsy, patterns characteristic of isolated disorders

appear systematically shifted towards relatively lowered performance abilities or

relatively spared verbal abilities. The similar “impact” exerted by the epilepsy on the

patterns of the various conditions suggested shared mechanisms.

van Iterson, L., De Jong, P. F., & Zijlstra, B. J. (2015). Pediatric epilepsy and comorbid reading disorders, math disorders, or autism spectrum disorders: Impact of epilepsy on cognitive patterns. Epilepsy and Behavior, 44, 159-168.

PATTERNS IN COMORBIDITIES

83

Introduction

Seizure conditions in children are heterogeneous disorders in terms of age at onset,

severity, type of seizures, response to medication, duration and cognitive outcomes (Berg

et al., 2008; Schmidt & Sillanpää, 2012). They are often accompanied by general

cognitive problems as a somewhat lowered IQ (Berg et al., 2008; Elger, Helmstaedter, &

Kurthen, 2004; Ellenberg, Hirtz, & Nelson, 1986; Hermann et al., 2008; Nolan et al.,

2003). Besides this general impact on cognition, studies have also suggested differential

effects on cognitive patterns. A number of studies on the Wechsler Intelligence Scales for

Children (WISC series) in mixed samples of children with epilepsy referred for

neuropsychological evaluation suggest that verbal abilities (Verbal IQ, VIQ or the factor

Verbal Comprehension Index, VCI) are relatively spared, while the performance abilities

(Performance IQ, PIQ or the factor Perceptual Organization, POI) are lowered. This

differential “impact” of epilepsy on the verbal and performance scales, suggesting a VIQ

> PIQ pattern seems independent of epilepsy variables such as side of seizure onset,

seizure type, number of anti-epileptic drugs (AEDs), or presence of MRI-abnormalities

(van Iterson & Augustijn, 2006; van Iterson, Zijlstra, Augustijn, van der Leij, & de Jong,

2014). In addition, while level of IQ was lower in children in special education as well as

in children with parents with lower education, the pattern displayed by VIQ and PIQ was

not related to type of education or to level of parental education (van Iterson et al., 2014).

Neuropsychological studies on epilepsy generally include data on verbal and performance

abilities as descriptives of the participants, even when VIQ – PIQ patterns are not the

focus of the study. Based on this information, the VIQ > PIQ pattern (or, similarly, a VCI

> POI pattern) is also observed in children with epilepsy in association with mixed

samples, frontal lobe epilepsies, Panayiotopoulos syndrome, benign epilepsy with centro-

temporal spikes (BECTS) and daytime seizures, the use of polytherapy, and interictal

discharges (Lopes, Simoes, & Leal, 2014; O'Leary, Burns, & Borden, 2006; Overvliet et

al., 2011; Smith, Elliott, & Lach, 2002; Tedrus, Fonseca, Melo, & Ximenes, 2009).

However, in other studies, the opposite pattern is observed. Specifically, VIQ < PIQ

patterns have been reported in mixed samples with learning problems and in association

with BECTS and night time seizures and older age at testing (Aldenkamp, Alpherts, De

Bruine-Seeder, & Dekker, 1990; Miranda & Smith, 2001; Northcott et al., 2007;

Overvliet et al., 2011; Vago, Bulgheroni, Franceschetti, Usilla, & Riva, 2008; Verrotti et


84

al., 2011). Also, some studies have presented data suggestive of similar VIQ and PIQ,

such as studies on lateralized seizures, studies on mixed samples, as well as studies on

samples with epilepsy which were either referred or not referred for psychological

evaluation (Braakman et al., 2012; Jones, Siddarth, Gurbani, Shields, & Caplan, 2010;

Northcott et al., 2007; Tedrus et al., 2009; Vago et al., 2008; van Mil et al., 2009;

Westerveld et al., 2000). Overall, although there is evidence of VIQ > PIQ patterns in

epilepsy, results across studies vary, even within a single epilepsy syndrome (as in

BECTS). These inconsistencies in the literature may be associated with differences across

samples in terms of duration of epilepsy: the VIQ > PIQ pattern is mostly seen in the

early stages of the epilepsy (van Iterson et al., 2014). These differences, however, could

also be related to differences associated with comorbidities in epilepsy.

The plea to study comorbidities in epilepsy is sounding increasingly louder

(Asato, Caplan, & Hermann, 2014; Helmstaedter et al., 2014). Studies have highlighted

the relevance of comorbidities in epilepsy indicating their high frequency of occurrence

(Berg, Caplan, & Hesdorffer, 2011; Fastenau, Shen, Dunn, & Austin, 2008; Russ, Larson,

& Halfon, 2012). In particular, learning, psychiatric, social or behavioural comorbidities

have been frequently reported in children with seizures (Austin & Fastenau, 2010;

Brooks-Kayal et al., 2013; Lin, Mula, & Hermann, 2012; Russ et al., 2012). Learning

problems are common (Austin, Huberty, Huster, & Dunn, 1999; Reilly & Neville, 2011;

Russ et al., 2012) and in an epidemiological study, Russ et al. (2012) indicated that the

adjusted relative risk ratio for various kinds of school problems in epilepsy was 6.7. The

rate of children with epilepsy with reading scores below the seventh percentile has been

reported to be 20.1%; specific reading problems (i.e., based on IQ- achievement

discrepancy) comorbid with epilepsy have been reported in 12.8% of children (Fastenau

et al., 2008). For math problems, the percentage of children with epilepsy scoring below

the seventh percentile was found to be 26.8%; and 20.1% had specific math problems

based on the IQ-achievement discrepancy (Fastenau et al., 2008). Autism spectrum

disorders (ASD) are also a major comorbidity in epilepsy. Russ et al. (2012) reported a

relative risk ratio of 15.5. Rates of co-occurrence of epilepsy and autism tend to vary

from 15% (Russ et al., 2012) to 30% (Tuchman, Alessandri, & Cuccaro, 2010). ASD in

epilepsy is most often seen in the presence of intellectual disabilities; it remains unsettled

whether rates of ASD are elevated in children with epilepsy with average IQs (Berg &

Plioplys, 2012). Importantly, although some comorbidities have been reported to occur

mostly in association with specific epileptic syndromes (Besag, 2009; Clarke et al., 2007),


85

overall, comorbidities have been found to occur across epilepsy syndromes (Berg et al.,

2011; Fastenau et al., 2008; Lin et al., 2012).

Children with epilepsy present with neuropsychological disorders of all kinds

(Braakman et al., 2012; Hoie, Mykletun, Waaler, Skeidsvoll, & Sommerfelt, 2006;

Northcott et al., 2007). These disorders, however, need not lead to the diagnosis of

comorbidities. The disorders may be considered the neuropsychological counterpart of

the epileptic condition reflecting the interference of the seizure condition with

performance on cognitive tasks, not necessarily clustering into a specific second

diagnosis. Such children will be referred to in the present paper as children with

“isolated” epilepsy. Some authors suggest that the focus on the medical condition

(epilepsy) and its treatment may be leading to under diagnosis and underreporting of the

comorbidity (Helmstaedter et al., 2014; Matsuo, Maeda, Sasaki, Ishii, & Hamasaki,

2010). Available studies have suggested that the combined presence of epilepsy and

learning or behavioural disorders are associated with overall lowered IQ (Hermann et al.,

2008). Not much is known as to whether the neurocognitive pattern (like the pattern

displayed by the verbal and performance abilities) seen in children with epilepsy and a

second diagnosis (a comorbidity) resembles the pattern seen in the neurodevelopmental

diagnosis when it occurs as a single diagnosis without epilepsy, that is, when it occurs as

an “isolated” condition.

Henceforth, the term “isolated” will also be used to denote children with a single

diagnosis of a developmental disorder (reading, math, ASD, or epilepsy), in contrast to

the child with a comorbidity. Similar to epilepsy, children with other developmental

disorders may also have other neuropsychological weaknesses which do not qualify for a

second diagnosis. Both isolated epilepsy as well as other neurodevelopmental conditions

occurring in isolation may be characterized by patterns of cognitive strengths and

weaknesses. As said, although the results of the literature remain inconclusive, for mixed

samples of children with epilepsy referred for neuropsychological evaluation, a VIQ >

PIQ pattern of has been found. For language based neurodevelopmental disorders, like

reading and spelling disorders, patterns of relative spared performance abilities and

relatively depressed verbal abilities have been found. Pelletier, Ahmand, and Rourke

(2001) reported that 61% to 78% of their samples with reading disabilities showed a VIQ

< PIQ discrepancy of at least 10 points. For children with math problems, large

discrepancies were seen favouring either the verbal or performance scale (Desoete, 2008),

but sometimes predominantly the verbal scale (Pelletier et al., 2001). In ASD, high rates


86

of children (41% – 50%) have been reported to have VIQ – PIQ discrepancies of 12 or

more IQ points in either direction (Black, Wallace, Sokoloff, & Kenworthy, 2009;

Charman, Pickles, Chandler, Loucas, & Baird, 2011). Relatively high scores on the

performance scale and strengths on specific performance subtests have been found in

mixed ASD samples (Charman et al., 2011; de Bruin, Verheij, & Ferdinand, 2006;

Scheirs & Timmers, 2009), and relatively high scores on the verbal scale have been

observed particularly in Asperger syndrome (Cederlund, 2004; de Bruin et al., 2006).

Thus, in ASD both verbal strengths and performance strengths can be seen, possibly with

a predominance for a VIQ < PIQ pattern.

It has been suggested that the manifestations of neurodevelopmental disorders in

epilepsy (comorbidities) may have both commonalities as well as differences to their

manifestation as isolated conditions (Lin et al., 2012). As in isolated reading disorders,

reading problems comorbid with epilepsy have been associated with lower verbal abilities

and difficulties with verbal memory and learning (Dunn et al., 2010; Vermeulen,

Kortstee, Alpherts, & Aldenkamp, 1994). The epilepsy syndrome most consistently

associated with reading disorders is BECTS (Clarke et al., 2007). Studies on BECTS have

provided some evidence for lowered verbal abilities, but these results have been reported

as being associated with older age and the presence of night time seizures (Northcott et

al., 2007; Overvliet et al., 2011; Tedrus et al., 2009; Vago et al., 2008; Verrotti et al.,

2011). For math disorders in epilepsy, no specific patterns have been described. Problems

with processing speed, younger age of epilepsy onset, symptomatic epilepsies,

generalized seizures and frequent interictal discharges have been identified as risk factors

for math disorders (Dunn et al., 2010; Fastenau et al., 2008; Nicolai et al., 2012; Rathouz

et al., 2014). Given that both PIQ weaknesses (van Iterson et al., 2014) and math

problems (Masur et al., 2013; Rathouz et al., 2014) have been reported early in the course

of the epilepsy, a VIQ > PIQ pattern would be more likely to be seen in math problems in

epilepsy than a VIQ < PIQ pattern. For ASD and epilepsy, associations between language

disorders have been described (Tuchman et al., 2010), but literature on patterns of verbal

and nonverbal abilities in epilepsy and ASD is still scarce. Some features of a disorder

may be masked and others may be emphasized in the light of epilepsy (Lin et al., 2012),

and more work has to be done to understand cognitive patterns seen in children with

epilepsy with or without a comorbid condition.

One aim of the current study was to compare cognitive patterns of children across

conditions, both isolated conditions (that is, without an additional comorbid diagnosis) as


87

well as conditions comorbid with epilepsy. Two main research questions were addressed.

The first research question focussed on the pattern of verbal and nonverbal abilities in

children with isolated epilepsy contrasted (a) to control children, and (b) to children with

other isolated neurodevelopmental disorders, in particular reading disorders, math

disorders and autism spectrum disorders. The first hypothesis was that children with

isolated epilepsy would show a VIQ > PIQ (or VCI > POI) pattern and that this pattern

would be different from control children or other isolated developmental disorders

(reading disorders, math disorders or ASD).

Table 6.1. Characteristics of the samples.Samples sizes, number of boys and age.

Sample 1: WISC-RNL Male Age

N N (%) Mean (SD) Isolated Epilepsy 39 22 (56.4) 12.3 (1.9) Isolated Reading Disorder 29 19 (65.5) 12.6 (0.8) Isolated math Disorder 27 18 (66.7) 12.9 (1.2) Isolated ASD 24 23 (95.8) 12.2 (1.3)

Sample 2: WISC-IIINL Male Age N N (%) Mean (SD)

Isolated Epilepsy 100 48 (48.0) 10.0 (2.6) Epilepsy + Reading Disorder 31 20 (64.5) 9.6 (2.7) Epilepsy + Math Disorder 17 5 (29.4) 9.2 (2.0) Epilepsy + ASD 21 18 (85.7) 9.4 (3.2) Control 81 40 (49.4) 9.4 (1.7)

Note. ASD = autism spectrum disorders. Sample 1 consists of children with epilepsy, reading, math or autism spectrum disorders “in isolation”. Sample 2 consists of non-referred control children, children with epilepsy in isolation, and children with reading, math or autism spectrum disorders comorbid with epilepsy.

The second research question addressed VIQ – PIQ discrepancies for children with

isolated epilepsy versus epilepsy with comorbid disorders. The aimed at studying (a) to

what extent isolated epilepsy and epilepsy with comorbid conditions differ in VIQ – PIQ,


88

and (b) whether VIQ – PIQ patterns in epilepsy depend on the type of comorbid disorder.

Do children with epilepsy show a different cognitive pattern in the light of comorbidities

like reading disorders, math disorders or autism spectrum disorders? Do developmental

disorders present with different patterns when accompanied by epilepsy? The second

hypothesis was that in the light of comorbidities, cognitive patterns will appear altered. If

this is the case, it will provide better understanding of the inconsistent results reported on

the literature. That is, if cognitive patterns in isolated epilepsy are different from patterns

seen in epilepsy with comorbidities, the variation in findings on VIQ – PIQ patterns could

be due to variation across samples reported in the literature in the type and proportion of

comorbid disorders. If patterns in comorbidities appear altered, the finding will also have

implications for the clinical diagnosis of the comorbidity and for its remediation. The

present study was based on two samples, one with isolated conditions and one with

comorbid conditions.

Methods

Participants

Except for the control children, all participating children had been referred for

special services including psychological assessment because of developmental concerns.

The children with epilepsy came from a tertiary centre for epilepsy and from a school

which provided special services to children with epilepsy associated to the centre and had

heterogeneous epileptic conditions. The children with specific learning disorders and

ASD came from schools providing special educational services for learning disorders and

for children with psychiatric and behavioural disorders, respectively. The Wechsler IQ

data for the current study were gathered from the files of the schools and the epilepsy

centre; over time, two different test versions of the Wechsler were used. For each version,

there were not sufficient numbers of children to fill each disorder (reading, math, ASD)

by comorbidity (with or without comorbid epilepsy) condition. Therefore, for this study

two separate samples were used. Sample 1 was tested with the Dutch version of the WISC-

R (to be called WISC-RNL), and consisted of four groups of children matched for age: 39

with isolated epilepsy, 29 with a reading disorder, 25 with a math disorder, and 24 with

an autism spectrum disorder. A portion of the first sample has been described earlier (van

Iterson & Kaufman, 2009).


89

Sample 2, tested with the Dutch WISC-III (WISC-IIINL), included 171 children with

epilepsy and 81 non-referred control children. The control children came from regular

schools and were included only if no disabilities were suspected by parents or teachers

and if their FS-IQ was between 76 and 130. The sample of 171 children with epilepsy

consisted of four groups of children: 100 with epilepsy without a comorbid disorder, 31

with a comorbid reading disorder, 19 with a comorbid math disorder, and 21 with

comorbid ASD. The majority of children were not included in samples reported upon in

earlier publications. However, in order to maintain adequate sample sizes of the children

with comorbidities, 9 children (5.3% of the present sample) were included which

overlapped with an earlier study (van Iterson et al., 2014).

Children were included only if they had taken the complete Wechsler Intelligence

Scales for Children and had a FS-IQ above 75. It should be noted that while FS-IQ was a

criterion for eligibility for special services for children with specific learning disorders

and for children with ASD, patterns displayed by the scales (e.g. VIQ – PIQ discrepancy)

was not.

Children with epilepsy had a confirmed diagnosis of epilepsy by a neurologist or

child epileptologist. Information on epilepsy was obtained from medical reports and

related to seizure type (focal or generalized seizures), side of seizure onset, localisation,

presence of abnormalities on neuroimaging (MRI+), epilepsy syndrome, age at onset of

epilepsy (AOE) and number of AEDs used. Epilepsy syndrome severity was rated on a

10-point scale (Dunn, Buelow, Austin, Shinnar, & Perkins, 2004), where 10 was the most

severe. Duration of epilepsy was calculated as the difference between AOE and age at

testing.

For inclusion in a sample with learning disorders, an expert in special education

verified the presence of significant and persistent achievement problems on the domains

of reading/spelling or math; or a qualified psychologist provided a diagnosis of specific

reading or math disorder. Children with both reading and math disorders were excluded.

C

OG

NIT

IVE

PATT

ERN

S IN

PA

EDIA

TRIC

EPI

LEPS

Y

Tabl

e 6.

2. S

eizu

re c

hara

cter

istic

s of t

he sa

mpl

es w

ith e

pile

psy.

Sa

mpl

e 1

Sa

mpl

e 2

Epile

psy

isol

ated

Ep

ileps

y is

olat

ed

Epile

psy

and

Rea

ding

Dis

. Ep

ileps

y an

d M

ath

diso

rder

s Ep

ileps

y an

d A

SD

M

ean

SD

Mea

n SD

M

ean

SD

Mea

n SD

M

ean

SD

N

39

100

31

19

21

Test

ver

siso

n W

ISC

-RN

L W

ISC

-IIIN

L W

ISC

-IIIN

L W

ISC

-IIIN

L W

ISC

-IIIN

L A

ge a

t ons

et o

f epi

leps

y 6.

7 4.

6 6.

0 3.

0 6.

2 3.

2 5.

3 2.

4 5.

7 3.

7 D

urat

ion

epile

psy

to te

st

5.7

3.7

4.1

3.2

3.4

2.4

3.9

3.0

3.6

2.1

AED

s trie

d 2.

4 1.

5 2.

2 1.

3 2.

0 1.

8 1.

9 1.

3 2.

3 1.

9 Ep

ileps

y sy

ndro

me

seve

rity

5.1

1.4

5.4

1.7

5.2

1.7

5.6

1.7

5.1

0.8

n

%

n %

n

%

n %

n

%

Seiz

ure

type

Gen

eral

ized

seiz

ures

6

15.4

24

24

.0

10

32.3

6

31.6

4

19.0

A

bsen

ces /

Aty

pica

l abs

ence

s 2

/ 1

5.1

/ 2.6

11

/ 3

11.0

/ 3.

0 3

/ 0

9.7

/ 0.0

3

/ 0

15.8

/ 0.

0 0

/ 0

0.0

/ 0.0

To

nic

clon

ic se

izur

es /

Myo

clon

ic

seiz

ures

0

/ 0

0.0

/ 0.0

2

/ 1

2.0

/ 1.0

1

/ 2

3.2

/ 6.5

1

/ 0

5.3

/ 0.0

2

/ 0

9.5

/ 0.0

Se

vera

l gen

eral

ized

seiz

ure

type

s 1

2.6

5 5.

0 3

9.7

1 5.

3 2

9.5

Gen

eral

ized

not

spec

ified

2

5.1

2 2.

0 1

3.2

1 5.

3 0

0.0

Fo

cal (

all f

ocal

) 25

64

.1

56

56.0

15

48

.4

11

57.9

15

71

.4

(Als

o) te

mpo

ral /

(als

o) fr

onta

l 6

/ 11

15.4

/ 28

.2

22 /

13

22.0

/ 13

.0

5 / 1

1 16

.1 /

35.5

1

/ 8

5.3

/ 42.

1 3

/ 5

14.3

/ 23

.8

(Als

o) p

arie

tal /

(als

o) c

entra

l 2

/ 5

5.1

/ 12.

8 10

/ 16

10

.0 /

16.0

1

/ 4

3.2

/ 12.

9 0

/ 1

0.0

/ 5.3

1

/ 1

4.8

/ 4.8

(A

lso)

occ

ipita

l 7

17.9

9

9.0

4 12

.9

1 5.

3 1

4.8

Foca

l: LH

/ R

H

7 / 5

17

.9 /

12.8

22

/ 13

22

.0 /

13.0

4

/ 7

12.9

/ 22

.6

4 / 1

21

.1 /

5.3

6 / 2

28

.6 /

9.5

Bila

tera

l of m

utifo

cal

13

33.3

21

21

.0

4 12

.9

6 31

.6

7 33

.3

U

ncer

tain

/ un

know

n 3

/ 5

7.7

/ 12.

8 19

/ 1

19.1

/ 1.

0 5

/ 1

16.1

/ 3.

2 1

/ 1

5.3

/ 5.3

2

/ 0

9.5

/ 0

MR

I - /

MR

I +

32 /

7 82

.1 /

17.9

83/1

7 83

.0 /

17.0

27/ 4

87

.1 /

12.9

16/ 3

84

.2 /

15.8

18/ 3

85

.7 /

14.3

90

PATT

ERN

S IN

CO

MO

RB

IDIT

IES

Ta

ble

6.2.

Sei

zure

cha

ract

eris

tics o

f the

sam

ples

with

epi

leps

y (c

ontin

ued)

.

Sa

mpl

e 1

Sa

mpl

e 2

Epile

psy

isol

ated

Ep

ileps

y is

olat

ed

Epile

psy

and

Rea

ding

Dis

. Ep

ileps

y an

d M

ath

diso

rder

s Ep

ileps

y an

d A

SD

Epile

psy

synd

rom

es

n %

n

%

n %

n

%

n %

Foca

l idi

opat

hic

BEC

TS /

Rol

andi

c ep

ileps

y (2

)a 3

7.7

5 5.

0 2

6.5

BEO

P / P

anay

otop

olou

s (3)

1

2.6

1 1.

0 1

3.2

1 4.

8

Foca

l sym

ptom

atic

B

y vi

rtue

of: e

tiolo

gy (5

) / lo

caliz

atio

n (7

) 5

/ 13

12.8

/ 33

.3

11 /

36

11.0

/ 26

.0

3 / 1

0 9.

7 / 3

2.3

1 / 5

5.

3 / 2

6.3

2 / 8

9.

5 /3

8.1

Cry

ptog

enic

loca

lisat

ion

rela

ted

(6.5

) 6

15.4

5

5.0

4 21

.1

1 4.

8

Gen

eral

ized

epi

leps

y

Gen

eral

ized

idio

path

ic

CA

E (3

) / JM

E (5

) 1

/ 0

2.6

/ 0.0

12

/ 2

12.0

/ 2.

0 5

/ 1

16.1

/ 3.

2 3

/ 0

15.8

/ 0.

0 0

/ 1

0.0

/ 4.8

O

ther

gen

er id

iopa

thic

epi

not

def

ined

abo

ve (5

) 3

7.7

11

11.0

3

9.7

4 19

.5

C

rypt

ogen

ic a

nd/o

r sym

ptom

atic

W

est s

yndr

. (9.

5) /

nons

peci

fic o

ther

aet

iolo

gy (

8)

1 / 0

1.

0 / 0

.0

0 / 1

0.

0 / 3

.2

1 / 2

5.

3 / 1

0.3

Ep

ileps

y sy

ndro

mes

und

eter

min

ed (f

ocal

/gen

eral

ized

)

Ep

ileps

y w

ith C

SWS

(8)

8 8.

0 2

6.5

LKS

, aty

pic

Rol

andi

c, p

seud

o Le

nnox

(8)

7 7.

0 1

3.2

1 5.

3

Oth

er (

1 - 9

) 3

7.7

3 3.

0

Unk

now

n 4

10.3

8

8.0

2 6.

5 2

10.3

4

19.0

91

C

OG

NIT

IVE

PATT

ERN

S IN

PA

EDIA

TRIC

EPI

LEPS

Y

Not

e. a =

epi

leps

y sy

ndro

me

seve

rity

scor

e in

bra

cket

s acc

ordi

ng to

Dun

n et

al.,

200

4.

Epile

psy

isol

ated

= e

pile

psy

and

cogn

itive

con

cern

s bu

t no

com

orbi

d di

agno

sis

of r

eadi

ng, m

ath,

ASD

. Epi

leps

y +

read

ing

= ep

ileps

y an

d

com

orbi

d re

adin

g di

sord

ers,

epile

psy

+ m

ath

= ep

ileps

y an

d co

mor

bid

mat

h di

sord

ers,

epile

psy

+ A

SD =

epi

leps

y an

d co

mor

bid

ASD

, AED

s

tried

= n

umbe

r of

ant

i-epi

lept

ic d

rugs

trie

d; S

ever

al g

ener

aliz

ed s

eizu

re t

ypes

= e

.g.,

myo

clon

ic s

eizu

res

and

abse

nces

, (A

lso)

tem

pora

l =

tem

pora

l sei

zure

s re

porte

d, p

ossi

bly

in a

dditi

on to

sei

zure

s fr

om a

noth

er lo

calis

atio

n, e

.g.,

tem

pora

l and

occ

ipita

l; L

H /

RH

= le

ft he

mis

pher

e

or ri

ght h

emis

pher

e se

izur

e on

set,

MR

I-: n

o ab

norm

aliti

es o

n ne

uroi

mag

ing

or n

o ne

uroi

mag

ing

avai

labl

e, M

RI+

: abn

orm

aliti

es re

porte

d on

neur

oim

agin

g. B

ECTS

: Ben

ign

Epile

psy

with

Cen

tro-T

empo

ral S

pike

s, B

EOP

= B

enig

n Ep

ileps

y w

ith O

ccip

ital P

arox

ysm

, TC

= to

nic

clon

ic

seiz

ures

, CA

E =

Chi

ldho

od A

bsen

ce E

pile

psy;

JM

E =

Juve

nile

Myo

clon

ic E

pile

psy;

CSW

S =

cont

inuo

us s

pike

and

wav

es d

urin

g sl

ow s

leep

,

LKS

= La

ndau

-Kle

ffne

r syn

drom

e. N

ote

that

rate

s pre

sent

ed u

nder

the

head

ing

of se

izur

e ty

pes a

nd u

nder

epi

leps

y sy

ndro

mes

may

som

etim

es

appe

ar in

cong

ruen

t (e.

g., a

typi

cal a

bsen

ce se

izur

es m

ay b

e cl

assif

ied

as b

elon

ging

to a

foca

l syn

drom

e, fo

cal s

eizu

res m

ay b

e se

en in

CSW

S).

92


93

Inclusion in the sample with specific learning disorders was based on three criteria, for

reading and math alike: (a) severity, defined as achievement scores below the 7th

percentile on reading, spelling or both reading and spelling for a specific reading

disorder, and on mathematics for a specific math disorder; (b) insufficient response to

intervention, i.e., persistence over time in spite of special remediation measures; and (c)

achievement not explained by a low IQ, for which FS-IQ > 75 was required. These

criteria have been maintained over time in The Netherlands in order to qualify for

diagnoses of specific learning disabilities (Pijl & Pijl, 1998; Resing et al., 2002; van Luit,

Bloemert, Ganzinga, & Mönch, 2012).

Children were included in the sample of children with ASD if they had a diagnosis

by a psychiatrist or by a qualified mental health psychologist according to DSM-IV

criteria. Diagnoses on autism, pervasive developmental disorders (PDD-NOS) or

Asperger syndrome as well as broad diagnoses of ASD were pooled into the diagnosis of

ASD. Three (14.3%) children in Sample 2 had been diagnosed with Asperger syndrome,

all others with ASD or PDD-NOS. Children with ASD and another behavioural

comorbidity (e.g., ADHD) were excluded. Table 6.1 shows data on age and sex of the two

samples. Table 6.2 displays data on the epilepsy characteristics for the groups with

epilepsy.

Wechsler Test Versions

Sample 1 was tested with the Dutch adaptation of the WISC-R (to be called WISC-RNL, van

Haasen et al., 1986), in use up to 2005. Sample 2 and the sample of control children, were

tested with the most recent WISC version in the Netherlands, the Dutch WISC-III (WISC-

IIINL, Wechsler, 2005). The test versions share the same two-scale structure, the verbal and

performances scales, and are composed of 5 verbal and 5 performance core subtests with

the same names. Both test versions also share two factor indexes verbal comprehension

index (VCI) and perceptual organization index (POI), which consist of the same subtests

(de Bruyn, Vandersteene, & van Haasen, 1986; Sattler, 1990, 2001; Wechsler, 2005). The

third factor, however, differs between the two samples: freedom from distractibility (FD)

is included in the WISC-RNL, and processing speed (PSI) in the WISC-IIINL. While the focus

in the present paper was on VIQ and PIQ, and VCI and POI will be reported as well, the

role of PSI on the pattern will be considered only briefly. The subtest substitution was

converted to a deviation quotient in order to provide an indication of speed for all

children (Table 6.3).


94

Analyses

The analyses for the two samples were run in parallel. For both samples,

ANOVAs with repeated measures were conducted with IQ scale (VIQ or PIQ, and VCI or

POI, respectively) as within-subject variable and type of disorder (epilepsy, reading

disorder, math disorder and ASD) as between-subject variable. Simple contrasts followed

to compare epilepsy with the other disorders. Thereafter, using ANOVA with planned

contrasts, post hoc analyses were conducted directly on the discrepancy scores (VIQ –

PIQ, VCI – POI) to contrast isolated epilepsy to each of the other disorders. The analyses

were repeated with age and sex as covariates. Similarly, an ANOVA with repeated

measures was done on the factor triad VCI – POI – PSI, to study the effect of processing

speed on isolated epilepsy in the second sample only. In addition, in the second sample

verbal and nonverbal abilities of isolated epilepsy were contrasted with those of the non-

referred control sample.

Results

This section starts with preliminary comparisons on age, sex, epilepsy variables and IQ of

the various groups. Then, results of repeated measures ANOVA are presented in which

VIQ – PIQ and VCI – POI patterns across groups were examined. The means and

standard deviation on the various IQ scales for the various disorders and for the control

groups are presented in Table 6.3. The results of the statistical analyses are reported in

Table 6.4.

The results on the cross-group differences in the VIQ – PIQ difference will be

presented in four sections. (1) First, differences in VIQ – PIQ pattern between the non-

referred control group and isolated epilepsy (Sample 2); (2) differences in VIQ – PIQ

pattern in isolated epilepsy versus isolated reading disorders, isolated math disorders and

isolated ASD (Sample 1); (3) differences in VIQ – PIQ pattern for isolated epilepsy

versus epilepsy with comorbid reading disorders, comorbid math disorders and comorbid

ASD (Sample 2); (4) differences in VCI – POI – PSI pattern in isolated epilepsy (Sample

2) to determine the role of processing speed in epilepsy; and (5) an exploratory overall

analysis of both samples that examines the independent contributions of type of disorder

(reading, math or ASD) and comorbid epilepsy status (absent or present) on the difference

between VIQ and PIQ. Results on VCI and POI are reported in Table 6.4 but will only be

described if the results differ from those for VIQ – PIQ.


95

Table 6.3. Means and SDs on the WISC-RNL and the WISC-IIINL.

Sample 1 Sample 2 "Isolated disorder" Disorder comborbid

with epilepsy Epilepsy Reading Math ASD Epilepsy Reading Math ASD Control

FS-IQ Mean 93.0 93.1 93.5 92.5 90.5 94.0 85.9 94.3 103.0 SD 11.0 11.5 9.1 10.0 11.4 10.1 9.6 13.6 10.7

VIQ Mean 96.4 90.5 93.9 91.6 93.8 94.6 91.6 96.1 102.5 SD 10.9 11.9 8.6 10.1 12.1 10.3 8.5 13.8 11.0

PIQ Mean 90.8 97.8 95.6 94.4 88.4 95.0 82.9 93.6 103.1 SD 13.5 13.1 11.1 14.6 12.4 10.8 11.9 12.9 12.4

VIQ–PIQ Mean 5.6 -7.3 -1.6 -2.8 5.3 -0.5 8.7 2.5 -0.6 SD 13.9 13.8 11.5 16.6 13.5 11.1 11.4 11.7 13.7

VCI Mean 98.0 89.7 92.8 92.5 94.8 95.3 95.6 96.8 102.7 SD 12.0 11.9 10.0 10.3 11.5 11.0 9.7 13.0 11.7

POI Mean 92.9 96.1 92.4 95.6 90.0 95.9 83.1 96.1 103.3 SD 13.3 12.4 12.3 17.1 12.4 10.4 12.4 12.8 12.5

VCI–POI Mean 5.1 -6.4 0.3 -3.0 4.8 -0.6 12.6 0.7 -0.6 SD 15.5 12.7 15.3 19.4 13.6 12.6 11.4 12.1 13.4

SU Mean 89.0 97.1 94.0 90.0 90.4 96.1 91.1 89.0 103.0 SD 11.4 15.8 12.4 12.9 14.5 13.2 10.9 13.3 15.3

PSI Mean 91.1 96.8 89.0 87.9 104.1 SD 14.6 13.2 12.8 13.8 14.8

VCI–PSI Mean 3.7 -1.6 6.6 8.8 -1.4 SD 17.1 14.3 16.5 13.2 18.2

POI–PSI Mean -1.1 -0.9 -5.9 8.1 -0.86 SD 15.7 15.2 16.1 15.6 18.3

Note. VCI = Verbal Comprehension Index, POI = Perceptual Organization Index, PSI = Processing Speed Index (WISC-IIINL only). SU = subtest substitution converted into a deviation quotient.

Preliminary analyses

Comparison of isolated epilepsy between Sample 1 and 2 with t or chi-square

tests showed that the first sample was older at the age of testing (t = 5.64, p < .001, d =

1.16) and the duration of epilepsy was also longer (t = 2.45, p = .015, d = .43). Otherwise,

statistically significant differences were not seen for sex, any epilepsy variable or any of

the Wechsler scales or factor indexes. MANCOVA, adjusting for age and duration of

epilepsy, revealed that the two samples with isolated epilepsy showed highly similar

patterns of verbal and performance abilities (for VIQ – PIQ, F(1, 132) = 0.035, p = .852

and for VCI – POI, F(1,132) = 0.003, p = .958). Similar results were found when the


96

square root of the duration of epilepsy (van Iterson et al., 2014) was entered in the

MANCOVA instead of duration of epilepsy.

Within each sample there were no age differences across disorder groups. For

Sample 2, ANOVA and chi-square tests did not reveal statistically significant differences

across the four epilepsy groups for any of the epilepsy variables. Chi-square tests showed

that boys were overrepresented in ASD in Sample 1 (χ2 (3) = 11.2, p = .011) and Sample

2 (χ2 (3) = 17.1, p = .001). Comparison for each disorder between Sample 1 and 2 showed

similar sex ratios for reading disorders and ASD. However, girls were overrepresented in

comorbid math (Sample 2) relative to isolated math disorders (Sample 1; χ2 (1) = 7.50, p

= .006). ANOVA and ANCOVA revealed that there were no differences in FS-IQ

between Samples 1 and 2 before or after adjusting for differences in age and sex.

Epilepsy in isolation versus non-referred controls

Repeated measures ANOVA (Table 6.4.3. showed a significant main effect for

group (F(1,179) = 62.27, p < .001, η2p= 0.26). The control children outperformed the

children with isolated epilepsy. A main effect was seen for the IQ scales, indicating that

VIQ was higher than PIQ (F(1,179) = 5.38, p = .022, η2p= 0.03). In addition, the

interaction of VIQ and PIQ by group was significant (F(1,179) = 8.64, p = .004, η2p=

0.05; for VCI and POI F(1,179) = 7.03, p = .009, η2p= 0.04). The epilepsy group had

higher verbal than performance abilities while the control sample had a flat pattern of

verbal and performance abilities. Thus, the results indicated that the control children had

higher overall scores on the Wechsler test, and that the children with epilepsy had a VIQ

> PIQ pattern not seen in the controls. There were no age and sex differences between

samples.

PATT

ERN

S IN

CO

MO

RB

IDIT

IES

Tabl

e 6.

4.1.

Res

ults

of re

peat

ed m

easu

res A

NO

VA

. Sam

ple

1.

Re

peat

ed m

easu

res A

NO

VA

Post

hoc

anal

yses

95

% C

I 95

% C

I 95

% C

I

df1

df2

F p

η p2

Con-

trast

SE

p U

L LL

Co

n-tra

st SE

p

UL

LL

Con-

trast

SE

p U

L LL

Sam

ple

1: Is

olat

ed D

isord

ers

Epile

psy

vs R

eadi

ng

Ep

ileps

y vs

Mat

h

Epile

psy

vs A

SD

VIQ

PIQ

W

ithin

subj

ects

VIQ

vs P

IQ

1 11

6 1.

35 .

247

0.01

Be

twee

n su

bjec

ts D

isord

er

3 11

3 0.

15 .

930

0.00

0

.53

2.49

.8

33

-4.4

5.

5 1.

11 2

.47

.654

-3

.8

6.0

-0

.63

2.63

.81

2 -5

.8

4.6

Inte

ract

ion

VIQ

vs P

IQ *

diso

rder

3

113

4.99

.00

3 0.

12

-12.

89 3

.43

<.00

1 -1

9.7

-6.1

-7

.26

3.59

.04

6 -1

4.4

-0.1

-8

.45

3.63

.02

2 -1

5.6

-1.3

V

CI P

OI

With

in su

bjec

ts V

CI v

s PO

I 1

116

0.48

.49

1 0.

00

Betw

een

subj

ects

Diso

rder

3

113

0.58

.63

2 0.

02

-2.5

6 2.

50

.310

-7

.5

2.4

-2.8

4 2.

51

.260

-7

.8

2.1

-1.3

7 2.

65

.605

-6

.6

3.9

Inte

ract

ion

VCI

vs P

OI *

diso

rder

3

113

3.25

.02

5 0.

08

-11.

53 3

.84

.003

-1

9.1

-3.9

-4

.76

4.03

.2

40

-12.

7 3.

2 -8

.12

4.07

.0

48

-16.

2 -0

.1

97

C

OG

NIT

IVE

PATT

ERN

S IN

PA

EDIA

TRIC

EPI

LEPS

Y

Tabl

e 6.

4.2.

Res

ults

of r

epea

ted

mea

sure

s AN

OV

A. S

ampl

e 2.

R

epea

ted

mea

sure

s AN

OV

A

Po

st h

oc a

naly

ses

95%

CI

95%

CI

95%

CI

df1

df2

F p

η p2

Con

-tra

st

SE

p U

L LL

C

on-

trast

SE

p

UL

LL

Con

-tra

st

SE

p U

L LL

Sa

mpl

e 2:

Com

orbi

ditie

s

Is

olat

ed E

pile

psy

vs E

pile

psy

and

Rea

ding

Dis

orde

rs

Is

olat

ed E

pile

psy

vs

Epile

psy

and

Mat

h D

isor

ders

Is

olat

ed E

pile

psy

vs

Epile

psy

and

ASD

V

IQ P

IQ

With

in su

bjec

ts

VIQ

vs P

IQ

1 17

0 11

.32

.001

0.0

6 B

etw

een

subj

ects

D

isor

der

3 16

8 3.

00 .

032

0.05

3

.70

2.08

.0

78

-0.4

7.

8 -3

.82

2.52

.13

2 -8

.8

1.2

3.7

7 2.

43

.123

-1

.0 8

.6

Inte

ract

ion

VIQ

vs P

IQ *

dis

orde

r 3

168

2.57

.05

6 0.

04

-5.7

9 2.

61

.028

-1

1.0

-0.6

3

.34

3.17

.29

3 -2

.9

9.6

-2.8

2 3.

05

.358

-8

.8 3

.2

VC

I PO

I W

ithin

subj

ects

V

CI v

s PO

I 1

170

12.5

6 .0

01 0

.07

Bet

wee

n su

bjec

ts

Dis

orde

r 3

168

2.56

.05

7 0.

04

3.19

2.0

3 .1

18

-0.8

7.

2 -3

.05

2.45

.21

5 -7

.9

1.8

4.0

1 2.

37

.092

-0

.7 8

.7

Inte

ract

ion

VC

I vs P

OI *

dis

orde

r 3

168

4.61

.00

4 0.

08

-5

.46

2.69

.0

44

-10.

8 -0

.1

7

.77

3.27

.01

8 1

.3

14.2

-4.1

0 3.

14

.194

-1

0.3

2.1

98

PATT

ERN

S IN

CO

MO

RB

IDIT

IES

Ta

ble

6.4.

3. R

esul

ts o

f rep

eate

d m

easu

res A

NO

VA

. Sam

ple

2.

R

epea

ted

mea

sure

s AN

OV

A

Po

st h

oc a

naly

ses

95

% C

I

df1

df2

F p

η p2

Con

-tra

st

SE

p U

L LL

Sa

mpl

e 2:

Iso

late

d Ep

ileps

y vs

Con

trols

Isol

ated

Epi

leps

y vs

Con

trols

V

IQ P

IQ

With

in su

bjec

ts

VIQ

vs P

IQ

1 17

9 5.

38

.022

0.0

3 B

etw

een

subj

ects

G

roup

1

179

62.2

7 <.

001

0.26

11

.71

1.48

<.0

01

8.8

14.6

In

tera

ctio

n V

IQ v

s PIQ

* g

roup

1

179

8.64

.0

04 0

.05

-5.9

7 2.

03

.004

-1

0.0

-2.0

V

CI P

OI

With

in su

bjec

ts

VC

I vs P

OI

1 17

9 4,

42

.037

0,0

2 B

etw

een

subj

ects

G

roup

1

179

50,9

0 <.

001

0,22

10

.59

1.48

<.0

01

7.7

13,5

In

tera

ctio

n V

CI v

s PO

I * g

roup

1

179

7,03

.0

09 0

.04

-5

.37

2.02

.0

09

-9.4

-1

.4

Not

e. R

eadi

ng/m

ath

= re

adin

g di

sord

ers/

mat

h di

sord

ers.

Rig

ht s

ide

of th

e ta

ble

pres

ents

the

resu

lts o

f po

st h

oc a

naly

ses:

sim

ple

cont

rast

s fr

om r

epea

ted

mea

sure

s fo

r th

e be

twee

n-su

bjec

t ef

fect

s an

d pl

anne

d co

ntra

sts

from

AN

OV

A (

on th

e di

scre

panc

y) f

or th

e in

tera

ctio

n ef

fect

s. LL

/UL:

low

er a

nd

uppe

r lim

its o

f the

95%

con

fiden

ce in

terv

al.

99


100

Verbal and performance abilities in isolated conditions. Epilepsy, Reading

Disorders, Math Disorders and ASD

Repeated measures ANOVA showed that the main effect of disorder group was

not significant (Sample 1, see Table 6.4.1), indicating that overall IQ scores were

highly similar across groups. Also, the overall difference between VIQ and PIQ was

not significant. But, as expected, there was a significant disorder by IQ-scale (VIQ –

PIQ) interaction, indicating that the discrepancy between VIQ and PIQ differed across

disorder groups. Planned contrasts revealed a significant difference in VIQ – PIQ

pattern between epilepsy and each of the disorders (reading disorder, math disorder

and ASD). In epilepsy, VIQ was higher than PIQ, whereas in the other disorders the

VIQ – PIQ pattern was opposite or flat. These results did not change when the

analyses were redone with age and sex as covariates.

Verbal and performance abilities in epilepsy. Isolated epilepsy and epilepsy

comorbid with Reading Disorders, Math Disorders and ASD

Repeated measures ANOVA revealed a significant main effect of disorder

group suggesting differences in overall IQ across groups (Sample 2, see Table 6.4.2).

Post hoc tests revealed no IQ differences when the comorbid conditions (epilepsy

with reading disorder, math disorder or ASD) were compared to isolated epilepsy.

Children with a comorbid reading disorder, however, outperformed children with a

comorbid math disorder (FS-IQ of 94.0 versus 85.9).

The interaction effect of disorder by IQ scale (VIQ – PIQ) fell short of

statistical significance (p = .056). However, the interaction of disorder by factor index

(VCI – POI) was significant indicating that the VCI – POI difference varied across

groups. Using planned contrasts, isolated epilepsy was compared to each comorbid

condition. A significant difference in VCI – POI pattern was found between the group

with isolated epilepsy and the group with epilepsy with comorbid reading disorders

(similar to the results for VIQ – PIQ). The higher VCI than POI pattern found in

isolated epilepsy was not seen in the group with an additional reading disorder.

Comparison of isolated epilepsy and epilepsy with a comorbid math disorder showed

that the VCI – POI difference was significantly higher in the group with a comorbid

math disorder (but not for the VIQ – PIQ discrepancy). No differences were found in

VCI – POI pattern between isolated epilepsy and epilepsy with comorbid ASD.

Overall, the results of these analyses suggest that in the sample with epilepsy, verbal


101

abilities were higher than performance abilities. However, this pattern is qualified by

a comorbid reading or math disorder. In comorbid reading disorders, the VIQ > PIQ

pattern is not found, while in math disorders, the VIQ > PIQ pattern seems

exacerbated. Comorbid ASD did not affect the pattern of verbal and performance

abilities. The results did not change when the analyses were redone with age and sex

as covariates.

The role of processing speed

Repeated-measures ANOVA with polynomial contrasts was conducted on the

factor triad VCI – POI – PSI in isolated epilepsy. It was hypothesized that, if lowered

performance abilities would be mainly due to lowered speed of processing, the factor

PSI, with its high reliance on speed, should be the lowest in the pattern and a linear

downward pattern should emerge. The analysis revealed a pattern best described as

quadratic (F (1,99) = 6.11, p = .015, ηp2 = 0.58); a linear pattern was also supported (F

(1,99) = 4.74, p = .032, ηp2 = 0.46). That is, the VCI was highest, and POI as well as

PSI were lowered, but PSI was relatively less lowered. Pairwise contrasts indicated

that VCI was higher than both POI (t = 3.53, p = .001, d = 0.43) and PSI (t = 2.18, p =

.032 , d = 0.28). No significant difference was seen for POI – PSI (t = -0.69, p = .492).

The impact of epilepsy and of diagnostic condition on verbal - performance

patterns

Results from Sample 1 and Sample 2 indicate that in isolated epilepsy PIQ is

lower than VIQ. In the other disorders (Sample 1), such pattern is not seen. In reading

disorders, VIQ is relatively lower than PIQ; and in math disorders and ASD, VIQ is

approximately equal to PIQ. In comorbid disorders (Sample 2), however, the VIQ –

PIQ patterns are different from the VIQ > PIQ pattern that characterizes isolated

epilepsy. In comorbid reading disorders, the pattern is relatively flat, and in comorbid

math disorders, an even greater VIQ > PIQ discrepancy emerges. In ASD, no major

changes are seen.

In a final exploratory analysis, we examined whether these VIQ – PIQ

differences across disorders (reading disorders, math disorders and ASD) were

similarly affected by status of epilepsy (presence or absence). For this analysis, the

disorder groups of the two samples were taken together. The control children (Sample

2) and the children with isolated epilepsy (Samples 1 and 2) were excluded. Sample 1


102

and Sample 2 differed in the type of Wechsler test that was administered. In taking

together groups from both samples, it is assumed that different types of Wechsler tests

do not affect the VIQ – PIQ discrepancies, although they may have an impact on mean

IQ differences. This assumption appears warranted, given that VIQ – PIQ

discrepancies were highly similar across Samples 1 and 2. The assumption will be

further considered in the Discussion. A 2 * 3* 2 factorial analysis was done, which

included two IQ scales (VIQ and PIQ) by three disorders (reading disorder, math

disorder, ASD) by two values of status of epilepsy (present or absent). Thereafter, the

results were redone including age and sex as covariates and no major changes were

seen on the main effects.

There was no significant difference in overall IQ between the isolated and

comorbid disorder (F(1,143) = 1.05, p = .308), which indicated there was no overall

effect of epilepsy status on IQ. There was also no significant difference in overall IQ

level across disorders (F(2,143) = 1.76, p = .175). However, the disorder by epilepsy

interaction was significant (F(2,143) = 3.11, p = .047, η2p= 0.04), due to a lower

overall IQ for math disorder comorbid with epilepsy.

Overall, the difference between VIQ and PIQ was not significant (F(1,143) =

.024, p = .877). More interestingly, significant interactions were found with epilepsy

and with type of disorder. The interaction of VIQ – PIQ with epilepsy status (F(1,143)

= 12.35, p = .001, η2p= 0.08) was due to a VIQ < PIQ pattern in the isolated disorders

and a VIQ > PIQ pattern in the comorbid disorders. The interaction between VIQ –

PIQ and disorder (F(2,143) = 4.19, p = .017, η2p= 0.06) indicated differences in VIQ –

PIQ patterns across disorders, irrespective of epilepsy status. Most importantly, the IQ

scales by disorder by epilepsy interaction was not significant (F(2,143) = 0.44, p =

.647), indicating that the VIQ – PIQ discrepancies across the disorders were similar

for the isolated and comorbid disorders, given the VIQ > PIQ pattern seen in epilepsy.

Put differently, the VIQ – PIQ discrepancy in each of the disorders is affected in a

similar way by the comorbid presence of epilepsy. This can also be seen in Figure 6.1,

where the lines are largely parallel, suggesting a systematic shift in the difference

between VIQ and PIQ due to comorbid epilepsy. A similar shift is visible in Figure

6.1 between the control group and the group with isolated epilepsy.


103

Figure 6.1.

Means after adjusting for age and sex with MANCOVA for VIQ – PIQ discrepancies (left) and VCI – POI (right) for the data of the two studies. No (add) diag = no (additional) diagnoses, denotes “isolated” epilepsy and controls. Open squares: children with isolated epilepsy (= no additional diagnosis), reading disorders (= Reading), math disorders (= Math) and ASD. Filled triangles = non-referred control sample (WISC-IIINL). Filled diamonds: children with epilepsy: isolated epilepsy, or epilepsy comorbid with reading disorders, math disorders and ASD. Note that positive values – values in the upper part of the figure – denote lowered performance/perceptual abilities and negative values – in the lower part of the figure – denote lowered verbal abilities, while overall FS-IQs do not differ.

Discussion

The present study further supported the finding that the cognitive pattern of referred

children with isolated epilepsy – that is, epilepsy without an additional diagnosis – is

characterized by relatively spared verbal abilities and relatively depressed

performance abilities (VIQ > PIQ). This pattern is not seen in children with isolated

neurodevelopmental disorders – reading disorders, math disorders, ASD – of similar

overall IQ. In children who had two conditions diagnosed jointly – epilepsy and either

reading disorders, math disorders or autism spectrum disorders – patterns are different.

control control

-10

-5

0

5

10

15

No (add)diag reading math ASD

No (add)diag reading math ASD

VIQ - PIQ VCI - POIΔ

in IQ

poi

nts


104

The VIQ > PIQ pattern appears mitigated in reading disorders and exacerbated in

math disorders.

Alternatively, from the perspective of the learning or behavioural condition,

the results suggest that in the presence of epilepsy, the cognitive patterns of

neurodevelopmental disorders are altered. This change is the direction of relatively

lowered performance abilities or relatively spared verbal abilities. A strength in the

performance abilities seen in isolated reading disorders appears levelled off; a flat

pattern seen in isolated math disorders is changed into a VIQ > PIQ pattern in

comorbid math disorders. Supplementary exploratory analyses further suggested that

the impact of epilepsy on VIQ – PIQ discrepancies is similar across the various

disorders. These results might suggest a common mechanism from the seizure

condition impinging on the comorbid neurodevelopmental disorder.

The study shed new light on the previous equivocal findings on the VIQ – PIQ

pattern in children with epilepsy, suggesting that patterns vary according to the

presence of children with comorbidities. By carefully checking for comorbid

disorders, the present study found higher verbal than performance/perceptual abilities

in isolated epilepsy. However, VIQ > PIQ patterns in isolated epilepsy were altered

when epilepsy appeared with a comorbidity. These results illustrate that VIQ – PIQ

differences across samples of children with epilepsy can differ if comorbidity with

other disorders is not taken into account.

The VIQ > PIQ pattern in isolated epilepsy was different from the pattern

observed in other isolated disorders. Depressed VIQ was seen in reading disorders,

and relatively flat VIQ – PIQ patterns were seen in math disorders and ASD, overall

in accordance with previous studies (de Bruin et al., 2006; Desoete, 2008; Pelletier et

al., 2001; Wechsler, 2005). Earlier studies comparing children with epilepsy with

other referred children have reported similar differences in cognitive patterns between

the referred children with and without epilepsy (Nicolai et al., 2012; van Iterson &

Kaufman, 2009; Vermeulen & Aldenkamp, 1995).

A VIQ > PIQ pattern was also found in an earlier study on paediatric epilepsy

(van Iterson et al., 2014), particularly in early onset epilepsy and in the early years of

the epilepsy, and fading away over time. The lowered scores on PIQ subtests might be

caused by impairments in visual perceptual abilities, perceptual reasoning, or

constructional and motor abilities, which are all involved in the subtests constituting

PIQ (and POI). Also, several of these subtests are timed, which means that quicker


105

speed leads to better scores. Earlier studies have, indeed, pointed toward lowered speed

and executive abilities in children with epilepsy (Gottlieb, Zelko, Kim, & Nordli, 2012),

impaired visual perceptual reasoning, visual attention, sustained attention, motor

abilities and motor speed (Bhise, Burack, & Mandelbaum, 2010; Braakman et al.,

2012). Lowered scores on speed, executive and visual tasks have also been reported in

new-onset epilepsy before the start of medication (Hermann et al., 2006; Oostrom, van

Teeseling, Smeets-Schouten, Peters, & Jennekens-Schinkel, 2005; Rathouz et al., 2014)

and persisting over time (Rathouz et al., 2014). Lowered scores on processing speed

were also seen in the present work. The most depressed scores were seen, however, on

the performance/ perceptual abilities, suggesting that the more complex constructional

abilities tapped by the scales may be most vulnerable to the epileptic condition, where

speed may be one of the constituents leading to low scores.

Compromised performance abilities have also been seen in children born

prematurely (Lee, Yeatman, Luna, & Feldman, 2011), children with traumatic brain

injury (Babikian & Asarnow, 2009), and children with lateralized perinatal brain

damage, regardless of the side of the lesion (Ballantyne, Spilkin, Hesselink, & Trauner,

2008). These results suggest that the performance scale appears particularly vulnerable

to neurological risks, including epilepsy.

An interesting finding of the present study is that comorbid epilepsy in various

disorders appeared associated with a similar “systematic” shift in the difference

between VIQ and PIQ compared to these disorders in isolated form. Though potentially

important, this finding should be considered with some caution. First, it should be

acknowledged that the sample sizes were probably too small to detect small interaction

effects. Second, the finding is based on combination of samples tested with two

different versions of the Wechsler scales and with an age difference of 3.1 years. In

merging the data into a final analysis, it was assumed that differences in test version and

age would not influence the results. Changes in test version can potentially be

associated with changes in verbal – performance patterns, given that Flynn effects may

affect the subtests differentially (Kaufman, 2010). In a previous study on epilepsy,

however, which explicitly modelled for effects of test version (Dutch WPPSI-R, WISC-R

and WISC-III) on VIQ – PIQ patterns, no effect of test version was seen (van Iterson et

al., 2014).

With regard to age differences, earlier studies on children with isolated reading

disorders, math disorders and autistic spectrum disorders with younger children or with


106

wider age ranges were congruent with the present results (de Bruin et al., 2006;

Desoete, 2008; Pelletier et al., 2001; Wechsler, 2005). Also, studies on serial testing

of referred children reported negligible mean differences (Canivez & Watkins, 1998)

in VIQ – PIQ discrepancy over time. A follow-up study on non-referred children

reported moderate to high correlations for the VIQ – PIQ discrepancy across ages, and

a higher stability over time for VIQ < PIQ discrepancies than for VIQ > PIQ

discrepancies (Moffitt & Silva, 1987). Therefore, although results should be treated

with caution, there are also arguments suggesting that differences in age and test

version did not unduly affect the results. Notably, the samples with isolated epilepsy –

regardless of age and test version – showed conspicuous similarities in cognitive

patterns.

In line with current knowledge, both epilepsy as well as the comorbidities are

understood as complex, multidimensional conditions in terms of aetiology and

presentation (Berg et al., 2010; Pennington, 2006; Walsh, Elsabbagh, Bolton, &

Singh, 2011). The epilepsy and the comorbidity may be independent conditions, or

they may be related conditions partly sharing underlying risk factors (Brooks-Kayal et

al., 2013). Several models on the causes of comorbidities have been proposed (Pal,

2011) which may all be valid in particular cases. According to one model, the seizure

condition could be understood as the cause of the comorbid disorder. The epileptic

networks could be interfering with cognitive networks involved in reading (for

example), causing a reading disorder (Pal, 2011). In this model, the VIQ – PIQ pattern

in a particular isolated disorder would be similar to the pattern of this disorder

comorbid with epilepsy – a finding not supported by the current study. A second

model suggests that there may be one or more causes leading to the epilepsy as well as

the comorbidity, which may present alone or in combination. According to the third

model (Pal, 2011), epilepsy and the comorbidity may or may not share a common

cause, but epilepsy might impact on the comorbidity, for example by aggravating it.

The current study was not designed to test these models, and does not permit

conclusions about their validity. However, the present study may contribute to the

understanding of comorbidities in epilepsy, suggesting that in familiarly unrelated

cases, isolated neurodevelopmental disorders show different cognitive patterns from

patterns in isolated epilepsy, and when neurodevelopmental disorders present together

with epilepsy, an altered cognitive pattern is seen relative to the isolated condition.

Cognitive patterns seen in isolated disorders appear systematically shifted towards


107

relatively lowered performance abilities (or relatively spared verbal abilities) when

they co-occur with epilepsy.

A limitation of the study is the inclusion of two samples. Although efforts were

done to select the children with developmental problems according to objective

criteria, new insights and new assessment tools develop over time and sharpen

diagnostic criteria for classification (Pijl & Pijl, 1998; Resing et al., 2002), leading to

differences. Some criteria remained stable over time, as the 7th percentile criterion to

determine a true weakness in learning disorders, even if subtyping of disorders has

progressed. In the diagnosis of ASD, the earlier reliance on subtypes as PDD-NOS

and Asperger subtype is leading to a broad categorization of “ASD”, possibly with the

advent of DSM-V. Despite these unavoidable differences, cognitive patterns found for

the isolated disorders resembled those described in the literature.

Overall, the present study suggests that in isolated epilepsy, the cognitive

pattern is characterized by VIQ > PIQ. In other developmental disorders, such a

pattern was not seen. When these disorders appear as comorbidities in epilepsy, the

patterns are altered, partly resembling the isolated condition, and partly differing from

the isolated condition. In clinical evaluations of children with epilepsy, and

independent of epilepsy syndrome, the possibility of comorbidities should be

considered. The most relevant clinical implication of the present study is that the

cognitive pattern seen in the disorder comorbid with epilepsy is likely to differ from

the pattern seen in the isolated condition. One possibility is that the difficulties

encountered by the child with epilepsy may be associated with specific “subtypes”, of

the disorder. It may be speculated that children with epilepsy and reading disorders,

problems with rapid naming may be more prominent than phonological disorders.

Regardless of whether the starting point is epilepsy or another developmental

disorder, if the disorder is accompanied by epilepsy, the clinician should take into

consideration that the cognitive pattern may be unlike the pattern seen in the isolated

condition. Remediation measures should therefore be tailored to fit the individual

profile of the child with epilepsy and a comorbid diagnosis.

Discussion

DISCUSSION

111

Discussion

The overarching topic of the present work was the pattern of verbal and nonverbal

development found in referred children with epilepsy. It dealt with intra-individual

differences – that is, high levels of functioning on one cognitive domain or one measure

and low levels on the other – between IQ scales, between subtests and between serial

measurements. While the main sample of interest was the sample with epilepsy, for some

studies, clinical comparison children with other developmental disorders or typically

developing control children were included as well. The research questions were all

addressed with Dutch versions of the same diagnostic instrument: the Wechsler

Intelligence Scales for Children.

Overview of Major Findings

This section will start with an overview of the findings of the diverse chapters and some

words on consistenncies and inconsistencies between the results. Thereafter, in the next

section, the patterns of cognitive development and cognitive change that emerge from

these findings will be detailed.

Subtest scatter.

The Introduction gave a brief outline on epilepsy in children in relation to

cognition and listed the research questions. Thereafter, in Chapters 2 and 3, intra-

individual subtest variability or “subtest scatter” was discussed. Subtest scatter was

compared across samples with neurodevelopmental disabilities on the verbal,

performance and full scales of the Wechsler Intelligence test for Children to study

whether elevated scatter is a sign of pathology, and to study whether it is elevated in

epilepsy. The results indicated that while intra-individual subtest variability was elevated

in clinical samples, elevated scatter was not an overall sign of pathology. Rather, large

intra-individual subtest variability was seen in some developmental disorders, but not in

others. It was seen mainly in children with psychiatric disorders on performance IQ and

full-scale IQ. Within psychiatric disorders, large intra-individual subtest variability was

seen particularly in children with autism spectrum disorders, where it was seen on all

scales. Children with learning disabilities were less likely to show increased variability.

Children with epilepsy as a whole did not show increased subtest scatter. Subsets of

children with epilepsy, however, specifically children with left hemisphere seizures, were


112

prone to show increased scatter on the verbal scale. This was especially so when

abnormalities were present on neuroimaging. Children with right hemisphere onset

seizures and MRI abnormalities, on the other hand, showed decreased variability on the

verbal and full scale.

Reliable Cognitive Change.

Chapter 4 aimed at establishing whether rates of individuals showing reliable

cognitive change in IQ scores over time were elevated in epilepsy. Children with epilepsy

were tested two times with the same test version with a mean interval of 2.3 years.

Cognitive change was tested against predetermined 90% cut-off scores (5% of scores

beyond this interval reflecting reliable cognitive gain, 5% reflecting cognitive loss) based

on data coming from referred children with neurodevelopmental disorders, but without

epilepsy. It was found that children with epilepsy were likely to show reliable cognitive

change in higher rates than expected for referred children. This change was seen mainly

as decline; the rates of cognitive gains did not differ from those expected. Decline was

seen on the verbal scale in 26% of the children and on the full scale in 16.4 %. This is a

fivefold rate of cognitive loss on the verbal scale, and a threefold rate of cognitive loss on

the full scale for children with epilepsy relative to referred children without epilepsy. On

the performance scale, rates were not found to be elevated (5.5%).

Relation of pattern of cognitive change to epilepsy variables.

In Chapter 5, a longitudinal study was conducted to describe the pattern of

cognitive change over time as a function of duration of the seizure condition and other

epilepsy variables associated with severity of the condition. A differential impact on the

verbal compared to the performance scale was found. The performance scale showed

already relatively lowered scores at initial testing; thereafter a further decline was seen.

The verbal scale was initially “spared” and showed a steep decline followed by more

gradual decline that continued for a prolonged period of time. On both scales, the largest

decline is seen in the first 40 to 50 months after the onset of epilepsy; thereafter the

decline was less pronounced. Early age at onset of epilepsy and longer duration were

associated with more decline. Other epilepsy-related variables associated with severity of

epilepsy, however, failed to show a relation with cognitive decline over time. Inclusion in

special education was associated with lower IQ scores, but not with different patterns of

decline. Likewise, higher parental education was associated with higher IQ scores, but

was not “protective” against decline. While the overall pattern was characteristic for

cognitive decline over time at group level, large variability was seen among children.

DISCUSSION

113

Comorbidities and the VIQ – PIQ discrepancy

In Chapter 6, the verbal and performance scales and the VIQ – PIQ discrepancy

were addressed once again to study the impact of comorbidities in epilepsy on cognitive

patterns. Besides children with isolated epilepsy, the study also included typically

developing control children, children with other developmental disorders and children

with comorbidities in epilepsy. Patterns in isolated epilepsy were studied in relation to

patterns seen in isolated learning disorders and ASD, and in relation to epilepsy comorbid

with learning disorders or ASD. “Isolated” disorders were those which were not

accompanied with a second diagnosis; the child may have had other cognitive or

behavioural problems but these did not qualify for a diagnosis. Children with epilepsy in

isolation showed a pattern which differed from the pattern seen in non-referred control

children, and in children with isolated reading disorders, math disorders and autism

spectrum disorders. A VIQ > PIQ pattern (with a mean discrepancy of 5 to 6 IQ points)

was seen in epilepsy but not in children with other disorders – and not in control children.

In typically developing children, the pattern between verbal and performance abilities was

flat, and in other developmental disorders a tendency for a VIQ < PIQ pattern was seen

(most clearly in reading disorders in which VIQ was 7.3 points lower than PIQ).

However, the pattern was different when epilepsy was accompanied by a comorbidity. In

epilepsy and comorbid math disorders, the advantage of the verbal scale seen in epilepsy

became more conspicuous; while in epilepsy and reading disorders (and epilepsy with

ASD), the advantage for the verbal scale was no longer seen. When the comorbidities

were taken as a starting point in the interpretation, it could be seen that the impact of the

epilepsy on the comorbidity was always in the direction of relatively “less lowered”

verbal abilities or relatively “less spared” performance abilities. The impact of epilepsy

on comorbidities seemed similar across the various diagnoses, suggesting a shared impact

of epilepsy on all conditions.

Consistencies and conflicting findings across studies.

Two conflicting findings could be seen across studies. First, while an effect of

seizure lateralization and of aetiology was found to be associated with intra-individual

subtest variability as scatter (Chapters 2 and 3), no association was found between these

epilepsy variables and verbal – performance discrepancies. Why would epilepsy variables

affect scatter but not verbal – performance discrepancies? Is the finding that seizure

lateralization or brain lesions influenced scatter a true finding? One approach to

understanding these findings would assume that there is no true difference and that


114

methodological differences across studies led to different results. Chapters 2 and 3 were

written according to a frequently applied methodology: various groups (here: left versus

right hemisphere onset seizure, presence or absence of lesions) were compared on a single

dimension (scatter) after ensuring that no major differences existed between possibly

confounding variables (like age at onset, age at testing). While valid and often applied,

such an approach may fail to account for the smaller cumulative effects of these variables,

and therefore lead to results which differ from studies which account for these variables

including them in the model, like regression-based analyses as applied in Chapter 5. This

means that more research is needed to replicate the results of Chapter 3, studying the

effects of test version, as well as of duration of epilepsy. The results on Chapter 3 provide

some indications that these variables may also be of influence on scatter. The other

approach, however, would assume that there is a true difference and that scatter is indeed

influences by seizure lateralization. In the next section it will be discussed that the

reorganization that takes place in the epileptic brain leads to sparing of verbal abilities

and possibly to more scattered patterns of verbal abilities – literally reflected in elevated

scatter in lesional left hemisphere seizures.

The second seemingly incongruent finding is that no elevated rates of children

showing reliable cognitive change on the performance scale was seen in Chapter 4, while

Chapter 5 reports a protracted decline on the performance scale over time. This

incongruence is easily solved. The cut-off score needed to qualify for reliable cognitive

change on the performance scale is 18 points, larger than the 14 needed on the verbal or

full scales, while performance IQ is already lowered at the time of first measurement.

More importantly, the children in Chapter 4 were followed for a period of time as short as

2.3 years, while the children in Chapter 5, who were slightly younger at first

measurement, were followed for a period of 2.8 years and some of them for as long as 5.1

years. Indeed, when a longer period is taken between test and retest, and cut-off scores are

adjusted for changes in test version, losses on the performance scale appear as well: 19%

of a sample of 26 children showed reliable loss on the performance scale (van Iterson &

Augustijn, 2013).

Otherwise, the results of the various studies appear remarkably consistent. Most

important, the inclusion of new subjects in Chapter 6, not reported upon in the earlier

chapters, yielded highly similar results (verbal > performance patterns) as the data in

earlier chapters (Chapters 2 and 5).

DISCUSSION

115

A Pattern of Cognitive Function and Cognitive Change in Children with Epilepsy

Taken together, a pattern of cognitive function and cognitive change over time in

children with epilepsy emerges – not described earlier. Children with epilepsy referred to

a tertiary centre due to concerns about their development are likely to show overall

lowered intellectual abilities compared to children without epilepsy and without

developmental concerns. These lowered abilities have been reported in children assessed

within the first months after the onset of epilepsy (B. Hermann et al., 2006), and could be

seen at first testing (Chapters 4, 5 and 6). IQs may often be unexpectedly low, and may be

discrepant with the school type the child was initially enrolled in (van Iterson, 2010),

suggesting that early in the course of epilepsy significant changes occur. These changes

have sometimes already been detected prior to the onset of the epilepsy in terms of school

failure (Hermann, Jones, Jackson, & Seidenberg, 2012; Schouten, Oostrom,

Jennekens-Schinkel, & Peters, 2001). At first testing, the full-scale IQ appears somewhat

lowered, but the two subscales are differentially affected. Soon after the epilepsy

surfaces, the performance IQ can be seen to be already depressed, while verbal IQ is

relatively “spared” (Chapter 5 and 6). It is suggested that the verbal scale at first testing

may be reflecting IQ scores that are closer in magnitude to the original, premorbid

cognitive potential of the child. The performance scale may be a better indicator of the

vulnerable reaction of the brain to the emerging seizure condition.

The sparing of verbal abilities.

Language abilities are known to be processed in their majority by the left

hemisphere. Particularly in young children, language abilities have shown great resilience

to brain disruptions. The most noticeable example is the fact that language functions can

be taken over by other brain areas in cases of severe disruption or brain development. A

change in lateralization of language functions is seen in children with early, e.g.,

prenatally, acquired left hemisphere brain lesions, when the non-affected right

hemisphere participates in language even to the extent of taking over most of the

language functions (Lidzba, Staudt, Wilke, & Krageloh-Mann, 2006; Loring et al., 1999;

Staudt et al., 2002). It signals the plasticity of the brain when it comes to retain and

“spare” the verbal abilities in the human being. Children with epilepsy may have

identifiable brain lesions, but most children with epilepsy do not; further, most children

with epilepsy, even in surgical series, will have language functions preserved in the left

hemisphere (Blackburn et al., 2007). There is an increasing body of evidence, however,

that children with epilepsy, including children with epilepsy syndromes of low severity,


116

are likely to show abnormal patterns of brain development (Braakman et al., 2013;

Hutchinson et al., 2010; Overvliet et al., 2013). Early sparing of verbal functions as

manifested by relatively high verbal IQ may point to a similar mechanism of

“prioritizing” the preservation of verbal abilities, not only in children with prenatal brain

lesions, but also in children with epilepsy. Although it remains speculative, this plasticity

in favour of the linguistic domain may partly explain why verbal abilities are relatively

resilient to epilepsy, particularly in children with early onset of epilepsy.

Of interest, the relative sparing of verbal abilities in epilepsy has also been found

in verbal memory tasks. Children with epilepsy have been found to outperform typically

developing children in verbal list learning in several studies, even if they show lowered

scores on virtually all other neuropsychological tasks, including visual memory

(Braakman et al., 2012; Høie, Mykletun, Waaler, Skeidsvoll, & Sommerfelt, 2006).

The vulnerability of the performance abilities.

The present work leads us to suggest that lowered performance abilities could

possibly be seen as an early cognitive marker of the vulnerability of the brain to the

underlying seizure condition. It is still incompletely understood why the performance

scale is most vulnerable to the seizure condition in its earliest stages and it is unknown

whether performance abilities may already have been lowered before the emergence of

the seizures. It has been suggested that latent changes occur in the brain before the onset

of epilepsy (Hermann et al., 2010). When the epilepsy surfaces, non-specific cognitive

problems become evident, which affect attention, executive functions, constructional

abilities, and visual-motor speed (Bhise, Burack, & Mandelbaum, 2010; Fastenau et al.,

2009; Hermann, Jones, Jackson, & Seidenberg, 2012; Hermann et al., 2006), even before

medical treatment is started (Bhise, Burack, & Mandelbaum, 2010). These abilities

underlie the performance scales, and may give rise to the lowered scores particularly on

performance abilities, and lead to the VIQ > PIQ pattern. The performance scale may also

be relying more on the ability of the child to deal with problem solving in novel

situations, and as such the epilepsy may be interfering with the ability to deal with

novelty in an adequate and speedy manner; it is possible that an intact functioning brain is

needed for performance on these tasks. VIQ > PIQ patterns have also been described in

other neurological samples, like children with prenatal acquired unilateral brain lesion,

regardless of side of lesion (Ballantyne, Spilkin, Hesselink, & Trauner, 2008; Lidzba et

al., 2006), and in children born preterm (Lind et al., 2011), confirming the vulnerability of

the performance scale in children with neurological conditions.

DISCUSSION

117

The VIQ > PIQ pattern may be considered at least to some extent as specific for

epilepsy and other children with neurological disorders. The pattern is different from non-

referred controls and from children with other developmental disorders. Typically

developing children show a flat pattern of cognitive abilities, while children with reading

disorders, math disorders and autism spectrum disorders tend to show “spared”

performance abilities, at least to some degree (Chapter 6). Epilepsy may co-exist with

another developmental disorder, like reading, math and ASD (Russ, Larson, & Halfon,

2012), and in these cases, the performance abilities also seem to be less spared.

Patterns at subtest level – the issue of scatter.

While at the level of the scales a difference between verbal and performance

abilities is observed in epilepsy, at the subtest level (Chapters 2 and 3), for the group with

epilepsy as a whole, elevated intra-individual variability (subtest scatter) is not seen.

Rather, subtest scatter was dependent on epilepsy variables like lateralization and brain

lesion and could be either increased (in lesional LH seizures) or decreased (in lesional RH

seizures), levelling off the scores of the whole group.

Increased subtest variability had been found in adults with brain abnormalities and

normal IQ (Ryan, Tree, Morris, & Gontkovsky, 2006) and in mentally deteriorating adults

in an early stage of the disorder (Reckess, Varvaris, Gordon, & Schretlen, 2014). These

findings suggest that, at least in adults, the combination of a neurological condition and

relatively spared cognitive abilities may lead to increased intra-individual subtest

variability. Children with brain lesions are likely to show a trajectory of brain

reorganization. In LH lesions, such reorganization may include changes in language

lateralization in some children (Lidzba et al., 2006) and, it may tentatively be said, that

these changes are associated with less consistent or more variable responses on the

different tasks sampled in the verbal scale, and therefore, more scatter.

From a psychometric point of view, a similar rationale could be given for verbal –

performance discrepancies and elevated scatter. In the standardization group of typically

developing children, the verbal and performance scale correlate substantially with each

other, the scales show high internal consistency, and the subtests within the scale show

variable – but overall high – correlations with the scale. The seizure condition and the

concomitant reorganizations of the brain may be reflecting themselves on the intelligence

scales as a “loosening” of the internal consistency within and between the subscales,

leading to overall lower correlations and larger differences among the various subtests. In

left hemisphere seizures, sparing of verbal abilities can be seen, but at the expense of a


118

scattered profile of verbal abilities; in right hemisphere seizures, a VIQ > PIQ pattern can

also be seen, without elevated scatter. This remains, however, a speculative explanation.

The children included in Chapter 3 showed great heterogeneity in brain lesions and the

observation that verbal scatter is increased or decreased in epilepsy depending on side of

brain lesion remains insufficiently understood and awaits replication.

Changes in patterns over time.

Over time, as the child with epilepsy grows older and the duration of epilepsy

increases, the pattern shown by the verbal and performance scale changes (Chapter 5).

The performance IQ, which was already low at the beginning, declines at a slow pace.

The verbal IQ, on the other hand, declines as well. This decline is faster in the earlier

years and goes on at a progressively slower pace over time – but continues over a

prolonged period of time. The detrimental impact of duration of epilepsy, particularly on

verbal abilities, has been seen in other studies as well (Caplan et al., 2008; Lopes et al.,

2013). Some authors have suggested that children with epilepsy may not keep up with the

increasing demands on integration of linguistic abilities and complex thinking as they

grow older, therefore showing lowered verbal scores (Addis, Lin, Pal, Hermann, &

Caplan, 2013). Particularly in older children developing epilepsy, previously acquired and

consolidated knowledge may generally remain preserved, but the epileptic condition may

be interfering with the acquisition of newer and more complex information. Given that

the intelligence test requires higher levels of knowledge and proficiency as the child

grows older, the failure to develop at the same pace as other children is reflected as a

lowered verbal IQ.

The lowered IQ over time seen in children with epilepsy may be interpreted as

having various gradations. First, most children with seizures retain most of the

consolidated knowledge and continue to develop, but at a slower pace than earlier. IQ

appears lowered over time (Chapter 5). As found in Chapter 4, while elevated rates of

children were seen showing cognitive loss, it can also be stated that even in a sample with

relatively complicated epilepsy, most children maintained an IQ within the boundaries

required to speak of “no reliable change in cognitive function”. Second, a significant

subset of children will show a significant deceleration of cognitive development and

learning, which will be detected by the intelligence scales as reliable cognitive loss.

Third, some children may lose abilities acquired earlier and, therefore will also exhibit

cognitive loss on the scales. The last two groups, those with a strong deceleration of

development and those with true loss of cognitive function, show reliable cognitive

DISCUSSION

119

change at retesting. Finally, some children may show a temporary loss followed by

recovery, and repeated testing over time may sample moments within this process of

losing, maintaining or recovering abilities.

With increased duration of epilepsy, the VIQ > PIQ pattern becomes less

conspicuous and tends to disappear. Decreased VIQ – PIQ gaps in children as a function

of duration of epilepsy, has also been observed elsewhere (Blackburn et al., 2007). When

the seizure condition remits, cognitive changes have occurred and are likely to continue –

at least in a portion of the children. Over time, the cognitive abilities seen in the child will

differ substantially from the premorbid abilities, both in terms of cognitive level and in

terms of cognitive pattern. This changing pattern is likely to have major impact on the

school career of the affected individual. Level of intelligence, and particularly verbal IQ,

is known to be strongly associated with school achievement (Glutting, Watkins, Konold,

& McDermott, 2006; van Haasen et al., 1986; Watkins & Glutting, 2000). Indeed, in a

study on secondary school children, failure to progress in school was found to be

associated with significant lowering of IQ relative to estimated premorbid IQ (van

Iterson, 2010). This failure to progress at school was seen in repetition of grades, being

set back to a lower type of school, or both. Lower VIQ and FS-IQ were also associated

with participation in special education (Chapter 5).

Some authors have described epilepsy as a life-long condition, with consequences

surpassing those of the seizures themselves and affecting long-term social outcome (C. S.

Camfield & Camfield, 2007). The changing level and the changing pattern imply that not

only the child, but also parents, schools and teachers have to adapt to the new level of

functioning as well as to the new pattern of strengths and weaknesses.

The Role of Epilepsy Variables.

Earlier studies on heterogeneous samples of epilepsy that used comprehensive

models failed to identify the impact of specific epilepsy variables on cognitive outcome

(Braakman et al., 2012; Oostrom, van Teeseling, Smeets-Schouten, Peters, & Jennekens-

Schinkel, 2005; Reijs et al., 2007). Overall, the present studies were in line with earlier

studies. For most epilepsy variables no clear impact on cognitive level or cognitive

change over time was found. Epilepsy variables influencing the verbal – performance

discrepancy are not easily identified. Most epilepsy variables did not show a relation with

pattern of IQ. For example, when topographical localization of seizure onset (frontal

versus temporal) was entered as a variable in the analysis reported in Chapter 5, no


120

differential impact was seen on cognitive level or cognitive course over time as a function

of seizure localization. This is congruent with studies reporting that dysfunctions

generally known as “frontal dysfunctions” may also be seen in children with temporal

lobe seizures, like attention and executive functions (Campiglia et al., 2014; Riccio,

Pliego, Cohen, & Park, 2014; Rzezak et al., 2007).

The present work failed to find a significant effect of syndrome severity. This lack

of association may be, at least partly, due to “restriction of the range” of syndrome

severity. On one end of the continuum, in the present work, more children with moderate

and severe epilepsies were included and relatively fewer children with epilepsies of low

severity. On the other side of the continuum of severity, only children who were able to

participate in (repeated) testing were included. Omitting children who were unable to take

the test may have excluded children with most severe syndromes. It is known that

children with epilepsy may show unstable courses of remission and relapse (C. Camfield,

Camfield, Gordon, Smith, & Dooley, 1993), and the present sample likely also included

such children, obscuring the possible effects of seizure freedom on cognitive

development.

Also, no significant role of medication on cognitive pattern and cognitive changes

was found when the number of antiepileptic drugs (AEDs) tried was included in the

analyses (Chapter 5). The role of number of AEDs used at the time of neuropsychological

testing was analyzed on children tested on the factor scores of the WISC-IIINL (van Iterson

& Augustijn, 2014). That study showed that a verbal > perceptual pattern could be seen

for the children with epilepsy, regardless of whether they were on or off medication. An

effect of medication appeared only in children using two or more antiepileptic drugs,

suggesting lowered processing speed. The results were in line with studies showing that

the impact of the seizure condition on cognitive function is larger than the added impact

of the AED. Only after the 3rd drug, the impact of AED on cognitive functions, like

executive functions and speed of processing, becomes evident (Witt & Helmstaedter,

2013).

The present studies also revealed some contribution of distinct epilepsy variables.

These concerned mainly time-related variables.

Age at onset of epilepsy: Age at onset shows an overlap with seizure syndrome

and epilepsy syndrome severity given that many epileptic syndromes have a time window

wherein they appear: most epileptic encephalopathies, considered the worst epilepsies,

surface in the first years of life (Covanis, 2012). In a study that examined age at onset of

DISCUSSION

121

epilepsy in association with response to medication, (Berg, Zelko, Levy, & Testa,

2012), early age of onset was associated with poor outcome only in children whose

seizures did not remit with medication. In the present study, with children with more

complicated epilepsies, age at onset was a fair estimator of cognitive trajectory over time

and, therefore, a fair marker of severity.

Duration: Duration of epilepsy was a significant predictor of cognitive level and

of cognitive decline over time. The best measure of duration was not linear, but

logarithmic. This means that soon after the onset of epilepsy, decline is most pronounced.

Thereafter, it continues in an increasingly slower pace – decline levels off. These results

imply that information about the duration of the seizure condition should be added to all

studies on epilepsy in children. In terms of epilepsy, the passing of time is characterized

by seizure remission in some children, a pattern of remissions followed by relapses in

other children, and no seizure control in still others (Geerts et al., 2010). The current

study presented a pattern of decline where no significant contribution of active versus

inactive seizures was found. Interpretation of these results is aided by the studies that

indicate lasting changes in brain development and brain organization, including

connectivity and neuronal density in children with various epileptic conditions, like

centrotemporal spikes or frontal lobe seizures (Braakman et al., 2013; Kanemura, Sano,

Tando, Sugita, & Aihara, 2012; Overvliet et al., 2013). Based on cases with duration of

epilepsy up to almost three years, Kanemura (2012) highlighted that seizures from frontal

lobe origin are associated with disturbance of prefrontal brain growth over time. Already

with this relatively short duration of seizures, a relationship between duration of seizures

and level of disturbed brain growth could be seen. Also, this disturbed brain growth lasted

beyond seizure remission. Thus, long-term changes in cognitive functions seen in the

present work may be the cognitive counterpart of the long-term changes in brain

development.

Overall, recent studies, including the present work, converge towards showing that, at

least to some extent, epilepsies of various degrees of severity, share neuropsychological

outcome patterns. There is an increasing number of publications reporting abnormalities

in brain structure and decreased or abnormal connectivity between brain areas in children

with epilepsy, expanding well beyond the areas involved in seizure generation and lasting

beyond seizure remission (Braakman et al., 2013; B. P. Hermann et al., 2010; Hutchinson

et al., 2010; Kanemura et al., 2012; Overvliet et al., 2013). The findings that brain


122

abnormalities occur at areas distant to the primary epileptogenic area may explain why

epilepsy variables like side and topographical site of epilepsy fail to show relationships

with cognitive level, cognitive pattern and cognitive course over time, and why long-

lasting cognitive changes can be seen in children with epilepsy. In addition, from the

literature it becomes increasingly clear that various (genetic) causes may clinically

converge to a single epileptic syndrome, as in juvenile myoclonic epilepsy (Koepp,

Thomas, Wandschneider, Berkovic, & Schmidt, 2014), and also that epileptic syndromes,

even those with a shared cause, can manifest as a spectrum, with various degrees of

severity (Rudolf, Valenti, Hirsch, & Szepetowski, 2009). These findings are in line with

the present results showing that cognitive decline may be associated will all kinds of

epilepsy.

The Role of Comorbidities.

In addition to time-related epilepsy variables, comorbidities also were found to

impact on the pattern of abilities. Without making reference to possible causative factors,

the results presented in Chapter 6 may be discussed within the framework of models of

comorbidities described in the literature (Pal, 2011). One model of Pal assumed that

comorbidities are caused by the epilepsy. Epileptic networks could be interfering with

cognitive networks related to a particular ability (e.g., reading or math), leading to a

reading or math problem. Some cases with comorbidities, like subsets of children with

math disorders or subsets of children with ASD, which may be characterized by higher

verbal than performance abilities, would fit this explanatory model. Overall, however, the

present data are not in line with this model and suggest that the cognitive networks

vulnerable to seizures are more likely the networks involved in the performance abilities,

rather than those involved in the verbal abilities, at least during the early stages of the

seizure condition.

According to a second model (Pal, 2011) the epilepsy and the behavioural disorder

may share risk factors, which may be present at any level of (biological) development

(Brooks-Kayal et al., 2013). The disorders may appear either in combination or alone, as

can be seen in families with members having either an isolated seizure disorder, an

isolated (language) disorder, or both (Clarke et al., 2007). Chapter 6, which included

familiarly non-related cases, suggests that the seizure condition and the behavioural

disorder show specific cognitive patterns. When seizure condition and comorbidity co-

occur in an individual, cognitive outcome patterns tend to have characteristics of both

DISCUSSION

123

disorders and, up to some extent, they seem to be the additive result of combining the

patterns of both isolated disorders.

The third model (Pal, 2011) is similar to the second, adding that when a

developmental disorder and epilepsy co-occur, they impact on each other, for example by

worsening the cognitive outcome. Worse outcomes in comorbidities have indeed been

reported (Hermann, 2008). It could have been expected that, for example in reading

abilities, a lowered verbal IQ characteristic of the reading ability, together with a lowered

performance IQ, characteristic epilepsy, would lead to an overall depressed IQ. Full-scale

IQ did not appear particularly lowered in comorbidities. Focus on the epileptic condition

may lead to under-referral for comorbidities (Helmstaedter et al., 2014). It is conceivable,

that in children with a relatively spared IQ lack of progress in a specific school area may

be reason for referral for diagnosis of specific learning problems; in contrast, in children

with epilepsy and a low IQ, lack of progress is more likely to be ascribed to the seizure

condition. In the present work, the lowered IQ was seen in comorbid math only (due to

lowered PIQ). The present data, however, indicate that epilepsy and comorbidity impact

on the cognitive pattern, showing characteristics of both disorders.

Limitations and Assets of the Studies

The data in the present studies were collected in clinical settings on an individual

basis from clinically referred children over a protracted period of time. This type of data

collection is associated with limitations and methodological challenges to ensure

generalizability of results.

Clinical data collection. Data on clinical groups are observational data, collected

in the clinical situation. If afterwards information is missing, for example because new

insights have demonstrated the relevance of specific variables, the missing information

can hardly be collected in retrospect. For example, data on parental education were not

available for the children tested earlier, given that data gathering on social economic

status was considered “not done” in the first stages of data collection. This means a

limitation in the presentation and interpretation of the data. An early decision was made

to collect data on the Wechsler Intelligence Scales from the children evaluated

systematically throughout a prolonged period of time. Doing so generated the research

hypotheses which were tested thereafter. Overall, however, similar results were found in

the present studies regardless of whether the newer or the older version of the WISC had

been used.


124

Single centre study. Participants with epilepsy were all collected within the

context of the tertiary epilepsy centre SEIN and its associated epilepsy school De

Waterlelie with its national school service centre, LWOE. This raises the question

whether the findings can be generalized to the rest of the country. The answer to this

question can confidently be affirmative: the two tertiary epilepsy centres which operate

nationwide can be considered largely equivalent; the two schools work together in the

services they provide. In addition, data coming from elsewhere, for example data on first

testing on children retested at the centre, were also included. Similarly, as detailed in the

next paragraph, the results should generalize to nations other than The Netherlands,

provided that the topic of interest is mixed samples of children with epilepsy referred to

tertiary centres due to developmental concerns.

Clinically referred children. In the present series, the children with epilepsy

included were all clinically referred children, for whom concerns about cognitive

development had risen and a risk for cognitive problems was seen. While the results of

the present series may not be valid for samples with uncomplicated epilepsies without

cognitive problems (B. Hermann et al., 2008; Jones, Siddarth, Gurbani, Shields, &

Caplan, 2010), the data may well generalize to the large number of children whose

epilepsy is not uncomplicated. Surveys conducted on large samples of children with

seizures show that the co-occurrence of epilepsy and school-related difficulties, far from

being exceptional, is very common. For example, Russ et al. (2012) presented data

showing that up to almost 75% of children with epilepsy acceded special (education)

services – which implies that large percentages had cognitive, learning or behavioural

problems, or combinations of these problems. The samples in the present work included a

clinically relevant broad spectrum of children, likely to represent these 75% which

needed assessment and specialized educational services.

Test versions. The present work refers mainly to two different test versions of the

WISC: the WISC-RNL and the WISC-IIINL. The WISC-IIINL is the most recent WISC available in

The Netherlands. Other countries however, like the US and the UK, have changed to the

WISC-IV (Wechsler, 2004), and the norm data collection for the standardization of the

WISC-V started in the US a few years ago (Alan S. Kaufman, 28.7.2013, San Diego,

personal communication), and the test was published in 2014. In this light, some WISC

data may be considered “historical”. After more than 60 years, the structure of the new

Wechsler test has undergone major changes. The WISC-IV no longer has the option for

calculating verbal and Performance IQ; instead the factor index scores verbal

DISCUSSION

125

comprehension and perceptual reasoning are given. In addition, the WISC-IV has a

Processing Speed factor and a Working Memory factor. The perceptual reasoning factor

of the WISC-IV has a lower reliance on motor dexterity and speed than either the

performance scale or the perceptual organzation factor of the WISC-III. The WISC-V took

this development towards more “pure” neuropsychological measures a step further, by

splitting the perceptual reasoning index into two factors – Visual Spatial and Fluid

Reasoning. The newest test versions will allow more fine-grained analyses of the

strengths and weaknesses of children with developmental disorders.

In spite of these changes, the verbal and performance-perceptual abilities have

continued to be the two core domains in cognitive measurement, also in the newest test

versions (Flanagan & Kaufman, 2009; Wechsler, 2004; Weiss, Saklofske, Schwartz,

Prifitera, & Courville, 2006). It remains open to what extent the results presented in this

work would have been different, had newer test versions been used. Interestingly, a recent

study on children with frontal and temporal seizures also presented data suggesting a

verbal > perceptual pattern of cognitive abilities (Campiglia et al., 2014).

The implications of these changes in test versions for the study of epilepsy are not

yet clear. For example, WISC-III studies have shown that the Working Memory index

and the Processing Speed index are affected in epilepsy (Berg et al., 2008b). While

of interest, such a pattern is unlikely to be specific for epilepsy, as there is evidence that

they can also be found in well-delineated samples with ADHD, as well as in mixed

samples with developmental disabilities (Devena & Watkins, 2012).

There are considerations in favour of the inclusion of traditional test versions.

Newer test versions are being developed at a quicker pace in the US and UK than in other

countries. Earlier WISC versions, like the WISC-III, are still being used internationally, and

if several test versions yield similar results (as seen for the WISC-R and WISC-III in terms of

cognitive patterns), generalizability of the results to other test versions or languages is

increased.

Second, it should be considered that the “broader” measures as included in the

traditional tests may have a higher ecological validity. For example, the constructional

tasks which also require motor ability and speeded responding in the performance scale –

due to their complexity – may be tapping an ability which is truly selectively lowered in

children with epilepsy and may be interfering with everyday functioning, more so than

more “pure” underlying neuropsychological functions. In this sense, one should consider

the possibility that the traditional measure may still be providing some kind of


126

information than the narrower, more “neuropsychological function”-like index scores no

longer provide. It has been suggested that while these narrower factors intend to measure

specific abilities, they may not be doing so in a sufficiently diversified way (Grégoire,

2013). Of interest in the present context, for example, the processing speed index is

limited to visual-motor speed, not speed of auditory processing (Grégoire, 2013).

Therefore, the finding that PSI is lowered in epilepsy, still does not provide information

as to whether auditory-verbal processing speed may also be compromised. A similar kind

of “loss” has been has been noted by clinical psychologists regarding all revisions of the

Wechsler scales following David Wechsler’s death in 1981: the elimination of clinically

rich items or even an entire subtest (Baron, 2005; Kaufman, 1994).

Classification of Epilepsy. New efforts toward a different conceptualization and

classification of epilepsy are being undertaken (Berg et al., 2010; Berg & Cross, 2010;

Berg & Scheffer, 2011). Indeed, the traditional classifications have been qualified as

“both antiquated and arbitrary” (Berg & Scheffer, 2011, p.1058). The new classification

system kept the dichotomy focal versus generalized seizures (but no longer focal versus

generalized epilepsies). The newly suggested aetiological classification acknowledges

genetic causes, structural, metabolic, or immunological causes and unknown causes. It

also acknowledges that in a single child more than one of these causes may be at stake,

like genetic and structural. A flexible approach depending on the individual needs is

encouraged. The studies presented in this thesis relied on the “traditional” classification

of seizures and epilepsies (Engel, 2006; I.L.A.E., 1981, 1989), which continued to being

used in The Netherlands. Actually, in The Netherlands, new guidelines for epilepsy were

launched recently, developed by the Dutch League Against Epilepsy (LIGA, 2013) on

behalf of the Dutch Society of Neurologists and based on this traditional classification.

The Dutch League decided to hold on to the traditional classification, while at the same

time acknowledging the existence of a revised classification and the utility of adding

information according to this new classification (Augustijn, 2014). Interestingly, results

presented in Chapter 5 of the present work, showing that progressive cognitive decline in

epilepsy can be seen regardless of epilepsy variables, are congruent with propositions of

the new classification which indicate that encephalopathic effects (i.e., cognitive decline)

may be seen in any form of epilepsy (Berg & Cross, 2010).

DISCUSSION

127

Future Directions

New developments. Special mention should be made of the rapid progress seen in

the field of epilepsy genetics (Addis et al., 2013; Koepp et al., 2014; Olson, Poduri, &

Pearl, 2014; Rudolf et al., 2009; Thomas & Berkovic, 2014) which undoubtedly will shed

new and possibly more decisive light on the results of the present papers. As mentioned in

Chapter 5, except for age at onset and duration of epilepsy, epilepsy variables did not

exert a significant contribution on the cognitive decline seen in epilepsy. Epilepsy

syndrome severity, for example, was not related significantly to cognitive decline. These

results suggest that some crucial information is still missing in the explanatory models.

New advances in genetics could be providing this key information as to why one child

previously diagnosed with a mild epileptic syndrome (e.g., absences or localization

related seizures) may develop in the expected manner, while another child will show

cognitive loss. The child showing cognitive decline may in fact have a genetic

abnormality or a pattern of abnormalities related to this cognitive loss.

The VIQ > PIQ difference as an early marker of a vulnerable brain. The

VIQ > PIQ pattern found in the previous chapters appears early in the course of epilepsy

(Chapter 5 and 6) and is largely specific for children with epilepsy. It is conceivable that

VIQ > PIQ could be a marker of the vulnerability of the brain to the underlying epileptic

condition which is about to emerge. It is proposed that this pattern may antedate the onset

of epilepsy. To examine this hypothesis, cognitive data of children before the onset of

epilepsy is of great value. One way of generating such data is the epidemiological study,

like the study conducted by Ellenberg et al. (1986). In the study of Ellenberg et al., from

45.000 children, 62 new cases with epilepsy had emerged between the first test at age 4

and the second test at age 7. A “not normal” status at age 4 was reported in 21% of these

children. Such a not normal status or a developmental problem may lead to testing before

the onset of epilepsy. Therefore, a second way of obtaining these data is to retrospectively

collect the data from children evaluated before the onset of epilepsy. The present studies

were limited to children assessed after the diagnosis of epilepsy was confirmed, but data

from children assessed for developmental problems before the emergence of epilepsy are

being collected as well. These data could be useful to test hypotheses associated with

cognitive patterns antedating epilepsy onset. Also, the cognitive status of children before

the onset of epilepsy could be studied in children who are at high developmental (e.g.,

genetic) risk for developing seizures.


128

Specific samples. Further research on patterns of cognitive development in

children with epilepsy could include specific samples, like children with night-time

epileptic activity and children who have had status epilepticus. Two clinical observations

on exceptional cases may generate new studies. First, the VIQ > PIQ pattern

characteristic of children with epilepsy was not observed clinically in the small subset of

children with epilepsy born to mothers with epilepsy and prenatally exposed to anti-

epileptic drugs. Earlier studies have suggested that lowered verbal abilities may be

characteristic of children who were exposed to AEDs in utero (Meador et al., 2012).

Further research should be conducted to elucidate whether the VIQ > PIQ pattern so often

seen in children with epilepsy is reversed in children born to mothers with epilepsy.

Second, elevated rates of mortality in childhood-onset epilepsy have been reported

in a long-term follow-up study (Sillanpaa & Shinnar, 2010). Risk factors associated with

mortality were a severe cognitive impairment, symptomatic cause and active seizures for

which no medication was taken. On an observational level, during the period of data

collection, (sudden) death in epilepsy has been reported in some of the children who

presented with reliable cognitive decline. This raises the question whether cognitive

decline may be a marker of an increased vulnerability of the child or youngster to

(sudden) death.

Epilepsy and autism. Chapter 2 revealed overlap between autism spectrum

disorders and children with left hemisphere seizures: both samples showed increased

intra-individual variability on the verbal and full scales. The sample with autistic

spectrum disorders showed increased variability on the performance scale as well. This

result may be of particular relevance given the increasing interest in the co-occurrence of

autism spectrum disorders and epilepsy (Berg & Plioplys, 2012; Besag, 2009; Tuchman,

Alessandri, & Cuccaro, 2010). The overlap between the two conditions is suggested as

being as high as 15 to 30% (Russ et al., 2012; Tuchman et al., 2010). The overlap is seen

particularly in children with intellectual disabilities and in epileptic encephalopathies

(Besag, 2009; Tuchman et al., 2010). Some authors (Berg & Plioplys, 2012) point out that

it is still unclear whether the overlap is mediated by intellectual disabilities or whether it

is also seen in children with normal intellectual abilities. The results described in the

present study suggest a shared feature – increased verbal subtest scatter – in children with

close to average IQs (FS IQ ~ 92). Further research in this area of increased variability

may aid in the understanding of mechanisms shared by both conditions. For example,

increased variability in ASD has been associated with the lowered problems with the

DISCUSSION

129

integration of information (or deceased central coherence) in the ASD literature (Joseph,

Tager-Flusberg, & Lord, 2002; Noens & van Berckelaer-Onnes, 2005). Further research

could be aimed at studying whether these concepts would also be helpful in describing

children with epilepsy. In addition, increased variability within tasks has also been

described in ASD (Geurts et al., 2008). Variability within tasks has been associated with

abnormal patterns of brain development (Towgood, Meuwese, Gilbert, Turner, &

Burgess, 2009) as well as with deficits of working memory (Kofler et al., 2014). Of

interest, no differences emerged in children with isolated epilepsy and epilepsy with

comorbid ASD in terms of cognitive patterns (Chapter 6), again suggesting similarities

between the two conditions which would warrant future research.

Relevance of the Results

Beyond the significance of the results already indicated, some additional aspects

should be highlighted which may be of interest for researchers and clinicians.

Cognitive change over time and first and later testing. The present papers

suggest that cognitive data in children with epilepsy change over time. For research, one

implication of this finding is that the samples described in papers should be well

delineated. Information on age at testing, age at seizure onset and information on duration

of epilepsy is essential. It is also important to know whether earlier clinical assessment

took place before or after the onset of epilepsy, whether the test reported is the first or a

later test, whether assessments have occurred outside the research setting, whether the

interval between tests is long enough to preclude practice effects, whether all subjects

were tested with the same test version, and how all of this information is being dealt with.

To date, this information is often insufficiently accounted for in the literature.

Intra-individual subtest variability. Two opposing problems emerge relative to

intra-individual subtest variability. The first relates to ignoring variability at the time of

using short forms of the WISC in research; the second relates to the overinterpretation of

intra-individual subtest variability in clinical practice.

The use of short forms. Researchers on epilepsy often limit their Wechsler testing

to short forms (Bailet & Turk, 2000; Berg et al., 2008a, 2008b; Gülgönen,

Demirbilek, Korkmaz, Dervent, & Townes, 2000). One of the criteria in the use of a

short form is that 81% or more of the estimates of a scale fall within the 90% confidence

interval of the full-length scale (Donders, Elzinga, Kuipers, Helder, & Crawford, 2012),

but no efforts are done by the authors to ensure that this criterion is met in the sample


130

with epilepsy. Chapters 2 and 3 indicated that elevated subtest scatter can be found in

specific subsamples, suggesting a differential impact depending on seizure onset side and

presence of MRI-abnormalities on the verbal scale. If replicated with other samples, a

consequence of this finding is that the use of short forms in WISC testing, either for

clinical or research purposes (particularly on lateralized epilepsy and surgical studies),

should be applied with great caution. The best practice is to avoid the use of short forms

and administer all subtests. If this is not possible, both clinicians and researchers should

be aware that in particular samples, a short form is more likely to yield estimation errors.

Overinterpretation. In The Netherlands, in clinical practice, some clinicians

refrain from calculating and reporting IQ values in their psychological reports, arguing

that the variability seen between subtests is too large while, in fact, the observed

variability is normal variability. Already in the 1970s, Kaufman (1976) warned for this

overinterpretation of normal subtest variability, and presented base rates tables for subtest

scatter (Kaufman, 1979). The values given in Chapters 2 and 3 may help Dutch clinicians

to base their conclusions in concert with the psychometrically valid cut-off values for

Dutch data. In addition, base rate tables are given in the Appendixes to provide the

clinician with empirical information.

Epilepsy, comorbidities and cognitive patterns. The impact of the comorbidities

on cognitive patterns in children with epilepsy as seen in Chapter 6 and described in the

literature (B. Hermann et al., 2008) suggest that information on comorbidities is relevant

in studies on epilepsy. This is true regardless of whether the starting point is epilepsy (and

its learning and behavioural comorbidities) of neurodevelopmental disorders (and

comorbid epilepsy).

Importantly, an impact of the epilepsy on the cognitive pattern was seen, similar

across conditions. In all cases, the performance strength tended to disappear or a

performance weakness tended to increase in the light of epilepsy. These changes in

patterns in the light of comorbidities are relevant for the clinician. When testing children

for comorbidities like specific learning disorders or with behavioural conditions, the

clinician should be aware that the pattern seen in the disorder may be different from the

pattern seen in the isolated condition, and that the remediation measures should be

adapted as well. Thorough neuropsychological evaluation of the comorbidity is

recommended.

The possibility that reliable cognitive decline has occurred in epilepsy (Chapter 4)

should also be borne in mind whenever testing children for comorbidities. In educational

DISCUSSION

131

practice, “insufficient response to intervention” may be seen in a child. This lack of

scholastic progress may lead to referral questions on “specific learning disabilities”. It

may indeed be the case that the child has the learning comorbidity. In these cases, the

second diagnosis is warranted. The clinician should ensure, however, that insufficient

school progress is not, in fact, epilepsy-related “stagnation of development”, that is,

reliable cognitive decline, which is observed at school as difficulties to progress on school

subjects. Retesting the child for its cognitive abilities and comparing the scores with

earlier scores is recommended before diagnosing specific learning disabilities. As found

in Chapters 4 and 5, cognitive decline can occur in any child with epilepsy.

Cognitive change over time and educational measures. The findings on

cognitive decline at group level and reliable cognitive decline on the verbal scale in as

many as one out of four children, have significant implications for the school career and

the educational facilities for the children with epilepsy. The measures should be tailored

to suit the needs of the individual child. This means remediation for the lowered

performance abilities at baseline, adaptation of the curriculum to the specific difficulties

in terms of speed and visual, spatial and motor abilities. Also, advantage should be taken

of the assets of the child, like the higher verbal abilities seen in many children, both in

order to support the performance abilities as well as to keep standards high and try to hold

decline. It also means searching for specific ways to monitor development and academic

achievement and adapt the curricular demands to the new level of the child. It also may

mean psycho-social coaching of the child and the parents to enable adaptation to the new

level. Most of all, the findings should encourage the continuation of the educational

measures. Follow-up assessments and help for children with epilepsy should be offered

well beyond seizure remission. Beyond educational facilities, the finding of reliable

cognitive decline in children should alert clinicians to continue searching for causes and

treatment possibilities.

Clinical Data: The Appendices

As said in the Introduction, research should also be “consumer friendly”. All

chapters, except Chapter 5 (and Chapter 6, for which they are provided in Appendix E),

provide some information on individuals: frequency of occurrence, sensitivity, specificity

or results from Receiver Operating Characteristics. In addition, the Appendices provide

the clinician with base rate tables on intra-individual subtest variability (scatter; Appendix

A), verbal – performance discrepancies (Appendix B), discrepancies between factor index

scores (Appendix C) and cognitive change after serial testing (Appendix D).

Summaries

SUMMARY

135

Summary

(English Summary)

Cognitive Patterns in Paediatric Epilepsy

Intra-individual variability, cognitive patterns and patterns of cognitive change in children

with epilepsy on the Wechsler Intelligence Scales for Children

The present work includes five studies on the cognitive patterns in children with epilepsy.

This work is based on clinical data. The testing of children with various kinds of

disorders during a protracted period of time suggests that children with epilepsy displayed

patterns of cognitive abilities that differed from the patterns seen in other disorders. Such

clinical experience has sometimes been called the “internal database” of the clinician.

However, to be of true value, the internal data base should be confirmed with an

“empirical database”. The aim of the present study was to describe cognitive patterns in

children with epilepsy who were tested with the Wechsler Intelligence scales for children

(WISC series).

There is a large body of evidence that cognitive problems exist in children with

epilepsy. These problems also include the verbal and non-verbal (performance) abilities

of a child with epilepsy. Verbal and performance abilities are core abilities that are

measured in the intelligence scales for children as Verbal IQ and Performance IQ,

abbreviated as VIQ and PIQ, respectively. They are sampled in a standardized, well

normed and internationally widely accepted manner in the WISC series. Verbal and

performance abilities are partly independent of each other, but they also show a

substantial correlation, suggesting that they are both measures of the general factor IQ.

The verbal and performance IQ scales are comprised of subtests. The subtests are also

partly independent of each other but correlate with each other as they are all understood

be a measure of, for example, the verbal abilities.

Verbal and performance abilities are also the core abilities investigated in the

present work. While well researched, relatively less is known about the cognitive patterns

displayed on the verbal and performance scales by children with epilepsy. Cognitive

patterns relate to profiles of strengths and weaknesses and are, therefore, measures of

variability.


136

One kind of variability is intra-individual subtest variability or subtest scatter. In

the literature – also on disorders other than epilepsy – it is still a matter of debate whether

increased intra-individual subtest variability is a sign of some kind of pathology or

whether it should be considered an expression of the statistical properties of the test. The

discussion remains important, because, for example in The Netherlands, it is common for

clinicians – also in the field of epilepsy – to say that the pattern displayed by a child

shows too much variability, and therefore does not provide a reliable measure of the

child’s abilities. They refrain from reporting IQ. In fact, for the Dutch Wechsler tests it is

largely unknown what can be considered normal scatter, and whether unusual scatter may

have any clinical diagnostic value in a child with epilepsy.

A second kind of variability relates to the pattern displayed by the verbal and

performance scales (VIQ and PIQ). A verbal – performance discrepancy suggests that a

particular domain is more compromised than the other. Existing studies lead to

conflicting results – even within a single epilepsy syndrome – as to which scale is most

lowered in epilepsy. Little is known about the differences in cognitive patterns between

children with epilepsy and other developmental disorders. Even less is known about

cognitive patterns in children affected by two conditions, that is, children with

comorbidities in epilepsy.

A third kind of variability relates to the changes over time that may occur in

children with epilepsy. Children with epilepsy often have seizures during a prolonged

period of time. During this period, they are expected to develop, while intermittent

epileptic discharges hamper cognitive functioning. The changes over time that may be

seen on the verbal and the performance scales, the possible differences in these changes

across the various cognitive domains, and the variables that affect these changes, are still

largely unknown.

The principal research questions of the present work relate to measures of cognitive

patterns or intra-individual variability. Do children with epilepsy present with more intra-

individual subtest variability than expected from the psychometric properties of the test?

Do children with epilepsy show lower scores on a particular measure and higher on

another? Do they show changes over time after serial testing? Do patterns change over

time?

SUMMARY

137

If increased intra-individual subtest variability can be found, particular cognitive

patterns or patterns of cognitive change over time – can variables be identified, associated

with these patterns?

The samples that participated in the different studies were referred children with epilepsy

who were tested because concerns had risen about their cognitive development. They

came to a tertiary epilepsy centre in The Netherlands or its associated school for epilepsy.

The school provided educational services, not only in the special school for epilepsy

proper, but in any primary or secondary school for regular or special education in the

northern half of the country. For clinical comparison, some samples of children with other

disorders (learning disorders, behavioural disorders) were included, as well as typically

developing children for whom no concerns about their development were known.

For all children, WISC data were collected. They related to the two most recent test

versions in The Netherlands, the WISC-RNL and WISC-IIINL, and sometimes to the WPPSI-RNL

as well (the superscript was applied to all Dutch test versions). For children with epilepsy,

additional data were collected from medical or neuropsychological reports. These data

concerned epilepsy variables, like age at onset of the epilepsy, type of seizures (focal

versus generalized), lateralization (left versus right hemisphere onset seizures), the

topographical localization of the seizures (e.g., frontal or temporal), the number of

different anti-epileptic drugs tried during the course of the epilepsy, and the presence of

brain lesions visible on MRI. Based on these data, other information could be extracted

such as the duration of the epilepsy or the severity of the epilepsy syndrome. Data on

participation in regular or special education and data on parental education were collected

as well.

After a general introduction, Chapter 2 studied intra-individual subtest variability, that is,

differences between the lowest and the highest subtest score. Based on 467 children with

developmental disorders (157 with epilepsy, 132 with learning disorders, 178 with

behavioural and psychiatric disorders) it was shown that scatter, while indeed somewhat

elevated in referred children, was not a general sign of pathology. Rather, differences

emerged depending on the sample studied. Children with epilepsy, and children with

learning disorders, did not display elevated subtest scatter. Elevated scatter was seen in

children with behavioural and psychiatric disorders (especially in children with autism


138

spectrum disorders.) Within the sample with epilepsy, increased variability appeared to be

dependent on lateralization: children with left hemisphere seizure activity showed

increased variability, not seen in children with right hemisphere seizures.

Chapter 3 studied subtest scatter in lateralized epilepsy in relation to brain lesions.

The study included 90 children with lateralized epilepsy, of whom 56 had epilepsy

emanating from the left hemisphere (LH; 22 of these children had shown abnormalities

on neuroimaging) and 34 had epilepsy from the right hemisphere (RH; 15 had MRI

abnormalities).It was found that there was a differential impact of side and lesion on

subtest scatter. Children with LH seizures and MRI lesions, displayed increased subtests

scatter; children with RH seizures and MRI lesions, displayed decreased scatter. In the

general Discussion it was speculated that in children with LH seizures and brain lesions,

reorganization of brain functions may lead to preservation of verbal abilities at the

expense of more scatter.

Chapter 4 studied the rate of children who showed reliable cognitive change at

retesting. Reliable cognitive change is a change in scores that is seen in less than 10% of

the children of a reference sample (5% should show gains, 5% losses). The cut-off values

were estimated based on a reference sample of referred children with developmental

disorders but without epilepsy. The data came from the literature, and were based on

children tested twice with Dutch versions of the WISC. Based on 73 children with epilepsy

tested two times with the same test version, after a mean interval of 2.3 years between test

and retest, 26% showed reliable cognitive change – as decline – on the verbal scale and

16.4% on the full scale. No increase in decline was seen on the performance scale in this

sample. Some children also showed reliable cognitive gains. The rate of children showing

gains, however, never surpassed the expected 5%.

Chapter 4 had suggested that there was an increased risk in epilepsy to show

cognitive decline over time. In Chapter 5, the variables associated with this decline were

studied. Based on 113 children tested two or three times with the Wechsler scales, decline

could once again be seen. The following variables did not contribute significantly to the

patterns of change: Hemispheric side or site of seizure onset, number of anti-epileptic

drugs tried over time, brain lesions, severity of the epileptic syndrome, or the actual

presence of seizures. In addition, interactions between these factors did not contribute to

the patterns change. The results were interpreted in the light of recent epilepsy studies,

which described lasting changes in the brain – for example in brain connectivity – in

areas distant from the sites of seizure onset, and lasting beyond seizure remission.

SUMMARY

139

The Chapter 5 study revealed interesting insights into the patterns of decline. A

lowered performance score could already be seen early in the course of the epilepsy,

while the verbal scale appeared relatively “spared.” Throughout the years, however, this

pattern changed. The verbal scale started to decline steeply; the performance scale also

declined, but less strongly; and the relative advantage of the verbal scale over the

performance scale could no longer be seen. The cognitive pattern changed over time as a

function of duration of epilepsy. At first, a steep decline could be seen, followed by a less

pronounced decline, which continued throughout several years. The result was a closing

of the VIQ > PIQ gap over time.

In addition to the variable “duration of epilepsy”, a second time-related variable

was found to contribute to the pattern: age at onset of epilepsy. Epilepsy starting early in

life was associated with a relatively higher VIQ at first testing and a relatively steeper

decline afterwards. Overall, there was large variability between children in terms of

cognitive patterns and patterns of decline.

Lowering of IQ scores could also be seen regardless of the school environment

(regular or special education), and regardless of the home environment (higher or lower

educated parents). As expected, IQ was higher in regular education relative to special

education, but decline was independent of school type. Higher parental educational level

was associated with higher IQ-scores in children, but was not “protective” against

decline.

The VIQ > PIQ pattern is not always seen in samples of children with epilepsy.

When epilepsy and a second disorder came together, the patterns appeared different.

Chapter 6 studied the cognitive patterns of 117 children with isolated disorders, including

epilepsy, but also reading disorders, math disorders and autism spectrum disorders. The

term “isolated” was applied to indicate that the children, while they might have had other

neuropsychological problems, did not have a diagnosis of a second disorder. This study

indicated that the children with epilepsy had a VIQ > PIQ pattern, while the other groups

had either a flat pattern of abilities or a pattern favouring the performance scale (most

clearly in reading disorders). Control children also had a flat pattern of abilities. In

addition, we included 171 children with epilepsy, of whom some had epilepsy as an

isolated condition, and others as a comorbid condition. The comorbidities were – again –

reading disorders, math disorders or autism spectrum disorders. Results of this study

indicated that the VIQ > PIQ pattern, characteristic for isolated epilepsy, was different in

the presence of a comorbidity. With comorbid reading disorders, the pattern appeared


140

attenuated; with comorbid math disorders, the pattern appeared exacerbated, with

comorbid autism spectrum disorders, no clear pattern emerged. Interestingly, when the

“other” disorder was taken as a starting point, and epilepsy co-occurred, there seemed to

be a systematic “shift” towards more compromised performance abilities or more spared

verbal abilities

Overall, a pattern of cognitive abilities and a pattern of cognitive changes emerged. In the

initial stages of the seizure condition, a pattern of relatively “spared” verbal abilities and

relatively “compromised” performance abilities is seen. The verbal abilities may be a

better indicator of the premorbid cognitive potential of the child. The performance

abilities may be an indicator of the vulnerable reaction of the brain to the seizure

condition. In some subsets of children, such as children with left hemisphere onset

seizures and brain lesions, the spared verbal abilities may appear with increased scatter on

the verbal scale. With increased duration of the epilepsy, the VIQ > PIQ pattern is likely

to change: the verbal scale starts to decline, steeply at the beginning, and more gradually

thereafter. The performance scale also declines further. Over time, the VIQ > PIQ gap is

no longer seen. The great majority of the children with epilepsy maintain a cognitive level

within a boundary wherein lowering of scores was seen, but not “reliable cognitive

change.” However, the percentage of children who showed such reliable change – as loss

– was elevated. This threefold decline on the full scale (and fivefold decline on the verbal

scale) often means that the children will fail to follow the educational trajectory they were

initially were enrolled in. They repeat grades, need special educational assistance, or, if in

secondary education, are set back to a lower type of education. Conceivably, these pattern

of change over time requires the children, the parents and the teachers to adapt to the new

situation. The VIQ > PIQ pattern seen in epilepsy appears modulated by two variables:

first, the duration of epilepsy which leads to the closing of the VIQ – PIQ gap. Second,

the presence of comorbidities which may decrease or augment the VIQ > PIQ difference.

The result of these studies lead us to hypothesize that lowered performance

abilities could be a cognitive marker of the vulnerability of the brain to the epileptic

condition, and it would be valuable to study whether this vulnerability can already be

seen before the emergence of the seizures proper.

The results of the different studies from the present work may be of utility for researchers,

policy makers and clinicians. For researchers, they results highlight that information on

SUMMARY

141

age at onset and duration of epilepsy, as well as on the presence of comorbidities, is

important when describing samples with epilepsy. In addition, the use of short forms

should be discouraged, given that subsets of children with epilepsy may show increased

subtest scatter and lead to increased measurement error. Wherever short forms cannot be

avoided for practical reasons, they should be tested for their utility in epilepsy.

Policy makers in education should take note that the cognitive trajectory seen in

children with epilepsy may be unlike the trajectory seen in other children with

developmental disorders. Children with epilepsy may therefore require specific expertise

from specialized school services, also beyond seizure remission.

When testing children with epilepsy, the clinician should be aware that children

with epilepsy may have a particular pattern of cognitive abilities, often VIQ > PIQ, and

that this pattern need not be maintained over time. Rather, it is likely to change over time.

When the child is diagnosed with a second condition (learning disorders, autism), the

clinician should keep in mind that the pattern displayed by the child may be different

from both the pattern characteristic for epilepsy as well as the pattern characteristic for

the other condition. The clinician should also bear in mind, that failure to progress in

school need not point to a specific learning disorder. Rather, it may point towards

cognitive decline (and therefore stagnation of school development), which may be much

more frequent in epilepsy than in other disorders without epilepsy. Increased intra-

individual subtest variability may be seen in some samples with epilepsy, and may prove

to have clinical value.

Finally, to make the results of the work “consumer friendly”, the appendices provide base

rate tables for children tested with various Dutch WISC test versions, (a) on subtest scatter;

(b) on discrepancies between scales and factor index scores; and (c) on change in IQ

scores over time. These tables are based on large numbers of clinically referred and non-

referred samples.

SAMENVATTING

143

Samenvatting

(Dutch Summary)

Cognitieve patronen in kinderen met epilepsie

Intra-individuele variabiliteit, cognitieve patronen en patronen van cognitieve verandering

in kinderen met epilepsie op de Wechsler Intelligence Scale for Children

Dit proefschrift bevat een vijftal studies over cognitieve patronen (cognitieve profielen)

bij kinderen met epilepsie.

De oorsprong van dit werk is klinisch van aard. Gedurende vele jaren van klinisch

neuropsychologisch werk met kinderen met allerlei verschillende

ontwikkelingsstoornissen viel op dat bij kinderen met epilepsie zich specifieke profielen

aftekenden, die verschilden van de profielen bij andere stoornissen. Klinische

opgebouwde ervaring is wel eens “de interne database” genoemd van de clinicus. Mooi

en belangrijk, maar pas werkelijk betekenisvol wanneer deze gestaafd kan worden aan

een “empirische database”. Het doel van dit onderzoek was om de patronen te beschrijven

bij kinderen met epilepsie die getest waren met de Wechsler Intelligentietest voor

kinderen (WISC).

In de wetenschappelijke literatuur is uitvoerig gedocumenteerd dat bij kinderen

met epilepsie vaak sprake is van cognitieve stoornissen. Tot de cognitieve vaardigheden

waarbij lagere scores worden gevonden, behoren in ieder geval ook de verbale en de niet-

verbale (performale) vaardigheden. Verbale en performale vaardigheden vormen de

hoofddimensies die door de intelligentietesten in kaart worden gebracht. Op een

gestandaardiseerde, goed genormeerde en internationaal erkende wijze gebeurt dat met de

verbale schaal (VIQ) en performale schaal (PIQ) van de Wechsler intelligentietests. De

verbale en de performale schaal zijn onderverdeeld in subtests. De verbale en performale

schalen zijn gedeeltelijk onafhankelijk van elkaar, maar vertonen ook een sterke

samenhang omdat ze beide de algemene vaardigheid weerspiegelen (totaal IQ of FS-IQ).

Iets vergelijkbaars kan gezegd worden van de subtests binnen een schaal: ze zijn

gedeeltelijk onafhankelijk, maar weerspiegelen allemaal de onderliggende vaardigheid, zo

weerspiegelen de verbale subtests het verbaal IQ.

Verbale en performale vaardigheden vormen ook de kern van het huidige werk. Er

is nog onvoldoende bekend welke relatieve sterktes en zwaktes op deze vaardigheden te


144

zien zijn bij kinderen met epilepsie. Het profiel van sterktes en zwaktes binnen het

testprofiel wordt hier aangeduid met “cognitieve patronen”, en heeft betrekking op de

variabiliteit van de testscores.

Een van de vormen van variabiliteit is de intra-individuele subtestvariabiliteit op

een schaal. Een subtestprofiel kan vlak zijn maar ook grillig – deze subtestvariabiliteit

wordt in de Engelstalige literatuur “subtest scatter” genoemd. Nog steeds is in de

literatuur – ook buiten de epilepsieliteratuur – discussie over over de vraag of verhoogde

intra-individuele variabiliteit een kenmerk is van een “stoornis”, of slechts een uiting is

van normale variatie op een test (zoals verwacht gezien de psychometrische

eigenschappen van de test). Toch is het belangrijk om dit te weten omdat clinici – ook in

de epilepsie – er snel toe neigen te spreken van een te grillig patroon van vaardigheden,

dat “het zicht op de werkelijke cognitieve vaardigheden belemmert”, en ertoe over gaan

het IQ niet te rapporteren. Dit doen ze zonder dat er voor de Nederlandstalige tests

bekend is, welke kritieke waarden nodig zijn om te kunnen spreken van zeldzaam hoge

subtestvariabiliteit (variabiliteit die bij minder dan 5% van de kinderen van een

normgroep voorkomt). Ook is niet bekend of verhoogde variabiliteit mogelijk van

klinische betekenis is bij het kind met epilepsie.

Een tweede type variabiliteit is het cognitief patroon of cognitief profiel dat zich

aftekent tussen de verbale en de performale vaardigheden, de verbaal – performaal

discrepantie die ontstaat doordat een cognitief gebied bij een kind met epilepsie relatief

sterker is aangedaan. Bestaande onderzoeken spreken elkaar tegen als het gaat om welke

schaalscore het meest verlaagd is bij kinderen met epilepsie, zelfs binnen één enkel

epilepsiesyndroom zijn er tegenstrijdige resultaten te zien. Er is weinig bekend over de

verschillen in cognitieve profielen tussen kinderen met epilepsie en kinderen en andere

stoornissen. Minder nog is er bekend over cognitieve profielen van kinderen met epilepsie

en een bijkomende stoornis, zogenaamde comorbiditeiten bij epilepsie.

Het derde type variabiliteit heeft betrekking op de mogelijke verschuivingen in IQ

die zich af kunnen spelen bij een kind met epilepsie door de tijd heen. Kinderen met

epilepsie hebben vaak gedurende verscheidene of zelfs vele jaren aanvallen. Ze moeten

zich ontwikkelen – doorgroeien – terwijl de epilepsie de hersenactiviteit met

onregelmatige tussenpozen verstoort. Over de mate waarin er in de loop der tijd

veranderingen optreden bij de verschillende cognitieve vaardigheden, of er verschillen

zijn in ontwikkelingsbeloop tussen de verschillende vaardigheden, en welke kenmerken

van epilepsie daarbij een rol spelen, is nog niet veel bekend.

SAMENVATTING

145

De belangrijkste vragen van deze verhandeling hebben betrekking op cognitieve patronen

van intra-individuele variabiliteit. Tonen profielen van kinderen met epilepsie meer

variabiliteit in testscores dan men op grond van de psychometrische eigenschappen van

de test zou verwachten? Gaat epilepsie samen met hoge scores op de ene taak of op het

ene meetmoment en lage op een andere taak of een ander meetmoment? Verandert dit

patroon in de loop van de tijd? Indien er sprake is van verhoogde intra-individuele

variabiliteit, van specifieke patronen, of van patronen van verandering door de tijd heen,

kunnen er variabelen worden geïdentificeerd die samenhangen met deze patronen?

De groep waarbij deze vragen werden bestudeerd bestond uit kinderen met epilepsie bij

wie vragen over hun cognitieve ontwikkeling waren gerezen. Ze waren aangemeld voor

neuropsychologisch onderzoek bij een tertiair centrum voor epilepsie, of de eraan

verbonden school voor kinderen met epilepsie. De school voor epilepsie biedt onderwijs

of onderwijsondersteuning aan kinderen met epilepsie. Niet alleen op de eigen school,

maar ook op iedere andere school voor regulier of speciaal primair of voortgezet

onderwijs in de noordelijke helft van het land. Ter vergelijking met de kinderen met

epilepsie, zijn in enkele studies ook kinderen met andere ontwikkelingsstoornissen

(leerstoornissen en gedragsstoornissen) opgenomen, evenals kinderen uit het

basisonderwijs bij wie er geen reden was tot zorg.

Bij alle kinderen werden WISC-gegevens verzameld. Het betrof daarbij de twee

meest recente versies van de Nederlandstalige WISC, de WISC-RNL of de WISC-IIINL, soms

ook de kleuterintelligentietest WPPSI-RNL (het superscript werd op alle Nederlandstalige

testversies toegepast). Bij kinderen met epilepsie werden daarnaast ook gegevens

verzameld die gerelateerd zijn aan de aard en de ernst van de epilepsie. Deze gegevens

kwamen van de medische of neuropsychologische rapporten en hadden betrekking op de

leeftijd bij aanvang van de epilepsie, de aard van de epilepsieaanvallen (focaal dan wel

generaliseerd), de lateralisatie van de epilepsie (uitgaande van de linker dan wel rechter

hersenhemisfeer), de topografische lokalisatie van de epilepsie (bijvoorbeeld frontaal,

temporaal), de aanwezigheid van hersenlaesies zichtbaar bij beeldvormend onderzoek

(MRI), het aantal typen medicijnen dat in de loop van de tijd was ingenomen. Aan de

hand van deze gegevens konden andere maten worden bepaald, zoals de duur van de

epilepsie op het moment dat de test werd afgenomen en de ernst van het


146

epilepsiesyndroom. Ook werden gegevens omtrent de deelname aan het speciaal

onderwijs en het opleidingsniveau van de ouders verzameld.

Na de algemene inleiding werd in Hoofdstuk 2 onderzocht of er bij kinderen met

epilepsie sprake is van toegenomen intra-individuele subtestvariabiliteit (subtest scatter).

In de steekproef waren 467 kinderen opgenomen, van wie 157 epilepsie hadden, 132

gedrags- of psychiatrische stoornissen, en 178 leerstoornissen. Door ook kinderen met

andere stoornissen op te nemen kon bekeken worden of grotere variabiliteit een kenmerk

kon zijn voor ontwikkelingsstoornissen. Het resultaat van het onderzoek was dat

verhoogde variabiliteit, hoewel inderdaad in enige mate aanwezig, niet als een algeheel

kenmerk van kinderen met ontwikkelingsstoornissen kon worden beschouwd. Er waren

verschillen te zien afhankelijk van de onderzochte groep of subgroep. Kinderen met

epilepsie, evenals kinderen met leerstoornissen, lieten geen verhoogde subtestvariabiliteit

zien. Kinderen met gedragsstoornissen (en onder hen, vooral de kinderen met autisme

spectrum problematiek) toonden wel verhoogde variabiliteit. Ook bij subgroepen van

kinderen met epilepsie was het beeld genuanceerder. Kinderen met linker hemisfeer

epilepsie toonden verhoogde variabiliteit, met andere woorden een grillig profiel, terwijl

dit bij kinderen met epilepsie startende in de rechter hemisfeer niet te zien was.

In Hoofdstuk 3 werd nader ingegaan op de invloed van de lateralisatie en

aanwezigheid van hersenlasesies op subtestvariabiliteit. De onderzoeksgroep bestond uit

90 kinderen met gelateraliseerde epilepsie. Van deze kinderen hadden 56 linkszijdige

epilepsie, van hen hadden 22 tevens een op het MRI zichtbare hersenlaesie. Vierendertig

kinderen hadden rechtszijdige epilepsie; van hen had 15 een MRI-laesie. Ook nu werd de

subtest scatter onderzocht. Uit dit onderzoek kwam naar voren dat er een differentieel

effect was op variabiliteit, afhankelijk van de zijde en de aanwezigheid van een leasie.

Kinderen met laesies op een MRI en epilepsie uitgaande van de linker hemisfeer

vertoonden verhoogde variabiliteit; terwijl kinderen met laesies op een MRI en epilepsie

uitgaande van de rechter hemisfeer verminderde variabiliteit vertoonden. In de algemene

discussie werd gespeculeerd dat reorganisatie van hersenfuncties bij kinderen met linker

hemisfeer en hersenlesies mogelijk ten koste gaat van een verhoogde grilligheid in het

subtestpatroon.

In Hoofdstuk 4 werd bij kinderen met epilepsie bekeken hoe vaak er sprake was

van een klinisch betekenisvolle verandering in IQ. De groep bestond uit 73 kinderen die,

met een tijdsinterval van 2,3 jaar, twee keer met dezelfde WISC versie werden onderzocht.

SAMENVATTING

147

Het betrof twee keer de WISC-RNL, danwel de WISC-IIINL. Een verandering wordt als

klinisch betekenisvol begrepen wanneer deze zeldzaam is en bij een vergelijkingsgroep in

slechts 10% van de kinderen voorkomt, bij 5% als een stijging en bij 5% als een daling

van het IQ. Als referentie voor de berekening van de kritieke waarden werd gebruik

gemaakt van een Nederlandstalige vergelijkingsgroep uit de literatuur. De

referentiegegevens hadden betrekking op kinderen met ontwikkelingsproblemen zonder

epilepsie, die op vergelijkbare wijze twee keer waren getest, eveneens met een

tijdsinterval van ruim twee jaar. Bij deze vergelijking toonden we aan dat er bij kinderen

met epilepsie drie keer zo vaak een betekenisvolle daling te zien was op totale schaal

(namelijk bij 16.4% van de kinderen), en vijf keer zo vaak (26%) op het verbale schaal.

Op de performale schaal werden in deze tussenliggende periode van ruim twee jaar geen

veranderingen gezien. Er waren ook kinderen die een stijging van het IQ lieten zien; deze

stijgingen kwamen echter niet vaker voor dan de verwachte 5%.

De resultaten van Hoofdstuk 4 gaven aan dat er bij kinderen met epilepsie een

verhoogd risico is op een daling van het verbale en totale IQ ten aanzien van kinderen met

andere ontwikkelingsproblemen. In Hoofdstuk 5 werd onderzocht, welke variabelen

bijdragen tot IQ-dalingen. De steekproef bestond uit 113 kinderen met epilepsie die twee

of drie keer waren onderzocht met de WPPSI-RNL, de WISC-RNL of de WISC-IIINL. Ook nu

werd een daling in IQ gevonden. Epilepsie gerelateerde variabelen, zoals de zijde van de

epilepsie, de lokalisatie van de epilepsie, de aard van de epilepsieaanvallen, het aantal

verschillende medicijnen dat in de loop der tijd was ingenomen, de aanwezigheid van

laesies in het brein, de ernst van het epilepsiesyndroom, of het feit of de aanvallen onder

controle waren bij hertest met de WISC (al dan geen aanvalsvrijheid bij de laatste test),

leverden geen van alle een statistisch betekenisvolle bijdrage aan de daling in IQ. Dit

resultaat was opmerkelijk, maar kon wel worden geïnterpreteerd binnen het licht van

recente studies die aantoonden dat er bij epilepsie veranderingen plaatsvinden in

netwerken van het brein, ook ver van de plek waar de aanval ontstaat, en dat die

veranderingen blijven voortbestaan ook nadat de aanvallen onder controle zijn.

Opvallend was dat een lage score al te zien was bij de eerste testafname,

voornamelijk op de performale schaal, terwijl de verbale schaal bij de eerste meting

betrekkelijk “gespaard” leek te blijven. Door de tijd heen veranderde dit beeld evenwel.

De verbale schaal daalde in een versneld tempo, de performale schaal daalde ook maar

minder sterk, en na verloop van tijd was de relatieve voorsprong van de verbale schaal ten


148

aanzien van de performale niet meer te zien. De variabele duur van de epilepsie bleek

daarmee een belangrijke variabele te zijn; de daling in de tijd had een logaritmische vorm.

Naast de variabele duur van de epilepsie, bleek ook een tweede aan tijd

gerelateerde variabele een bijdrage te leveren aan het IQ-patroon en de daling van de

verbale en de performale schaal. Het ging daarbij om de leeftijd van de aanvang van de

epilepsie. Een jongere leeftijd was gerelateerd aan een relatief hogere beginscore op de

verbale schaal, en een relatief sterkere daling van de verbale schaal door de tijd heen.

Bij dit onderzoek werd tevens gekeken naar de bijdrage van onderwijstype

(regulier of speciaal onderwijs) en van het opleidingsniveau van de ouders, aan het

intelligentieprofiel van het kind. Zoals verwacht, hadden kinderen die geplaatst waren in

het speciaal onderwijs, lagere IQ-scores dan kinderen in het reguliere onderwijs. Plaatsing

in het speciaal onderwijs had evenwel geen relatie met de daling van het IQ. Kinderen

van ouders met een hogere opleiding scoorden hoger; maar ook hier bleek dat een hoog

opleidingsniveau geen “beschermende” factor was tegen een daling van IQ. Anders

gezegd: zowel op het reguliere onderwijs als op het speciaal onderwijs konden dalingen

in IQ worden gezien, evenzeer bij kinderen van hoog opgeleide ouders als bij die van

minder hoog opgeleide ouders.

In Hoofdstuk 6 werden de discrepanties tussen de verbale en de performale

schaal opnieuw bekeken, nu in het licht van de aanwezigheid van comorbide stoornissen.

Een comorbiditeit is een tweede gediagnosticeerde aandoening, waarvan er een relatie

met de eerste aandoening wordt verondersteld, omdat de aandoening in combinatie vaker

voorkomt dan men op basis van toeval zou verwachten. Twee steekproeven werden in

deze studie opgenomen. De eerste steekproef bestond uit 117 kinderen die slechts één

diagnose hadden. Deze diagnose betrof epilepsie, specifieke leesstoornissen, specifieke

rekenstoornissen, dan wel autisme spectrum stoornissen. Deze kinderen werden met de

term “geïsoleerde” stoornissen aangeduid, hoewel ze natuurlijk ook verdere

neuropsychologische problemen konden hebben, die echter niet tot een tweede diagnose

hadden geleid. Het profiel dat de verbale en de performale schalen te zien gaven werd

vergeleken. Uit dit deelonderzoek kwam naar voren dat kinderen met epilepsie een profiel

van relatief betere verbale dan performale vaardigheden lieten zien (VIQ > PIQ), dat niet

getoond werd door de andere groepen. De andere groepen hadden een vlak profiel (zoals

ook controle kinderen zonder problemen) of een profiel waarbij de verbale vaardigheden

zwakker waren dan de performale (vooral de kinderen met leesstoornissen). In die zin

leek het VIQ > PIQ profiel vrij specifiek te zijn voor kinderen met epilepsie. De tweede

SAMENVATTING

149

steekproef bestond uit 171 kinderen met epilepsie. Ook hier waren kinderen opgenomen

met geïsoleerde epilepsie, maar ook kinderen met epilepsie en cormorbide

leesstoornissen, rekenstoornissen of autisme spectrum stoornissen. Opnieuw werden de

patronen van kinderen met geïsoleerde epilepsie vergeleken met de patronen van de

kinderen met leer- of gedragsstoornissen, nu als comorbide stoornissen. Hieruit kwam

naar voren, dat de voor kinderen met geïsoleerde epilepsie kenmerkende VIQ > PIQ

discrepantie, was “afgevlakt” als het kind naast epilepsie een leesstoornis had. De

discrepantie was versterkt te zien als het kind naast epilepsie een rekenstoornis had. Er

werden geen verschillen gezien tussen geïsoleerde epilepsie en epilepsie met autisme.

Geconcludeerd werd dat bij kinderen bij wie twee stoornissen bij elkaar kwamen –

epilepsie en nog een tweede stoornis – het voor epilepsie kenmerkende patroon van

verbale en performale vaardigheden veranderd was. Omgekeerd, gegeven de stoornis

zonder epilepsie, kon gezegd worden dat de “impact” die de epilepsie op het cognitieve

profiel had, bij alle beelden vergelijkbaar leek, namelijk een “systematische

verschuiving” in de richting van een lager performaal IQ danwel in de richting van een

meer gespaard verbaal IQ.

Samenvattend leidden deze resultaten tot een schets van de ontwikkeling van de

cognitieve vaardigheden van kinderen met epilepsie, die op deze wijze niet eerder in de

literatuur gemaakt was. Al gauw na de aanvang van de epilepsie, zijn cognitieve

achterstanden te zien. Deze zijn het best zichtbaar op de performale schaal. De verbale

schaal blijft op dat moment relatief gespaard. Een VIQ > PIQ patroon wordt zichtbaar,

waarbij de verbale score mogelijkerwijs een betere maat is van het oorspronkelijke

potentieel van het kind, en de performale score mogelijk een betere maat is van de

kwetsbare reactie van het brein op de epilepsie. Na verloop van tijd begint het IQ te dalen,

zowel het verbale als het performale IQ. De daling is het duidelijkst te zien in de eerste

(vijf) jaren na aanvang van de epilepsie, maar kan daarna nog lange tijd voortduren. De

daling is het sterkst op de verbale schaal, en na enige tijd wordt het verschil tussen de

verbale en de performale schaal niet meer gezien. De meeste kinderen laten bij een hertest

van ruim twee jaar een score zien die binnen de kritieke waarden valt. Het aandeel

kinderen, dat een klinisch zeldzame (en daarom als betekenisvol beschouwde) daling te

zien geeft op de verbale schaal of op de totale schaal, is evenwel verhoogd. Op de totale

schaal is deze drie keer zo hoog als de verwachte daling bij andere kinderen met

ontwikkelingsproblemen maar zonder epilepsie, en op de verbale schaal vijf keer zo hoog.


150

Bij het langer duren van de epilepsie worden ook frequent dalingen op de performale

schaal gezien.

Het profiel dat kenmerkend is voor epilepsie, VIQ > PIQ, wordt door twee

belangrijke factoren gemoduleerd. De eerste is de duur van de epilepsie. Naarmate de

epilepsie langer duurt, verdwijnt het voordeel van de verbale boven de performale schaal.

De tweede is de aanwezigheid van comorbide problemen. Bij kinderen met

leesstoornissen én epilepsie wordt het VIQ > PIQ niet gezien, maar is er een vlak profiel;

bij kinderen met rekenstoornissen én epilepsie wordt het VIQ > PIQ profiel juist vergroot.

De klinische relevantie van deze resultaten kan op verschillende niveaus worden

besproken. Allereerst is het voor ouders, leerkrachten en onderwijsbegeleiders van belang

om te weten dat er bij kinderen met epilepsie sprake kan zijn van dalingen in cognitieve

vaardigheden, die niet op elk deelgebied even groot zijn. Met name de dalingen in het

verbale IQ zullen consequenties hebben voor de schoolcarrière. Doublures, teruggezet

worden naar een lagere vorm van (voortgezet) onderwijs, plaatsing in het speciaal

onderwijs behoren allemaal tot de mogelijkheden waarvoor een kind met epilepsie een

verhoogd risico heeft, en waarin het met zorg begeleid moet worden. Aanpassingen in het

onderwijs (zoals preteaching) zouden zo vroeg mogelijk ingezet moeten worden.

Kinderneurologen zouden alert moeten zijn op informatie die duidt op stagnaties.

Stagnatie zou een aanleiding moeten zijn voor nader onderzoek naar de cognitieve

ontwikkeling. Ook kan het een aanleiding zijn om op zoek te gaan naar de achterliggende

etiologie.

In de tweede plaats kunnen de resultaten betekenisvol zijn voor beleidsmakers in

het onderwijs. De onderzoeken tonen aan dat het ontwikkelingsbeloop van kinderen met

epilepsie verschilt van dat van kinderen met andere ontwikkelingsstoornissen. Zowel het

epilepsie-specifieke cognitieve profiel alsook het beloop door de tijd heen (cognitieve

achterstanden kunnen verergeren, ook nadat de aanvallen wegblijven) vragen aangepaste

voorzieningen en expertise gericht op kinderen met epilepsie.

Voor onderzoekers op het gebied van epilepsie is het van belang dat variabelen als

leeftijd bij aanvang van epilepsie, duur van epilepsie en aanwezigheid van comorbide

stoornissen, die alle van invloed zijn op het cognitieve profiel, opgenomen worden in de

beschrijving van de onderzochte steekproeven. Afhankelijk van de lateralisatie van de

epilepsie en van de integriteit van het brein, zal meer of minder variabiliteit in de

testscores te zien zijn. Toegenomen variabiliteit brengt met zich mee dat bij gebruik van

SAMENVATTING

151

verkorte testversies de kans op meetfouten toeneemt. Verkorte versies kunnen het beste

vermeden worden. Indien ze toch gebruikt worden, zouden ze op zijn minst getoetst

moeten worden op bruikbaarheid voor de onderzochte groepen.

Niet in de laatste plaats kunnen de resultaten van de verschillende onderzoeken

ook van waarde zijn voor de (neuro)psycholoog en orthopedagoog die diagnostisch werk

verricht in de klinische setting. De informatie kan een bijdrage leveren tot de beschrijving

en interpretatie van testresultaten en de klinische besluitvorming. (1) Pas wanneer

subtestvariabiliteit 8 punten of meer op de verbale schaal beslaat, 10 of meer op de

performale, en 11 of meer op de totale schaal, kan er gesproken van zeldzaam grote

variabiliteit. Deze zeldzaam grote variabiliteit kan klinische betekenis hebben. (2) Pas

wanneer bij hertest de daling in (WISC-RNL, WISC-IIINL) IQ 14 punten of meer op de verbale

schaal bedraagt, 18 of meer op de performale, en 14 of meer op de totale schaal, kan er

gesproken worden van een zeldzame en klinische betekenisvolle verandering in IQ. Bij

tussentijdse verandering van test van de WISC-RNL naar de WISC-IIINL zijn deze waarden

voor verlies aan vaardigheden 19, 18 en 17 IQ-punten. (3) Significante verbaal –

performaal verschillen hoeven niet te worden beschouwd als een reden om het totaal IQ

achterwege te laten. In plaats daarvan kunnen ze worden geïnterpreteerd als klinisch

relevant. De aanwezigheid van een comorbide stoornis kan samengaan met een ander

patroon van vaardigheden. Het profiel van een kind met een leesstoornis of rekenstoornis

kan er anders uit komen te zien wanneer er ook epilepsie in het spel is. De aard van het

verschil is telkens in de richting van een minder gespaarde performale schaal, of een

relatief beter behouden verbale schaal. (4) Tot slot worden in de Appendices “base rate”

tabellen weergegeven, gebaseerd op klinische data van verscheidene Nederlandstalige

WISCs. Deze base rate tabellen geven de frequentie van voorkomen van bepaalde groottes

van subtestvariabiliteit, discrepanties tussen schalen en factoren, en verschillen in IQ bij

een herstest zoals in de klinische setting waargenomen bij ruim duizend Nederlandse

kinderen.

RESUMEN

153

Resumen

(Spanish Summary)

Patrones cognitivos en epilepsia de niños con epilepsia.

Variabilidad intra-individual, perfiles cognitivos y patrones de cambio en niños con

epilepsia en las Escalas de Inteligencia Wechsler para niños

Este trabajo reúne cinco estudios sobre los perfiles cognitivos de niños y niñas con

epilepsia.

El origen del trabajo es clínico. El estudio diagnóstico clínico con niños con

distintos tipos de trastornos del desarrollo indicaba que los perfiles cognitivos de niños

con epilepsia se diferenciaban de los perfiles que se encontraban en otros desórdenes.

Este tipo de observaciones, que se han llamado “la base de datos clínica” (Baron, 2005)

del neuropsicólogo clínico, cobra importancia cuando se logra confirmar con una “base

de datos empírica”. El objetivo del presente estudio era, pues, describir los perfiles

cognitivos de niños con epilepsia evaluados con el instrumento de mayor uso en la

evaluación neuropsicológica de niños: las Escalas de Inteligencia de Wechsler para Niños

(el test de WISC).

Ya se ha estudiado ampliamente que niños con epilepsia corren el riesgo de tener

problemas cognitivos. Estudios existentes también incluyen datos sobre las habilidades

verbales y no verbales (llamadas de ejecución) del niño con epilepsia. Las habilidades

verbales y de ejecución son, tradicionalmente, las principales habilidades evaluadas por

los tests de inteligencia. Abreviadas como VIQ y PIQ en la literatura internacional (y

como cociente intelectual [CI] verbal y CI de ejecución en la de lengua española), forman

conjuntamente la escala total (FS-IQ o bien CI total).

Las habilidades verbales y de ejecución se evalúan de forma estandarizada,

escrupulosamente normada e internacionalmente aceptada con las varias escalas del

WISC. Las escalas verbal y de ejecución son escalas independientes, pero altamente

correlacionadas, sugiriendo que ambas hacen referencia a un constructo común, la

inteligencia general. Igualmente, las (sub)pruebas son independientes, pero se relacionan

entre sí y se relacionan con un constructo, por ejemplo cinco subpruebas verbales

conforman la escala verbal.


154

Las habilidades verbales y de ejecución también forman el núcleo del presente

trabajo. Aunque se han estudiado bastante ya, se sabe relativamente menos sobre los

perfiles cognitivos en niños con epilepsia. El término perfil cognitivo – valores altos en

una medida y bajos en otra – se refiere a medidas de variabilidad dentro de un mismo

individuo.

Un tipo de variabilidad se refiere a la variabilidad intra-individual entre las

subpruebas, denominada subtest scatter en la literatura inglesa. En la literatura – también

fuera del ámbito de la epilepsia – aún se debate si la variabilidad entre subpruebas se ve

incrementada en trastornos, y, por ende, si se puede interpretar como algún tipo de

patología, o, si debe considerarse meramente una manifestación de las cualidades

psicométricas de la prueba.

La discusión se torna importante cuando se observa que psicólogos clínicos, por

ejemplo en el área de epilepsia en Holanda, indican que el perfil de un niño muestra una

alta variabilidad y que por ello la escala no es una buena medida de las habilidades del

examinado. Prosiguen a omitir el CI en el informe. Esto llama la atención porque para las

pruebas neerlandesas se desconoce cuáles son los valores críticos para poder hablar de

“variabilidad alta”, y si esa alta variabilidad tiene algún valor para el diagnóstico clínico.

Un segundo tipo de variabilidad en el perfil se refiere a la discrepancia entre la

escala verbal y la de ejecución (VIQ – PIQ). Una discrepancia alta entre la escala verbal

y de ejecución indicaría que un área se ve más comprometida que la otra. Los estudios

existentes – incluso dentro de un mismo síndrome epiléptico – se contradicen. Es poco lo

que se sabe sobre las diferencias entre los patrones cognitivos de niños con epilepsia y

niños con otros trastornos del desarrollo. Aún menos se sabe del perfil cognitivo del niño

que se ve afectado por dos condiciones a la vez, es decir, niños con comorbilidades de la

epilepsia.

Un tercer tipo de variabilidad se relaciona con los cambios en el curso del tiempo,

que pueden ocurrir en niños con epilepsia. Los niños con epilepsia suelen tener crisis

epilépticas durante un periodo prolongado. Durante este periodo, se espera de ellos que se

desarrollen, mientras que las crisis intermitentes interfieren con el funcionamiento

cognitivo. Los cambios a lo largo del tiempo que se puedan producir en las escalas de

inteligencia se desconocen en su mayoría. Igualmente, se ignora si los posibles cambios

que se puedan dar en el curso del tiempo difieren entre la escala verbal y la de ejecución.

Y de ser así, cuáles variables afectan estos cambios.

RESUMEN

155

La principal cuestión a tratar en los diversos estudios es el perfil cognitivo en

niños con epilepsia. Los temas a investigar se relacionan con la variabilidad dentro de un

mismo individuo: (1) la variabilidad entre las subpruebas, (2) la discrepancia entre la

escala verbal y la de ejecución (VIQ – PIQ), (3) las diferencias entre la primera

evaluación y la segunda en niños evaluados más de una vez.

En caso de encontrar incrementos en la variabilidad dentro del mismo individuo,

en la forma de perfiles o perfiles de cambio, surge la pregunta de si se pueden identificar

variables que se asocian con tales perfiles o perfiles de cambio.

Estos temas se estudiaron en muestras grandes de niños/as con epilepsia evaluados

porque existía preocupación sobre su desarrollo cognitivo. Se habían presentado a un

instituto de epilepsia neerlandés o a la escuela para niños con epilepsia asociada con el

instituto. Éste atiende a personas con epilepsia a nivel terciario y cubre la mitad

septentrional de los Países Bajos. La escuela atiende a niños con epilepsia tanto dentro de

su propio plantel como en cualquier escuela regular o especial, primaria o secundaria en

la mitad norte del país, siempre y cuando haya una indicación para ello. Además, a modo

de comparación, en algunos estudios se incluyeron muestras de niños con trastornos del

desarrollo de otra índole, tales como trastornos específicos de aprendizaje y trastornos

psiquiátricos y de conducta, al igual que niños sin trastornos.

Para todos los niños, se recogieron datos de las escalas del WISC neerlandesas

más recientes, el WISC-RNL y el WISC-IIINL, y en algunos casos también el WPPSI-RNL para

preescolares (el volado NL se aplicó a todas las pruebas neerlandesas). Para los niños con

epilepsia, se recolectaron datos adicionales de informes médicos y neuropsicológicos.

Estos se refieren a variables de epilepsia, tales como la edad del inicio de la epilepsia, el

tipo de crisis (parciales o generalizadas), la lateralidad (actividad epiléptica con origen en

el hemisferio izquierdo, LH, o derecho, RH), la localización (por ejemplo, del lóbulo

frontal o temporal), el número de medicamentos tomados en el curso del tiempo, la

presencia de lesiones cerebrales detectadas por resonancia magnética (RM). Con base en

estos datos, se extrajo información adicional tal como la duración de la epilepsia o la

severidad del síndrome epiléptico. Para los niños descritos en el Capítulo 6. Se recogieron

datos sobre comorbilidades. También se recogieron datos sobre participación en

educación regular o especial y datos sobre el nivel educativo de los padres.


156

Luego de la Introducción general, en el Capítulo 2, y con base en 467 niños, se

estudió el tema de la variabilidad intra-indvidual entre las subpruebas. El estudio se basó

en 157 niños con epilepsia, 132 con trastornos del aprendizaje, y 178 con trastornos

psiquiátricos. Encontramos que la variabilidad se encuentra sólo levemente elevada en los

grupos clínicos en general, y en ese sentido no aparece como un indicador general de

patología, sino que depende del grupo diagnóstico estudiado. Los niños con epilepsia, al

igual que los niños con problemas de aprendizaje, no presentan una variabilidad elevada.

Los niños con trastornos psiquiátricos, y entre ellos ante todo los niños con trastornos en

el espectro autista, por otro lado, sí muestran variabilidad elevada. Dentro del grupo con

epilepsia, se percibió un incremento de la variabilidad en epilepsia de lateralidad

izquierda, pero no en epilepsia de lateralidad derecha.

El Capítulo 3 estudió la variabilidad intra-individual entre subpruebas en relación

con lesiones cerebrales. El estudio incluyó 90 niños con epilepsia lateralizada. De ellos,

56 niños tenían epilepsia que emanaba del hemisferio izquierdo (de ellos 22 tenían

lesiones en RM), y 34 tenían epilepsia del hemisferio derecho (15 con lesiones en RM).

Se encontró que la variabilidad entre las subpruebas se encuentra incrementada en

epilepsia de lateralidad izquierda, en especial en casos con lesiones cerebrales detectadas

en imágenes por resonancia magnética, mientras que en casos de lateralidad derecha y

lesiones, la variabilidad se encuentra disminuida. En la Discusión general se especula que

una posible reorganización cerebral, en el caso de lesiones del hemisferio izquierdo, se

produzca a favor de la conservación de las habilidades verbales pero a costa de un

incremento de la variabilidad.

En el Capítulo 4 se estableció el porcentaje de niños que presentaban cambios

confiables de inteligencia en la segunda evaluación con la misma prueba (bien fuera el

WISC-RNL o el WISC-IIINL) en 73 niños con epilepsia. Cambios confiables se definieron

como aquellos que se presentan en sólo 10% de los niños con trastornos de desarrollo

(pero sin epilepsia), en el 5% como incrementos en el cociente intelectual y en el 5%

como pérdidas (deterioro) del cociente intelectual. Encontramos que, en la escala verbal,

el porcentaje de niños con pérdidas era del 26, en la escala total, del 16.4, es decir cinco

veces más que lo esperado en la escala verbal y tres veces más en la total. En la escala de

ejecución, los cambios no eran diferentes a los esperados. Igualmente, algunos niños

presentaban incrementos en el cociente intelectual; estos porcentajes nunca superaron los

esperados.

RESUMEN

157

Dados estos resultados, en el Capítulo 5 estudiamos las variables que podían estar

influyendo en el deterioro de las habilidades cognitivas de los niños con epilepsia. Con

base en 113 niños evaluados dos o tres veces con las escalas de Wechsler, no se encontró

una contribución significativa para la mayoría de las variables de epilepsia. Así mismo,

no se encontró una contribución para la lateralidad, la localización, el tipo de crisis, el

número de fármacos antiepilépticos tomados en el curso de la epilepsia, la presencia de

lesiones en RM, la severidad del síndrome epiléptico, ni la presencia de crisis en la última

evaluación (libre o no libre de crisis). Tampoco se halló una contribución significativa

para la interacción de las variables. Estos resultados fueron interpretados a la luz de

estudios que han demostrado cambios cerebrales duraderos – por ejemplo en la

conectividad cerebral – en áreas distantes a aquellas en las que se origina la actividad

epiléptica; cambios que perduran aunque la epilepsia haya entrado en remisión.

El patrón de declive encontrado en el Capítulo 5 fue llamativo. Al comienzo, se ve

que la escala verbal supera la escala de ejecución. Luego, se aprecia un cambio que no se

describe con una curva lineal sino logarítmica. En los primeros años se ve un deterioro

pronunciado, posteriormente, el ritmo se desacelera, pero el descenso continúa por un

tiempo prolongado. Los cambios se ven en ambas escalas, pero es más pronunciado en la

escala verbal, por lo que, con el transcurso del tiempo, la ventaja inicial de la escala

verbal sobre la de ejecución tiende a desaparecer. El perfil cambia en el curso del tiempo

en función de la duración de la epilepsia.

Además de la variable “duración de la epilepsia”, se encontró otra, también

asociada con el factor tiempo, que contribuye al perfil cognitivo: la edad de inicio de la

epilepsia. Un inicio temprano se asocia con habilidades verbales inicialmente mayores,

una mayor discrepancia VIQ > PIQ, y un mayor declive posterior. Cabe anotarse que se

vio una gran variación entre los niños en términos de perfiles y perfiles de cambio, lo que

significa que si bien estos resultados describen el grupo, un caso individual puede mostrar

un patrón de cambio diferente.

Otras variables, tales como la participación en enseñanza regular o especial, y el

nivel socioeconómico (medido a partir del nivel educativo de los padres) se vieron

relacionados con el nivel cognitivo, pero no con los cambios a través del tiempo. Es decir,

los niños en educación especial tienen un nivel más bajo, pero no muestran mayor

deterioro; igualmente, los niños de nivel socioeconómico alto tienen un CI más alto, pero

el nivel socioeconómico no los “protege” del deterioro.


158

El perfil VIQ > PIQ, aunque característico en niños con epilepsia, no se percibe en

todos los niños con epilepsia. En el Capítulo 6, la discrepancia entre la escala verbal y de

ejecución se estudió en dos muestras de niños con trastornos aislados y con

comorbilidades. La primera muestra incluyó 117 niños con trastornos “aislados”, es decir

niños con un solo diagnóstico (aunque podían tener otros problemas neuropsicológicos

que no condujeron a un segundo diagnóstico). Se incluyeron niños con epilepsia, con

trastornos de lectura, trastornos de matemáticas, y trastornos en el espectro autista. Se vio

que, en los niños con epilepsia, el perfil VIQ > PIQ (las habilidades verbales superaban

las de ejecución) era específico – los demás niños no mostraban este perfil. Los restantes

niños (y los del grupo control) mostraban un perfil plano o bien un perfil opuesto, VIQ <

PIQ, principalmente en trastornos de lectura. Evaluamos una segunda muestra de 171

niños, todos con epilepsia, pero algunos de ellos con doble diagnóstico de epilepsia y

comorbilidades (nuevamente: trastornos de lectura, de matemáticas, autismo).

Encontramos que los perfiles habían cambiado, y que eran parcialmente similares y

parcialmente diferentes a perfiles de los trastornos aislados. El impacto de este cambio era

similar en todos los trastornos. Donde se conjugan epilepsia y otro diagnóstico, se ve un

relativo deterioro de la escala de ejecución, mientras que la escala verbal se ve

relativamente resguardada.

En resumen, los estudios permiten urdir una descripción no conocida anteriormente sobre

el desarrollo del perfil cognitivo de niños con epilepsia a través del tiempo. Inicialmente,

el perfil VIQ > PIQ sugiere que las habilidades verbales se conservan y las de ejecución

se ven afectadas. Este perfil nos lleva a sugerir que la escala verbal es un indicador del

potencial original del niño con epilepsia, mientras que la escala de ejecución refleja la

vulnerabilidad del cerebro ante el trastorno epiléptico. Es posible que esta baja inicial de

la escala de ejecución sea un marcador cognitivo y sería interesante estudiar si ya se

percibe ante la inminencia de la epilepsia, anterior a su presentación. A través de los

primeros años, el nivel cognitivo desciende, más en los niños con epilepsia temprana que

en aquellos con epilepsia tardía, y ante todo en la escala verbal. A los dos años y medio,

un cuarto de los niños referidos ha sufrido un cambio significativo en la escala verbal.

Con el curso del tiempo, el deterioro se torna más lento, pero continúa a lo largo de los

años; es independiente de la persistencia de crisis, e incluye también cambios

significativos en la escala de ejecución. La presencia de comorbilidades influye en el

perfil en el sentido de que VIQ > PIQ cambia, modulado por el segundo trastorno. En

RESUMEN

159

epilepsia y trastornos de lectura, el perfil se torna plano; en epilepsia y trastornos de

matemáticas, el perfil VIQ > PIQ se hace más marcado aún, mientras en epilepsia y

autismo, no hay diferencias significativas entre los perfiles.

Al final del trabajo, los apéndices presentan “base rate tables”, que son tablas con los

porcentajes de niños que presentan diferencias de cierta magnitud en los WISCs. Las

tablas se basan en un total de más de mil niños neerlandeses, referidos por trastornos de

desarrollo (trastornos de aprendizaje, trastornos de conducta y psiquiátricos, y, ante todo,

epilepsia). También se incluyen datos de 88 niños control. Las tablas hacen relación a los

diferentes temas estudiados: (1) variabilidad entre las subpruebas o subtest scatter, (2)

discrepancias entre VIQ – PIQ y entre los factores y, (3) cambios de nivel cognitivo a lo

largo de varias evaluaciones neuropsicológicas, incluyendo cambios que se observan tras

el uso de una prueba de Wechsler diferente en la segunda evaluación, tal como el WISC-

RNL o WPPSI-IIINL seguido del WISC-IIINL. Las tablas incluyen datos de niños evaluados

posteriormente a la recolección de datos para los diferentes capítulos. En este sentido, las

tablas se relacionan a los capítulos pero a la vez son más extensas.

Appendices

APPENDICES

163

Appendices

It may be argued that summary statistics coming from the studies already include data on

individual children. Bringing these data to view, however, is considered of relevance for

the clinician. It is observed that Dutch clinicians often refrain from reporting IQs when

the discrepancy between the IQ scales is significant. Significant differences between the

verbal and the performance abilities, rather than being a reason for excluding the values

from the reports, may have importance for clinical and remediation purposes, and should

be understood rather than avoided. Similarly, subtest variability may be a relevant

descriptive of the child. It is also observed that IQs are excluded from reports when

subtest scatter appears increased while, in fact, in The Netherlands data providing the

clinician with psychometrically sound information on cut-offs or frequency of occurrence

are mostly lacking. Thus, the data of the Appendices aim at providing some of this

information.

Insights on the cognitive pattern and on magnitudes of cognitive change in

epilepsy may provide better understanding as to whether the pattern seen in a particular

child with epilepsy is common in children with epilepsy. Certainly, it should always be

borne in mind that an individual child may show a different pattern than the pattern

displayed frequently by other children. While the studies indicate the presence of distinct

cognitive patterns, of changing patterns over time and of changing patterns in

comorbidities, the studies also suggest that there is large variability between individuals

with epilepsy. This being said, the data analysed in the previous chapters allow for the

construction of tables, which may be of utility for the clinician.

In the Appendices A – D, the data are presented as base rate data. They relate to

the topics and participants discussed in the previous chapters. Efforts were done go

beyond these data and to report on larger samples than those discussed. Data collection on

epilepsy continued after the writing of the chapters, allowing for inclusion of more

participants than those presented in the chapters. For example, while the study on subtest

scaled-score range (Chapter 2) related to WISC-RNL only, in the base rate tables (Appendix

A), data on children with epilepsy and non-referred children tested with the WISC-IIINL are

included as well. Also, for the tables on VIQ – PIQ discrepancies (Appendix B), besides

the data from the participants discussed in Chapter 6, data from the participants of


164

Chapters 2 and 5, as well as newly collected data, were included as well. Data on

cognitive change at retesting were extended with data after a change of test version

(Appendix C). The vaster collection of WISC-IIINL data allowed for the construction of

tables on the factor index scores as well (Appendices C and D).

Appendix E provides the ROC curves which relate to the VIQ – PIQ and VCI –

POI data on children with isolated disorders versus comorbid disorders related in Chapter

6, as well as the rate of children showing significant discrepancies.

After testing a child with epilepsy, the clinician observed intra-individual subtest

variability and a VIQ – PIQ discrepancy of a particular magnitude. If case of a

reassessment, the clinician observed a difference between the first and the second test.

The clinician may want to know, how frequent these values were found in clinical or

standardization samples:

(1) Intra-individual subtest variability (subtest scatter) is the difference between

the highest and lowest subtest-scaled score in the verbal, performance and full

scales. How often is the observed scatter seen clinical comparison samples?

(Appendix A). Base rate data are given for the samples discussed in Chapter 2,

and additional WISC-IIINL data for children with epilepsy.

(2) The verbal – performance discrepancy is the directional VIQ – PIQ difference.

A positive value denotes VIQ > PIQ. A negative value denotes PIQ > VIQ.

How often does a directional difference between the verbal and performance

scale occur in clinical samples? In Appendix B, expected values are given, as

well as observed values from the samples discussed in Chapter 2, and

additional WISC-IIINL data on children with epilepsy. The discrepancy

between factor scores is establised pairwise (VCI – POI, VCI – PSI or POI –

PSI). How often does a specific difference between factor index scores occur?

In Appendix C, data are given on children with epilepsy tested with the WISC-

IIINL.

(3) A child has already been tested earlier (T1), and is tested for the second time

(T2), with an interval between the two tests of 12 months or more. For the

verbal, the performance and full scales (and factor index scores), T2 – T1 is

established. What percentage of children with epilepsy show this change of

APPENDICES

165

cognitive function? Data are provided for the verbal, performance and full

scales, as well as for the factor index scores (VCI, POI, PSI) in Appendix D.

Data are given for the children discussed in Chapter 4, additionaldata for a

larger sample tested with the WISC-IIINL, and small samples after changes in test

version (from WISC-RNL to WISC-IIINL; from WPPSI-IIINL to WISC-IIINL).


166

Appendix A

Subtest Scaled-Score Range (Subtest Scatter)

Subtest scaled-score range relates to the difference between the highest and the lowest

subtest score in a profile. For example, if the highest score on the verbal scale is 11 and

the lowest is 6, scatter is 5. The subtest scaled-score range of the principal data discussed

in Chapters 2 are presented in the following tables as base rate data. In addition to the

discussed samples, data on the WISC-IIINL in children with epilepsy are also provided.

Table A.1. gives the descriptives of the samples. Base rate data are displayed for the

verbal and performance scales in Table A.2 and for the full scale in Table A.3.

Expected means and (modified) standard deviations for subtest scaled score range were

established based on formula’s from Silverstein (1987) and tables from Owen (1962), and

applied as follows:

Meanscatter =

Where:

σ = 3 (SD of a WISC subtest)

ρ = mean intercorrelation of subtests (for 5 verbal subtests of the Dutch WISC-RNL = 0.55)

E(W) = value from Owen (1962, Table 6.2, p.140), for n = 5 (5 verbal subtests), = 2.326

SDscatter =

Where:

σ = 3 (SD of WISC subtest)

ρ = mean intercorrelation of subtests (for 5 verbal WISC-RNL = 0.55)

[σ2 E(W)] = a value from (1962, Table 6.2, p.140) for n = 5 (5 verbal subtests) = 0.747

The tables read as follows: on the verbal scale (A.2., upper panel), a scaled score range of

5 points or more was found on the WISC-IIINL in 54.5% of non-referred children and in

48.4% of the sample with epilepsy and close to normal IQ. On the full scale (A.3), scaled

score range of 8 points or more was found on the WISC-IIINL in 51.1% of non-referred

children and in 48.1% of the sample with epilepsy and close to normal IQ.

APP

END

IX A

Tabl

e A

.1. M

ixed

refe

rred

and

non

-ref

erre

d sa

mpl

es. C

hara

cter

istic

s of t

he sa

mpl

es.

Ex

pect

ed

N

ot re

ferr

ed

LD

Psyc

hiat

ry

Ep

ileps

y St

anda

rdiz

atio

n sa

mpl

es

Not

refe

rred

LD

Ps

ychi

atry

Ep

ileps

y Ep

ileps

y Ep

ileps

y Te

st W

ISC

-RN

L W

ISC

-III

NL

WIS

C-I

IIN

L W

ISC

-RN

L W

ISC

-RN

L W

ISC

-RN

L W

ISC

-III

W

ISC

-III

N

19

61

1239

88

13

2 17

8 15

7 22

1 63

Se

lect

ion

Stan

dard

izat

ion

sam

ples

75

< FS

-IQ

< 1

30

FS-I

Q >

75

FS-I

Q >

75

FS-I

Q >

75

FS-I

Q >

75

FS-I

Q ≤

75

Age

6:

0 to

16:

11

6:0

to 1

6:11

9.

0 (1

.8)

12.8

(1.3

) 10

.9 (2

.7)

9.7

(2.7

) 9.

9 (2

.9)

10.5

(2.7

) V

IQ

100.

0 (1

5.0)

10

0.0

(15.

0)

102.

5 (1

0.7)

93

.3 (1

0.8)

93

.8 (1

1.5)

95

.3 (1

2.1)

94

.4 (1

1.0)

67

.9 (1

2.4)

PI

Q

100.

0 (1

5.0)

10

0.0

(15.

0)

103.

6 (1

2.5)

97

.3 (1

2.3)

95

.7 (1

3.5)

91

.0 (1

1.9)

89

.3 (1

2.2)

64

.2 (1

1.6)

FS

IQ

100.

0 (1

5.0)

10

0.0

(15.

0)

103.

2 (1

0.6)

94

.6 (1

0.8)

93

.9 (1

0.7)

92

.5 (1

0.7)

91

.1 (1

0.7)

63

.6 (7

.6)

Ran

ge F

S-IQ

78

- 12

6 77

- 12

4 76

- 12

7 76

- 12

5 76

- 13

1 47

- 75

V

erba

l sca

tter

4.7

(1.8

) 4.

7 (1

.8)

4.8

(2.0

) 4.

8 (2

.0)

5.0

(2.0

) 5.

0 (2

.0)

4.6

(1.9

) 4.

3 (1

.7)

Perf

orm

ance

scat

ter

5.8

(2.1

) 5.

7 (2

.1)

6.3

(2.2

) 5.

8 (2

.3)

6.5

(2.6

) 6.

0 (2

.4)

5.4

(2.0

) 4.

5 (1

.7)

Full-

Scal

e sc

atte

r 7.

3 (1

.9)

7.4

(1.9

) 7.

6 (2

.1)

7.4

(2.1

) 8.

0 (2

.2)

7.7

(2.3

) 7.

0 (2

.1)

6.0

(2.2

) A

OE

5.6

(3.3

) 6.

0 (3

.3)

4.6

(3.5

) D

urat

ion

of e

pile

psy

4.0

(3.2

) 3.

9 (2

.9)

5.9

(3.7

)

Not

e. L

D =

spec

ific

lear

ning

dis

abili

ties.

PSY

: psy

chia

tric

diso

rder

s. EP

I = e

pile

psy.

AO

E =

age

at o

nset

of e

pile

psy.

Shad

ed =

sam

ples

dis

cuss

ed in

Cha

pter

167


168

Table A.2. Base rate tables referred and non-referred samples. Subtest Scaled-Score range

(subtest scatter) on 5 subtests of the verbal scale and on 5 subtests of the performance

scale.Cumulative percentages.

Verbal Scale Not referred LD Psychiatry Epilepsy

WISC-IIINL WISC-RNL WISC-RNL WISC-RNL WISC-III WISC-III Range FS-IQ > 75 FS-IQ > 75 FS-IQ > 75 FS-IQ > 75 FS-IQ ≤ 75 Range

13 13 12 12 11 2.3 0.8 0.6 0.5 11 10 3.4 2.3 1.1 2.5 1.4 10 9 5.3 6.2 6.4 3.2 1.6 9 8 8.0 8.3 12.9 15.3 8.1 6.3 8 7 20.5 17.4 25.3 21.0 16.3 9.5 7 6 27.3 31.1 36.0 25.0 28.1 17.5 6 5 54.5 53.0 51.1 54.1 48.4 42.9 5 4 72.7 74.2 77.5 75.2 68.8 66.7 4 3 93.2 90.9 90.4 91.1 85.1 90.5 3 2 97.7 97.0 87.8 98.7 98.6 95.2 2 1 100 99.2 100 100 100 100 1 0 100 0

Performance Scale Not referred LD Psychiatry Epilepsy


16 16 15 0.6 15 14 0.8 0.6 14 13 1.5 2.8 1.3 13 12 1.1 3.4 3.8 12 11 5.7 2.3 6.2 5.7 0.9 11 10 8.0 7.6 12.9 7.0 2.3 10 9 14.8 11.4 23.0 13.4 5.0 9 8 30.7 24.2 30.9 20.4 14.9 4.8 8 7 43.2 32.6 46.1 37.6 28.1 11.1 7 6 60.2 49.2 63.5 52.2 48.4 30.2 6 5 76.1 69.7 77.0 70.1 66.5 49.2 5 4 89.8 87.1 8.2 89.2 82.4 68.3 4 3 97.0 97.0 96.1 96.8 94.1 87.3 3 2 100 99.2 98.9 99.4 98.6 98.4 2 1 100 100 100 100 100 1 0 0

APPENDIX A

169

Table A.3. Base rate tables on referred and non-referred samples. Subtest scaled-score range (subtest scatter) on 10 subtests of the Full scale.

Full Scale Not referred LD Psychiatry Epilepsy


16 16 15 0.6 15 14 1.1 1.5 0.6 0.5 14 13 3.0 4.5 2.5 0.9 13 12 3.8 6.7 6.4 3.2 12 11 9.1 6.8 12.9 10.8 4.5 11 10 14.8 15.9 27.0 18.5 10.9 4.8 10 9 36.4 26.5 42.7 33.1 21.7 11.1 9 8 51.1 47.0 58.4 50.3 42.1 22.2 8 7 69.3 61.4 74.7 73.2 62.0 36.5 7 6 81.8 81.8 87.6 81.5 76.9 60.3 6 5 96.6 93.9 97.8 91.7 88.7 74.6 5 4 97.7 99.2 100 98.1 95.0 87.3 4 3 100 100 100 99.5 98.4 3 2 100 2 1 100 1 0 0


170

Appendix B

Base Rate Tables: Verbal – Performance Discrepancies (VIQ – PIQ)

Verbal – performance discrepancies relate to the VIQ – PIQ difference. If a child scores

VIQ 105 and PIQ 85, the 20-point difference favours the verbal scale. In appendix B,

base rate data for the verbal – performance discrepancies (VIQ – PIQ) are pesented. Table

B.1. gives the characteristics of the samples included. Table B.2. and B.3 display the base

rate data.The differences may favour the verbal scale (B.2.) or the performance scale

(B.3). The data presented are (a) expected rates for the WISC-RNL and the WISC-IIINL,

(b) observed data from the mixed samples of children with learning disabilitises,

psychiatric and behavioural disorders and epilepsy discussed in Chapter 2, and (c)

observed WISC-IIINL data for children with epilepsy (children included in Chapters 4 –

6).

Expected rates. Based on the overall correlation between the two scales, Sattler (1990,

p.819) provides the formula to calculate the expected proportions for different

magnitudes: , and therefore: ;

wherein discrepancy is the difference between VIQ and PIQ of interest, sd = SD is the

standard deviation of the test (SD =15 for the Wechsler tests), and r the correlation

between the two scales which is rV-P= .58 for the WISC-RNL (de Bruyn, Vandersteene, &

van Haasen, 1986, p144) and is rV-P= .56 for the WISC-IIINL (Wechsler, 2005, Table

3.12). For each z- value, the equivalent proportion from a standard normal distribution

was established. For the WISC-RNL, the thus calculated expected values and observed

values by Van Haasen et al. (1986, p 177) for the standardization group were found to be

virtually identical.

The tables read as follows: on the WISC-IIINL, a VIQ – PIQ discrepancy of 20 points or

more favouring the verbal scale is expected in 7.8% of the children of the standardization

sample, and is actually observed in 11.5% of children with epilepsy (Table B.2.). On the

WISC-IIINL, a VIQ – PIQ discrepancy of 20 points or more favouring the performance

scale is expected in 7.8% of the children of the standardization sample, and is actually

observed in 2.8% of children with epilepsy (Table B.3.).

APP

END

IX B

Tabl

e B

.1. C

hara

cter

istic

s of t

he sa

mpl

es u

sed

for T

able

s B.2

and

B.3

.

Sam

ple

Stan

dard

izat

ion

N

ot

refe

rred

LD

Ps

ychi

atric

Epile

psy

Test

W

ISC

-RN

L W

ISC

-IIIN

L W

ISC

-IIIN

L W

ISC

-RN

L W

ISC

-RN

L W

ISC

-RN

L W

ISC

-IIIN

L W

ISC

-IIIN

L N

19

61

1239

88

13

2 17

8 16

7 25

3 72

Se

lect

ion

stan

dard

izat

ion

FS-I

Q <

130

FS

-IQ

> 7

5 FS

-IQ

> 7

5 FS

-IQ

> 7

5 FS

-IQ

> 7

5 FS

-IQ

=<

75

Age

6:

0 to

16

:11

6:0

to

16:1

1 9.

0 (1

.8)

12.8

(1.3

) 10

.9 (2

.7)

9.5

(2.5

) 9.

7 (2

.8)

10.2

(2.6

)

VIQ

10

0.0

(15.

0)

100.

0 (1

5.0)

10

2.5

(10.

7)

93.3

(10.

8)

93.8

(11.

5)

95.8

(12.

3)

94.9

(11.

3)

69.5

(8.8

)

PIQ

10

0.0

(15.

0)

100.

0 (1

5.0)

10

3.6

(12.

5)

97.3

(12.

3)

95.7

(13.

5)

90.9

(12.

9)

90.3

(12.

8)

65.8

(8.7

)

FSIQ

10

0.0

(15.

0)

100.

0 (1

5.0)

10

3.2

(10.

6)

94.6

(10.

8)

93.9

(10.

7)

92.6

(11.

7)

92.0

(11.

2)

64.4

(7.5

) V

IQ-P

IQ

0.0

(15.

0)

0.0

(15.

0)

-1.1

(13.

6)

-4,0

5 (1

2.6)

-1

,88

(15.

7)

4.9

(13.

7)

4.4

(13.

2)

3.8

(11.

6)

Ran

ge F

S-IQ

78

- 12

6 77

- 12

4 76

- 12

7 76

- 12

5 76

- 13

1 47

- 75

A

ge a

t ons

et o

f epi

leps

y 5.

6 (3

.2)

5.9

(3.2

) 4.

6 (3

.4)

Dur

atio

n of

epi

leps

y

3.9

(3.3

) 3.

8 (2

.9)

5.7

(3.5

) N

ote.

LD

= m

ixed

sam

ple

lear

ning

dis

orde

rs. P

sych

iatri

c =

mix

ed sa

mpl

e ps

ychi

atric

dis

orde

rs.

171

C

OG

NIT

IVE

PATT

ERN

S IN

PA

EDIA

TRIC

EPI

LEPS

Y

Ta

ble

B.2

. Bas

e ra

te ta

ble

VIQ

> P

IQ.

Ver

bal -

Per

form

ance

disc

repa

ncy.

VIQ

- PI

Q f

avou

rs th

e V

erba

l Sca

le

Ex

pect

ed

N

ot re

ferr

ed

LD

Psyc

hiat

ry

Ep

ileps

y

VIQ

> P

IQ

WIS

C-R

NL

WIS

C-I

IIN

L W

ISC

-III

NL

WIS

C-R

NL

WIS

C-R

NL

WIS

C-R

NL

WIS

C-I

IIN

L V

IQ >

PIQ

FS

-IQ

> 7

5 FS

-IQ

> 7

5 FS

-IQ

> 7

5 FS

-IQ

> 7

5 FS

-IQ

=<

75

25

3.4

3.8

8.

4

4.2

25

24

4.0

4.4

9.6

5.9

24

23

4.7

5.2

5.7

4.5

12.0

6.

7 23

22

5.

5 5.

9 1.

5 15

.0

7.5

5.6

22

21

6.3

7.8

8.

0

2.3

5.

6

15.6

9.

5

21

20

7.4

7.8

3.

0

18.0

11

.5

6.9

20

19

8.4

8.9

10.2

6.

2 18

.6

13.0

8.

3 19

18

9.

5 10

.0

4.5

7.9

19.2

14

.6

11.1

18

17

10

.7

11.3

5.

3 9.

6 16

.6

12.5

17

16

12

.3

12.7

6.1

11

.8

23

.4

19.0

13

.9

16

15

13.8

14

.2

11.4

7.

6 14

.0

24.6

19

.8

20.8

15

14

15

.4

16.1

12

.5

9.1

15.7

26

.3

22.9

23

.6

14

13

17.1

17

.9

15.9

10

.6

16.3

28

.1

25.3

25

.0

13

12

19.2

19

.8

17.0

17

.4

29.9

27

.3

28.4

12

11

21

.2

21.8

18.2

20.8

32.9

29

.6

27.9

11

10

23

.3

23.9

21.6

12.9

23.0

37.7

33

.6

29.2

10

9

25.8

26

.1

14.4

24

.2

39.5

36

.4

33.3

9

8 28

.1

28.4

22

.7

17.4

27

.9

41.9

40

.3

36.1

8

7 30

.5

30.9

26

.1

18.2

30

.3

44.3

42

.7

41.7

7

6 33

.0

33.4

20.5

34.3

47.9

48

.6

45.8

6

5 35

.9

35.9

33.0

23.5

37.1

52.1

51

.8

48.6

5

4 38

.6

39.9

35

.2

25.0

40

.4

53.9

53

.4

50.0

4

3 41

.3

41.7

37

.5

27.3

44

.9

56.9

56

.1

54.2

3

2 44

.0

44.4

40

.9

31.8

45

.5

56.7

58

.9

2 1

47.2

47

.2

42.0

34

.8

47.2

60

.5

60.1

58

.3

1 0

50.0

50

.0

45

.5

37

.1

51

.1

63

.5

65.2

69

.4

0

172

APP

END

IX B

Ta

ble

B.3

. Bas

e ra

te ta

ble

VIQ

< P

IQ.

Ver

bal -

Per

form

ance

disc

repa

ncy.

VIQ

- PI

Q f

avou

rs th

e Pe

rfor

man

ce S

cale

Expe

cted

Not

refe

rred

LD

Ps

ychi

atry

Epile

psy

V

IQ -

PIQ

W

ISC

-RN

L W

ISC

-III

NL

WIS

C-I

IIN

L W

ISC

-RN

L W

ISC

-RN

L W

ISC

-RN

L W

ISC

-III

NL

VIQ

< P

IQ

FS-I

Q >

75

FS-I

Q >

75

FS-I

Q >

75

FS-I

Q >

75

FS-I

Q =

< 75

-25

3.4

3.8

9.1

8.4

-25

-24

4.0

4.4

9.8

9.0

1.8

1.6

-24

-23

4.7

5.2

10.6

10

.1

2.4

2.0

2.8

-23

-22

5.5

5.9

4.5

11.8

4.

2 -2

2 -2

1 6.

3 7.

8 5.

7 11

.4

13.5

2.

4 -2

1 -2

0 7.

4 7.

8

15.2

2.

8

-20

-19

8.4

8.9

12.9

4.

8 3.

2 -1

9 -1

8 9.

5 10

.0

8.0

13.6

5.

4 3.

6 -1

8 -1

7 10

.7

11.3

15

.9

19.1

6.

6 -1

7 -1

6 12

.3

12.7

12

.5

17.4

21

.3

4.7

4.2

-16

-15

13.8

14

.2

14

.8

22

.5

7.

8 6.

3

-15

-14

15.4

16

.1

17.0

21

.2

24.2

9.

0 6.

7 8.

3 -1

4 -1

3 17

.1

17.9

20

.5

25.3

9.

6 8.

7 -1

3 -1

2 19

.2

19.8

23

.9

22.0

26

.4

10.2

9.

7 -1

2 -1

1 21

.2

21.8

28.4

25.8

28.7

13.8

10

.3

11.1

-1

1 -1

0 23

.3

23.9

33.0

31.1

30.3

15.0

11

.1

12.5

-1

0 -9

25

.8

26.1

36

.4

32.6

33

.1

16.2

13

.0

13.9

-9

-8

28

.1

28.4

7.

5 34

.8

36.0

17

.4

15.0

16

.7

-8

-7

30.5

30

.9

38.6

37

.9

37.6

18

.6

17.8

19

.4

-7

-6

33.0

33

.4

40

.9

40

.9

39

.3

21

.0

21.3

22

.2

-6

-5

35.9

35

.9

44.3

45

.5

22.8

24

.1

23.6

-5

-4

38

.6

39.9

47

.7

49.2

42

.1

26.3

27

.3

25.0

-4

-3

41

.3

41.7

48

.9

53.8

46

.1

29.9

28

.9

-3

-2

44.0

44

.4

54.5

57

.6

46.6

34

.1

33.2

27

.8

-2

-1

47.2

47

.2

62.9

48

.9

36.5

34

.8

30.6

-1

0

50.0

50

.0

58

.0

65

.2

52

.8

39

.5

39.9

41

.7

0

173


174

Appendix C

Base Rate Tables for the Discrepancies Between Factor Index Scores Verbal

Comprehension (VCI), Perceptual Organization (POI) and Processing Speed (PSI):

(VCI – POI , VCI – PSI, POI – PSI)

Chapter 6 analyzed the VCI – POI discrepancies alongside with the VIQ – PIQ

discrepancies, and discussed the POI – PSI briefly as well. The Discussion highlighted

that, throughout the decades, the WISC series have changed from relying mostly on the

scales (VIQ and PIQ) to relying increasingly on the factor scores (VCI, POI, PSI). In the

following tables, empirical base rate data for non-referred, typically developing control

children and for children with epilepsy tested with the WISC-IIINL are provided for pairwise

comparisons of the factor index scores. The data are based on the participants discussed

in Chapters 4 to 6, as well as data collected later on. No expected values are included,

given that the Dutch WISC-IIINL test manual does not provide the correlations between the

factor scores.

Table C.1. gives the descriptives of the samples. Tables C.2. displays the base rate

data for VIQ – POI, Table C.3. for VIQ – PSI and Table C.4. for PSI – PSI.

For example, a child (FS-IQ 93) gained scores on the WISC-IIINL factors as follows:

VIQ 100, POI 90, PSI 80.The tables read: A VCI > POI difference of 10 points or more is

observed in 19.3% of the non-referred children and in 31.7% of the children with epilepsy

(Table C.2., left panel). A VCI > PSI difference of 20 or more points is observed in

14.8% and 16.7% of the non-referred children and children with epilepsy, respectively

(Table C.3., left panel). A POI > PSI difference of 10 or more points is observed in 22.7%

and 28.0% of the non-referred children and children with epilepsy, respectively (Table

C.4., right panel).

APPENDIX C

175

Table C.1. Characteristics of the samples used for Tables C.2 and C.3.

Sample Not referred Epilepsy

Test WISC-IIINL WISC-IIINL WISC-IIINL N 88 246 66 Selection FS-IQ < 130 FS-IQ > 75 FS-IQ =< 75 Age 9.1 (1.9) 10.1 (3.1) 10.4 (2.9) VCI 102.3 (11.5) 95.4 (11.2) 71.0 (9.2) POI 103.3 (12.3) 91.2 (12.8) 67.0 (10.0) PSI 104.7 (14.6) 91.6 (14.2) 72.8 (12.2) VCI - POI -1.0 (13.2) 4.2 (13.8 4.0 (12.9) VCI - PSI -2.4 (18.3) 3.8 (17.4) -1.6 (13.2) POI - PSI -1.4 (18.0) -0.3 (16.3) -5.6 (13.2) AOE - 6.1 (3.4) 4.5 (3.4) Duration of epilepsy - 4.0 (3.2) 5.9 (3.8)

Note. VCI = Verbal Comprehension Index. POI = Perceptual Organization Index, PSI = Processing Speed Index. VCI – POI: mean difference between VCI and POI. AOE = Age at onset of epilepsy.


176

Table C.2. Base rate tables VCI > POI and VCI < POI.

VCI - POI favours Verbal Comprehension VCI - POI favours Perceptual Organization Non referred Epilepsy Non referred Epilepsy WISC-IIINL WISC-IIINL WISC-IIINL WISC-IIINL VCI - POI 75 <FS-IQ > 130 FS-IQ >75 FS-IQ <= 75 VCI - POI 75 <FS-IQ > 130 FS-IQ > 75 FS-IQ <= 75

> 53 0.4 -54 0.4 Ζ Ζ 40 -40 0.8 39 1.1 1.2 -39 38 -38 37 2.0 -37 36 2.4 -36 35 -35 34 2.3 2.8 1.5 -34 33 3.3 -33 32 -32 31 3.4 -31 30 3.7 -30 1.2 29 4.1 -29 28 4.5 -28 27 4.9 -27 1.1 1.6 26 5.7 -26 25 3.0 -25 2.0 3.0 24 6.5 4.5 -24 23 .5 7.3 -23 2.4 22 5.7 8.5 6.1 -22 4.5 4.5 21 6.8 9.8 9.1 -21 6.1 20 8.0 10.6 13.6 -20 19 9.1 13.4 15.2 -19 6.8 3.3 7.6 18 12.5 15.9 16.7 -18 9.1 4.1 9.1 17 17.1 21.2 -17 11.4 5.3 16 19.1 22.7 -16 15 21.5 24.2 -15 13.6 6.9 10.6 14 13.6 22.8 -14 15.9 9.8 13 23.2 -13 18.2 10.2 12 15.9 24.8 -12 19.3 12.1 11 17.0 28.0 -11 20.5 10.6 15.2 10 19.3 31.7 27.3 -10 27.3 12.6 16.7 9 20.5 35.8 34.8 -9 29.5 13.4 18.2 8 22.7 38.6 40.9 -8 33.0 15.4 7 23.9 40.7 43.9 -7 39.8 17.5 19.7 6 28.4 43.1 45.5 -6 43.2 19.5 5 31.8 45.1 48.5 -5 44.3 23.2 4 33.0 49.6 53.0 -4 46.6 26.8 21.2 3 54.1 59.1 -3 51.1 30.1 22.7 2 36.4 55.7 62.1 -2 55.7 32.1 28.8 1 37.5 59.8 68.2 -1 36.2 30.3 0 44.3 63.8 69.7

APPENDIX C

177

Table C.3. Base rate tables VCI > PSI and VCI < PSI.

VCI -PSI favours Verbal Comprehension VCI -PSI favours Processing Speed Non referred Epilepsy Non referred Epilepsy WISC-IIINL WISC-IIINL WISC-IIINL WISC-IIINL VCI - PSI 75 <FS-IQ > 130 FS-IQ > 75 FS-IQ <= 75 VCI - PSI 75 <FS-IQ > 130 FS-IQ > 75 FS-IQ <= 75

>=53 0.8 -54 Ζ ζ 40 3.3 -40 39 -39 3.4 1.6 38 -38 4.5 37 -37 36 1.1 3.7 -36 2.0 35 4.5 -35 2.8 34 2.3 4.9 -34 3.0 33 4.5 -33 5.7 3.3 32 5.3 -32 31 5.7 -31 3.7 30 6.1 -30 9.1 29 6.9 -29 4.5 28 6.8 7.7 -28 27 8.0 8.1 -27 10.2 4.9 26 9.1 8.9 1.5 -26 4.5 25 9.3 -25 6.1 24 10.2 10.2 -24 5.7 23 11.4 13.4 3.0 -23 11.4 9.1 22 14.6 4.5 -22 12.5 6.1 21 13.6 15.4 -21 20 14.8 16.7 6.1 -20 7.3 19 17.1 -19 14.8 8.9 10.6 18 17.9 -18 15.9 9.3 12.1 17 18.7 7.6 -17 23.9 11.4 13.6 16 17.0 22.0 9.1 -16 26.1 13.4 15 19.3 24.8 -15 28.4 13.8 16.7 14 26.0 -14 30.7 14.6 18.2 13 23.9 28.9 12.1 -13 15.9 12 32.5 -12 17.1 11 34.6 15.2 -11 31.8 19.5 10 25.0 37.8 -10 35.2 21.1 9 26.1 41.1 21.2 -9 23.2 21.2 8 42.7 24.2 -8 37.5 25.6 25.8 7 28.4 43.9 25.8 -7 39.8 26.8 28.8 6 31.8 46.7 -6 40.9 27.6 31.8 5 33.0 49.6 27.3 -5 47.7 30.9 34.8 4 26.4 52.0 28.8 -4 52.3 33.3 37.9 3 37.5 54.5 40.9 -3 54.5 33.7 42.4 2 56.1 43.9 -2 58.0 36.6 48.5 1 38.6 58.9 47.0 -1 59.1 38.6 50.0 0 40.9 61.4 50.0


178

Table C.4. Base rate tables POI > PSI and POI > PSI.

POI - PSI favours Perceptual Organization POI - PSI favours Processing Speed Non referred Epilepsy Non referred Epilepsy WISC-IIINL WISC-IIINL WISC-IIINL WISC-IIINL POI - PSI 75 <FS-IQ > 130 FS-IQ > 75 FS-IQ <= 75 POI - PSI 75 <FS-IQ > 130 FS-IQ > 75 FS-IQ <= 75

47 1.1 0.4 -48 1.1 Ζ ζ 40 -40 1.2 1.5 39 -39 1.6 38 1.2 -38 37 -37 2.3 36 4.5 -36 35 -35 34 5.7 2.0 -34 33 6.8 -33 32 -32 3.4 31 2.4 -31 2.4 30 -30 29 2.8 -29 3.3 28 8.0 -28 4.5 4.1 27 4.5 -27 5.7 4.5 3.0 26 9.1 4.9 -26 6.8 6.1 25 5.7 1.5 -25 8.0 7.6 24 6.1 -24 9.1 6.9 9.1 23 8.9 -23 10.2 8.5 10.6 22 9.3 3.0 -22 9.3 12.1 21 10.2 4.5 -21 14.8 11.4 20 10.2 11.4 -20 15.9 13.8 19 11.4 13.0 -19 17.0 14,2 16.7 18 12.5 14.6 -18 15.4 19.7 17 13.6 15.4 -17 18.2 18.3 22.7 16 14.8 18.3 -16 19.3 19.9 28.8 15 15.9 19.1 -15 30.3 14 20.7 -14 21.6 22.4 31.8 13 18.2 21.5 -13 25.0 23.6 36.4 12 22.0 -12 26.1 26.0 37.9 11 20.5 25.6 10.6 -11 29.5 28.0 39.4 10 22.7 28.0 13.6 -10 31.8 28.9 9 23.9 30.1 16.7 -9 37.5 30.9 42.4 8 27.3 32.9 19.7 -8 38.6 31.7 43.9 7 29.5 34.6 -7 33.7 47.0 6 31.8 38.2 21.2 -6 45.5 39.0 50.0 5 41.1 22.7 -5 46.6 40.2 4 37.5 41.5 -4 47.7 42.3 54.5 3 43.5 27.3 -3 54.5 45.9 2 40.9 44.7 30.3 -2 47.6 57.6 1 43.2 46.7 33.3 -1 56.8 49.2 60.6 0 40.9 50.8 39.4

APPENDIX D

179

Appendix D

Base Rate Tables: From Test 1 to Test 2 - Cognitive Change After Serial Testing

Chapter 4 studied reliable cognitive change in children with epilepsy tested twice with the

same test version, either the WISC-RNL or the WISC-IIINL. In the following tables, the results

of cognitive change after serial testing will be presented. The data relate to children with

epilepsy tested twice with the Wechsler scales. The children may have been tested twice

with the same test version (T1 = T2, as in Chapter 4) or a change of test version may have

occurred (T1 ≠ T2, from WISC-RNL to WISC-IIINL or from WPPSI-IIINL to WISC-IIINL). Cut-offs

were calculated based on empirical data for referred children with developmental

disorders but without epilepsy (Schittekatte, 2005). No data on typically developing

children are provided. It should be noted that samples are relatively small, which means

that the results should be interpreted with caution.

Table D.1. gives the descriptives of the samples. Base rates data on cognitive

gains and losses are presented for the verbal scale in Table D.2, for the performance scale

in Table D.3., for the full scale in Table D.4. Table D.5 provides the descriptives for the

sample used for the factor index scores (VCI, POI, PSI). Base rate data in cognitive gains

and losses are presented in Table D.6 for the factor index scores.

For example, a child with epilepsy may have shown a VIQ 99 at first testing (T1)

with the WISC-IIINL, and a VIQ 85 at second testing, also with the WISC-IIINL. The tables

read as follows: After testing a child with epilepsy twice with the WISC-IIINL, a loss of 14

or more on the verbal scale would be expected in 5% of referred children (Table D.1.),

but was actually seen in 22.5% of the children (last column D.2.)


180

Table D.1. Characteristics of the samples used for Tables D.2. to D.4.

Epilepsy

Test T1 WISC-RNL/IIINL WPPSI-IIINL WISC-RNL WISC-IIINL Test T2 WISC-RNL/IIINL WISC-IIINL WISC-IIINL WISC-IIINL N 41 + 32 = 73 30 26 80 Test version T1 = T2 T1 ≠ T2 T1 ≠ T2 T1 = T2

T1 Age at onset of epilepsy 5.4 (3.0) 2.6 (1.8) 5.0 (1.9) 5.5 (2.8) Age 9.1 (2.2) 5.7 (0.6) 8.9 (2.0) 8.5 (2.1) Duration epilepsy at T1 3.7 (3.0) 3.1 (1.8) 3.9 (2.5) 3.0 (2.5) VIQ 90.1 (15.5) 85.8 (13.0) 88.4 (13.7) 91.2 (12.0) PIQ 87.5 (16.3) 81.4 (15.5) 80.7 ( 16.0) 86.1 (14.5) FSIQ 88.1 (15.9) 81.4 (13.7) 82.9 (13.7) 87.7 (13.2) VIQ-PIQ 3.4 (13.5) 4.4 (14.9) 7.7 (15.1) 5.2 (11.7)

T2 Age 11.4 (2.3) 8.7 (1.7) 13.3 (2.8) 11.1 (2.2) Duration epilepsy at T2 6.0 (3.1) 6.1 (2.7) 8.4 (3.4) 5.6 (2.7) VIQ 83.7 (16.3) 80.9 (15.8) 79.0 (14.5) 84.4 (12.7) PIQ 86.9 (17.8) 75.9 (13.4) 74.8 (14.5) 83.9 (15.4) FSIQ 83.7 (17.0) 76.3 (13.6) 74.2 (14.6) 82.4 (14.1) VIQ-PIQ -3.2 (13.7) 5.0 (15.4) 4.3 (12.7) 0.5 (10.7)

ΔT2-T1 Interval T1 T2 2.3 (1.2) 3.0 (1.7) 4.4 (2.2) 2.6 (1.3) VIQ -7.2 (11.3) -4.9 (12.9) -9.4 (8.2) -6.9 (9.2) PIQ -0.6 (11.9) -5.5 (12.8) -6.0 (13.4) -2.1 (9.8) FSIQ -4.4 (11.1) -5.1 (11.7) -8.7 (9.8) -5.3 (9.3) VIQ-PIQ -6.6 (11.4) 0.6 (12.6) -3.4 (12.6) -4.7 (9.0)

cut-off RCI gain / loss VIQ IQ points 14 / 14 N.A. 10 / 19 14 / 14 PIQ IQ points 18 / 18 N.A. 18 / 18 18 / 18 FS-IQ IQ points 14 / 14 N.A. 11 / 17 14 / 14

Note. T1 = test 1, T2 = test 2. ΔT2-T1 = difference T2 minus T1 . Shaded = discussed in Chapter 4. RCI = Reliable cognitive change cut-off scores for the 90% confidence interval. If the same test version (WISC-RNL or WISC-IIINL) is given at T1 and T2, on the verbal scale a 14 or more point gain or 14 or more point loss is expected in 5% of the children. If a change of test version has occurred (WISC-RNL at T1 and WISC-IIINL at T2), cut-off scores are adjusted. On the verbal scale, in 5% of the children a gain is expected of 10 or more points, in 5% a loss of 19 or more points. For the rationale and formula’s, see Chapter 4. Minimum time interval between T1 and T2 is 12 months. Age, age at onset, duration of epilepsy and interval are expressed in years.

APPENDIX D

181

Table D.2. Cognitive gains and cognitive losses on the verbal scale.

Cognitive gains Verbal Scale Cognitive losses Verbal Scale

Epilepsy Epilepsy

T1 WISC-

RNL/IIINL WPPSI-

IIINL WISC-

RNL WISC-IIINL T1

WISC-RNL/IIINL

WPPSI-IIINL

WISC- RNL

WISC-IIINL

T2 WISC-

RNL/IIINL WISC-IIINL WISC-IIINL

WISC-IIINL T2

WISC-RNL/IIINL WISC-IIINL

WISC-IIINL

WISC-IIINL

Δ T1 = T2 T1 ≠ T2 T1 ≠ T2 T1 = T2 Δ T1 = T2 T1 ≠ T2 T1 ≠ T2 T1 = T2 ≥30 ≤30 5.5 2.5 29 -29 3.8 28 -28 7.7 27 3.3 -27 26 -26 6.8 3.8 25 -25 5.0 24 -24 23 -23 6.7 7.5 22 -22 11.5 8.8 21 1.4 -21 8.2 13.3 20 -20 9.6 19 -19 11.0 11.3 18 -18 12.3 16.7 17 -17 13.7 20.0 12.5 16 -16 16.4 23.3 15 1.3 -15 20.5 15.4 17.5 14 6.7 -14 26.0 22.5 13 10.0 -13 28.8 31.3 12 2.7 13.3 -12 37.0 30.0 33.8 11 -11 34.6 10 4.1 -10 38.4 50.0 36.3 9 2.5 -9 41.1 33.3 53.8 37.5 8 23.3 -8 45.2 40.0 57.7 40.0 7 6.8 3.8 -7 53.4 50.0 61.5 53.8 6 8.2 5.0 -6 54.8 60.0 55.0 5 11.0 7.5 -5 65.4 57.5 4 13.7 10.0 -4 69.2 3 15.1 15.0 -3 61.6 84.6 63.7 2 21.9 7.7 22.5 -2 65.8 63.3 70.0 1 23.3 30.0 26.3 -1 69.9 66.7 92.3 72.5 0 76.7 70.0 73.8


182

Table D.3. Cognitive gains and cognitive losses on the performance scale.

Cognitive gains Performance Scale Cognitive losses Perfromance Scale Epilepsy Epilepsy

T1 WISC-

RNL/IIINL WPPSI-

IIINL WISC-

RNL WISC-IIINL T1

WISC-RNL/IIINL

WPPSI-IIINL

WISC- RNL

WISC-IIINL

T2 WISC-


WISC-IIINL T2


WISC-IIINL

WISC-IIINL

Δ T1 = T2 T1 ≠ T2 T1 ≠ T2 T1 = T2 Δ T1 = T2 T1 ≠ T2 T1 ≠ T2 T1 = T2

≥30 ≤30 29 1.4 -29 6.7 28 -28 7.7 27 -27 10.0 11.5 26 1.6 -26 25 -25 24 2.7 -24 4.1 15.4 1.3 23 2.5 -23 5.5 22 4.1 -22 2.5 21 -21 3.8 20 3.3 -20 19.2 19 5.5 3.8 -19 18 -18 13.3 17 6.8 -17 8.2 20.0 5.0 16 8.2 3.8 -16 9.6 6.3 15 -15 11.0 23.3 7.5 14 9.6 7.7 5.0 -14 12.3 26.7 30.8 11.3 13 11.0 -13 15.1 34.6 13.8 12 -12 17.5 11 13.7 11.5 7.5 -11 19.2 30.0 42.3 23.8 10 15.1 6,7 10.0 -10 20.5 46.2 9 16.4 10.0 17.5 -9 23.3 33.3 31.3 8 13.3 15.4 -8 50.0 32.5 7 20.5 16.7 23.1 18.8 -7 26.0 36.3 6 21.9 20.0 20.0 -6 27.4 46.7 53.8 37.5 5 26.0 26.9 26.3 -5 28.8 40.0 4 32.9 23.3 30.8 28.7 -4 31.5 53.3 57.7 43.8 3 37.0 26.7 34.6 30.0 -3 3.2 56.7 61.5 48.8 2 42.5 30.0 38.5 35.0 -2 60.0 52.5 1 53.4 33.3 37.5 -1 35.6 66.7 55.0 0 46.6 62.5

APPENDIX D

183

Table D.4. Cognitive gains and cognitive losses on the full scale.

Cognitive gains Full Scale Cognitive losses Full Scale Epilepsy Epilepsy

T1 WISC-

RNL/IIINL WPPSI-

IIINL WISC-

RNL WISC-IIINL T1

WISC-RNL/IIINL

WPPSI-IIINL

WISC- RNL

WISC-IIINL

T2 WISC-


WISC-IIINL T2


WISC-IIINL

WISC-IIINL

Δ T1 = T2 T1 ≠ T2 T1 ≠ T2 T1 = T2 Δ T1 = T2 T1 ≠ T2 T1 ≠ T2 T1 = T2

≥30 ≤30 4.1 1.3 29 -29 6.7 3.8 28 -28 27 -27 26 -26 7.7 25 -25 24 -24 11.5 23 -23 2.5 22 -22 3.8 21 1.3 -21 10.0 15.4 20 -20 13.3 5.0 19 -19 6.8 16.7 6.3 18 -18 8.2 20.0 23.1 8.8 17 1.4 -17 9.6 23.3 10.0 16 3.3 -16 12.3 26.9 12.5 15 -15 13.7 15.0 14 2.7 6.7 -14 16.4 17.5 13 4.1 2.5 -13 23.3 26.3 12 -12 26.0 26.7 30.8 28.7 11 5.5 5.0 -11 30.0 34.6 30.0 10 6.8 10.0 3.8 6.3 -10 27.4 42.3 31.3 9 8.2 13.3 -9 32.9 33.3 46.2 40.0 8 12.3 -8 35.6 57.7 42.5 7 16.7 7.5 -7 39.7 36.7 48.8 6 16.4 11.3 -6 41.1 43.3 52.5 5 20.5 23.3 7.7 17.5 -5 43.8 50.0 61.5 58.8 4 24.7 21.3 -4 46.8 53.3 3 28.8 22.5 -3 49.3 76.9 61.3 2 31.5 26.7 15.4 25.0 -2 54.8 63.3 62.5 1 32.9 36.7 28.7 -1 61.6 65.0 0 67.1 71.3


184

Table D.5. Cognitive Change on the Factor Index scores of the WISC-IIINL. Characteristics of the samples used for Table D.6.

Epilepsy

Test T1 WISC-IIINL

Test T2 WISC-IIINL N 60 Test version T1 = T2

T1 Age at onset of epilepsy 6.2 (2.9) Age 8.7 (2.2) Duration epilepsy at T1 2.5 (2.2) VCI 94.3 (12.3) POI 90.3 (15.1) PSI 87.3 (16.1)

T2 Age 11.4 (2.2) Duration epilepsy at T2 5.2 (2.3) VCI 87.7 (12.2) POI 86.7 (16.1) PSI 86.5 (14.7)

ΔT2-T1 Interval T1 T2 2.7 (1.2) VCI 6.5 (9.8) POI 3.6 (9.7) PSI 0.8 (12.5)

cut-off RCI gain / loss VCI IQ pointsa 14 / 14 POI IQ pointsa 18 / 18

PSI IQ pointsb 15 / 15 Note. Based on children tested twice with the WISC-IIINL for whom all factor index scores were available. a = estimated cut-off scores. The empirical values for the WISC-IIIUS (Canivez & Watkins, 1998), converted into cut-off scores (as in Chapter 4), lead to equal values for the verbal scale and verbal factor (VIQ and VCI) as well as for the performance scale and perceptual factor (PIQ and POI). Therefore, for the Dutch situation, cut-offs for VCI and POI are set equal to those calculated for VIQ and PIQ, respectively. b = estimated cut-off score for PSI. Given that the tasks for PSI are identical for the Dutch and American WISC-III, the cut-off estimated for the WISC-IIIUS is applied to the WISC-IIINL

APPENDIX D

185

Table D.6. Base Rate Table on Cognitive Gains (T1 > T2) (left) and Losses (right) on the

Factor Index scores Verbal Comprehension (VCI), Perceptual Organization (POI) and

Processing Speed (PSI)

Cognitive gains on Factor Indexes Cognitive loss on Factor Indexes Epilepsy Epilepsy

T1 WISC-IIINL T1 WISC-IIINL T2 WISC-IIINL T2 WISC-IIINL Δ T1 = T2 Δ T1 = T2 VCI POI PSI VCI POI PSI

42 1.7 -37 1.7 ζ ζ -33 1.7 ζ

30 3.3 -30 3.3 29 -29 28 -28 5.0 27 -27 6.7 26 -26 25 -25 24 1.7 -24 3.3 23 -23 1.7 22 -22 5.0 5.0 21 -21 20 3.3 -20 10.0 8.3 19 -19 18 5.0 -18 10.0 17 -17 6.7 11.7 16 1.7 6.7 -16 13.3 8.3 13.3 15 -15 18.3 10.0 15.0 14 10.0 -14 21.7 18.3 18.3 13 -13 23.3 12 5.0 -12 20.0 20.0 11 6.7 11.7 -11 25.0 25.0 10 5.0 8.3 -10 30.0 21.7 9 10.0 -9 33.3 30.0 23.3 8 13.3 16.7 -8 41.7 33.3 7 20.0 -7 45.0 46.7 25.0 6 8.3 15.0 26.7 -6 53.3 28.3 5 10.0 31.7 -5 55.0 51.7 33.3 4 13.3 21.7 -4 58.3 3 16.7 25.0 38.3 -3 61.7 55.0 38.3 2 28.3 41.7 -2 68.3 40.0 1 43.3 -1 71.7 56.7 0 83.3 71.7 56.7


186

Appendix E

Isolated Epilepsy, Isolated Developmental Disorders and Comorbidities in Epilepsy:

ROC Images for Chapter 6

The verbal – non-verbal discrepancies (VIQ – PIQ and VCI – POI) of the data in Chapter

6 were further analysed with Receiver Operating Characteristics (ROC).

ROC is a measure of diagnostic accuracy used to differentiate two samples from

each other. ROC uses all possible cut-off scores of the discrepancy (VIQ – PIQ or VCI –

POI) to compare two samples. Specificity (true positive scores, for example the group

with isolated epilepsy showing a specific VIQ – PIQ discrepancy) is contrasted to 1-

specificity (“false positive scores”, the control group showing this discrepancy). The Area

under the Curve (AUC) is calculated. AUC can take a value between .0 and 1.0, where

random scores appear close to .5. Scores between .5 and .7 denote a low discriminatory

accuracy, values between .7 and .9 denote moderate accuracy, and values above .9 denote

a high accuracy (Watkins, Glutting, & Youngstrom, 2005). A value of, for example, .640

for the contrast between non-referred controls and isolated epilepsy, can be read as

follows (Devena & Watkins, 2012): if a non-referred child is randomly selected from a

sample of non-referred children and a child with epilepsy is randomly selected from a

sample with epilepsy, the child with epilepsy would have a higher VIQ – PIQ (VIQ >

PIQ) difference about 64% of the time.

Isolated epilepsy was contrasted to control children (Figure E.1.) and isolated

epilepsy from Sample 1 was contrasted to isolated epilepsy from Sample 2 (Figure E.2.).

The isolated conditions were contrasted to the conditions comorbid with epilepsy:

isolated reading disorders versus comorbid reading disorders (Figure E.3), isolated math

disorders versus comorbid math disorders (Figure E.4), and isolated autism spectrum

disorders (ASD) versus comorbid ASD (Figure E.5.).

In addition, percentages of children showing “significant” directional VIQ – PIQ

discrepancies of 15 or more IQ points are presented in Table E. Based on the correlation

between the scales and the formula given by Sattler (1990, p.819); and a SDWechsler IQ =

15, the expected rate for a cut-off of 15 points was calculated: 13.8% for the WISC-RNL and

14.2% for the WISC-IIINL. The values are tested against the expected value (13.8% was

applied to all) with chi-square test.

APPENDIX E

187

Figure E.1.

Isolated epilepsy versus non-referred control

children.

Non-referred control children n = 82, isolated

epilepsy n = 139. VIQ – PIQ: AUC = .640, SE =

0.039, p = .001, [95%CI: .56, .72]. VCI – POI:

AUC = .637, SE = 0.042, p = .002, [95%CI: .55,

.72]

Figure E.2.

Isolated epilepsy.Two different samples.

Isolated epilepsy Sample 1, n = 39, WISC-RNL;

isolated epilepsy Sample 2, n = 100, WISC-IIINL.

VIQ – PIQ: AUC = .479, SE = 0.056, p = .708,

[95%CI: .37, .59]. VCI – POI: AUC = .483, SE =

0.057, p = .761, [95%CI: .37, .59]. Note equality

between the two samples.

Figure E.3.

Reading disorders. Isolated reading disorders

versus reading disorders comorbid with

epilepsy.

Sample 1, isolated reading disorders, n = 29;

Sample 2, reading disorders comorbid with

epilepsy, n = 31.VIQ – PIQ: AUC = .680, SE =

0.072, p = .017, [95%CI: .54, .82]. VCI – POI:

AUC = .651, SE = 0.072, p = .044, [95%CI: .51,

.79]


188

Figure E.4.

Math disorders. Isolated math disorders

versus math disorders comorbid with epilepsy.

Sample 1: Isolated math disorders, n = 25;

Sample 2, comorbid math disorders, n = 19. VIQ

– PIQ: AUC = .734, SE = 0.076, p = .009,

[95%CI: .58, .88]. VCI – POI: AUC = .733, SE =

0.085, p = .009, [95%CI: .59, .89]. Note the

moderate discriminatory value.

Figure E.5.

ASD: Isolated autism spectrum disorders

versus autism spectrum disorders comorbid

with epilepsy.

Sample 1: Isolated ASD, n = 24; Sample 2,

comorbid ASD n = 21. VIQ – PIQ: AUC = .589,

SE = 0.086, p = .306, [95%CI: .42, .76]. VCI –

POI: AUC = .531, SE = 0.088, p = .724, [95%CI:

.36, .70].

Table E. Rate of children showing VIQ – PIQ discrepancies of 15 or more points.

VIQ < PIQ VIQ > PIQ Sample % χ2 p % χ2 p Non-referred control 13.6 0.003 .954 12.3 .144 .704 Isolated epilepsy 5.0 8.975 .003 20.9 5.830 .016 Reading disorders (isolated) 27.6 4.633 .031 6.9 1.162 .281 Reading disorders (comorbid) 9.7 1.133 .287 6.4 2.385 .125 Math disorders (isolated) 8.0 0.707 .400 12.0 0.068 .794 Math disorders (comorbid) 0.0 - - 36.8 8.480 .004 ASD (isolated) 29.2 4.764 .029 20.8 0.998 .318 ASD comorbid) 4.8 1.442 .230 19.0 0.486 .486 Note. Chi-square testing against the expected value of 13.8%. Comorbid conditions = comorbid with epilepsy.

REFERENCES

189

References

Addis, L., Lin, J. J., Pal, D. K., Hermann, B., & Caplan, R. (2013). Imaging and genetics of language and cognition in pediatric epilepsy. Epilepsy Behav, 26(3), 303-312.

Aldenkamp, A. P., Alpherts, W. C., De Bruine-Seeder, D., & Dekker, M. J. (1990). Test-retest variability in children with epilepsy--a comparison of WISC-R profiles. Epilepsy Research, 7(2), 165-172.

Angold, A., Costello, E. J., & Erkanli, A. (1999). Comorbidity. Journal of Child Psychology and Psychiatry, 40(1), 57-87.

Asato, M. R., Caplan, R., & Hermann, B. P. (2014). Epilepsy and comorbidities -What are we waiting for? Epilepsy and Behavior, 31C, 127-128.

Augustijn, P. B. (2014). De nieuwe aanvals- en epilepsieclassificatie is niet beter dan de oude. Epilepsie, 12(3), 18 - 20.

Austin, J. K., & Fastenau, P. S. (2010). Are seizure variables related to cognitive and behavior problems? Developmental Medicine & Child Neurology, 52, 4-9.

Austin, J. K., Huberty, T. J., Huster, G. A., & Dunn, D. W. (1999). Does academic achievement in children with epilepsy change over time? Developmental Medicine & Child Neurology, 41(473-479).

Babikian, T., & Asarnow, R. (2009). Neurocognitive outcomes and recovery after pediatric TBI: Meta-analytic review of the literature. Neuropsychology, 23(3), 283-296.

Bailet, L. L., & Turk, W. R. (2000). The impact of childhood epilepsy on neurocognitive and behavioral performance: a prospective longitudinal study. Epilepsia, 41(4), 426-431.

Ballantyne, A. O., Spilkin, A. M., Hesselink, J., & Trauner, D. A. (2008). Plasticity in the developing brain: intellectual, language and academic functions in children with ischaemic perinatal stroke. Brain, 131(Pt 11), 2975-2985.

Baron, I. S. (2005). Test review: Wechsler Intelligence Scale for Children - Fourth Edition (WISC-IV). Child Neuropsychology, 11, 471 - 175.

Berg, A. T., Berkovic, S. F., Brodie, M. J., Buchhalter, J., Cross, J. H., van Emde Boas, W., . . . Scheffer, I. E. (2010). Revised terminology and concepts for organization of seizures and epilepsies: report of the ILAE Commission on Classification and Terminology, 2005-2009. Epilepsia, 51(4), 676-685.

Berg, A. T., Caplan, R., & Hesdorffer, D. C. (2011). Psychiatric and neurodevelopmental disorders in childhood-onset epilepsy. Epilepsy and Behavior, 20(3), 550-555.

Berg, A. T., & Cross, J. H. (2010). Towards a modern classification of the epilepsies? The Lancet Neurology, 9, 459-461.

Berg, A. T., Hesdorffer, D. C., & Zelko, F. A. (2011). Special education participation in children with epilepsy: what does it reflect? Epilepsy and Behavior, 22(2), 336-341.

Berg, A. T., Hesdorffer, D. C., & Zelko, F. A. (2011). Special education participation in children with epilepsy: What does it reflect? Epilepsy and Behavior, 22(336-341).

Berg, A. T., Langfitt, J. T., Testa, F. M., Levy, S. R., DiMario, F., Westerveld, M., & Kulas, J. (2008a). Global cognitive function in children with epilepsy: a community-based study. Epilepsia, 49(4), 608-614.

Berg, A. T., Langfitt, J. T., Testa, F. M., Levy, S. R., DiMario, F., Westerveld, M., & Kulas, J. (2008b). Residual cognitive effects of uncomplicated idiopathic and cryptogenic epilepsy. Epilepsy and Behavior, 13(4), 614-619.


190

Berg, A. T., & Plioplys, S. (2012). Epilepsy and autism: Is there as special relationship? Epilepsy and Behavior, 23, 193-198.

Berg, A. T., & Scheffer, I. E. (2011). New concepts in classification of the epilepsies: Entering the 21st century. Epilepsia, 52(6), 1058-1062.

Berg, A. T., Smith, S. N., Frobish, D., Beckerman, B., Levy, S. R., Testa, F. M., & Shinnar, S. (2004). Longitudinal assessment of adaptive behavior in infants and young children with newly diagnosed epilepsy: influences of etiology, syndrome, and seizure control. Pediatrics, 114(3), 645-650.

Berg, A. T., Testa, F. M., & Levy, S. R. (2011). Complete remission in nonsyndromic childhood-onset epilepsy. Annals of Neurology, 70(4), 566-573.

Berg, A. T., Zelko, F. A., Levy, S. R., & Testa, F. M. (2012). Age at onset of epilepsy, pharmacoresistance, and cognitive outcomes: a prospective cohort study. Neurology, 79(13), 1384-1391.

Besag, F. M. C. (2009). The relationship between epilepsy and autism: a continuing debate Acta Paediatrica, 98, 618 - 620.

Bhise, V. V., Burack, G. D., & Mandelbaum, D. E. (2009). Baseline cognition, behavior, and motor skills in children with new-onset, idiopathic epilepsy. Developmental Medicine & Child Neurology, 52(1), 22-26.

Bjornaes, H., Stabell, K., Henriksen, O., & Loyning, Y. (2001). The effects of refractory epilepsy on intellectual functioning in children and adults. A longitudinal study. Seizure, 10(4), 250-259.

Black, D. O., Wallace, G. L., Sokoloff, J. L., & Kenworthy, L. (2009). Brief report: IQ split predicts social symptoms and communication abilities in high-functioning children with autism spectrum disorders. Journal of Autism and Developmental Disorders, 39(11), 1613-1619.

Blackburn, L. B., Lee, G. P., Westerveld, M., Hempel, A., Park, Y. D., & Loring, D. W. (2007). The Verbal IQ/Performance IQ discrepancy as a sign of seizure focus laterality in pediatric patients with epilepsy. Epilepsy and Behavior, 10(1), 84-88.

Blackburn, L. B., Lee, G. P., Westerveld, M., Hempel, A., Park, Y. D., & Loring, D. W. (2007). The Verbal/Performance IQ discrepancy as a sign of seizure focus laterality in pediatric patients with epilepsy. Epilepsy and Behavior, 10, 84 - 88.

Boone, D. E. (1993). WAIS--R scatter with psychiatric inpatients: II. Intersubtest scatter. Psychol Rep, 73(3 Pt 1), 851-860.

Bourgeois, B. F., Prensky, A. L., Palkes, H. S., Talent, B. K., & Busch, S. G. (1983). Intelligence in epilepsy: a prospective study in children. Annals of Neurology, 14(4), 438-444.

Braakman, H. M., Ijff, D. M., Vaessen, M. J., Debeij-van Hall, M. H., Hofman, P. A., Backes, W. H., . . . Aldenkamp, A. P. (2012). Cognitive and behavioural findings in children with frontal lobe epilepsy. European Journal Paediatric Neurology, 16(6), 707-715.

Braakman, H. M., Vaessen, M. J., Jansen, J. F., Debeij-van Hall, M. H., de Louw, A., Hofman, P. A., . . . Backes, W. H. (2013). Frontal lobe connectivity and cognitive impairment in pediatric frontal lobe epilepsy. Epilepsia, 54(3), 446-454.

Brooks-Kayal, A. R., Bath, K. G., Berg, A. T., Galanopoulou, A. S., Holmes, G. L., Jensen, F. E., . . . Scharfman, H. E. (2013). Issues related to symptomatic and disease-modifying treatments affecting cognitive and neuropsychiatric comorbidities of epilepsy. Epilepsia, 54 Suppl 4, 44-60.

REFERENCES

191

Bulteau, C., Jambaque, I., Viguier, D., Kieffer, V., Dellatolas, G., & Dulac, O. (2000). Epileptic syndromes, cognitive assessment and school placement: a study of 251 children. Developmental Medicine and Child Neurology, 42(5), 319-327.

Camfield, C., Camfield, P., Gordon, K., Smith, B., & Dooley, J. (1993). Outcome of childhood epilepsy: a population-based study with a simple predictive scoring system for those treated with medication. The Journal of Pediatrics, 122(6), 861-868.

Camfield, C. S., & Camfield, P. R. (2007). Long-term socia outcomes for children with epilepsy. Epilepsia, 48(suppl.9), 3-5.

Campiglia, M., Seegmuller, C., Le Gall, D., Fournet, N., Roulin, J. L., & Roy, A. (2014). Assessment of everyday executive functioning in children with frontal or temporal epilepsies. Epilepsy Behav, 39, 12-20.

Canivez, G. L., & Watkins, M. W. (1998). Long-term stability of the Wechsler Intelligence Scale for Children - Third edition. Psychological Assessment, 10(3), 285 - 291.

Caplan, R., Siddarth, P., Stahl, L., Lanphier, E., Vona, P., Gurbani, S., . . . Shields, W. D. (2008). Childhood absence epilepsy: behavioral, cognitive, and linguistic comorbidities. Epilepsia, 49(11), 1838-1846.

CBS. (2007). Centraal Bureau voor de Statistiek (Central Statistics Bureau). Standaard Onderwijs Indeling, Onderwijsniveaus van de SOI 2006. (Standard classification of education). http://www.cbs.nl/NR/rdonlyres/6091E080-373F-4A5E-B08C-706D544835FA/0/2006soiniveauindeling201112.pdf. Retrieved May 28 2012

Cederlund, M. (2004). One hundred males with Asperger syndrome: a clinical study of background and associated factors. Developmental Medicine & Child Neurology, 46, 652-660.

Charman, T., Pickles, A., Chandler, S., Loucas, T., & Baird, G. (2011). IQ in children with autism spectrum disorders: data from the Special Needs and Autism Project (SNAP). Psychological Medicine, 619 - 627.

Chelune, G. J. (2002). Making neuropsychological outcomes research consumer friendly: a commentary on Keith et al. (2002). Neuropsychology, 16(3), 422-425.

Chelune, G. J., Naugle, R. I., Lüders, H., Sedlak, J., & Awad, I. A. (1993). Individual change after epilepsy surgery: practice effects and base-rate information. Neuropsychology, 7(1), 41-52.

Clarke, D. F., Roberts, W., Daraksan, M., Dupuis, A., McCabe, J., Wood, H., . . . Weiss, S. K. (2005). The prevalence of autism spectrum disorder in children surveyed in an tertiary care epilepsy clinic. Epilepsia, 46(12), 1970-1979.

Clarke, T., Strug, L., Murphy, P., Bali, B., Carvalho, J., Foster, S., . . . Pal, D. (2007). High risk of reading disability and speech sound disorder in rolandic epilepsy families: case-control study. Epilepsia, 48(12), 2258-2265.

Cohen, J. (1988). Statistical Power Analyses for the Behavioral Sciences (2nd ed.). Mahwah: Lawrence Erlbaum.

Cormack, F., Cross, J. H., Isaacs, E., Harkness, W., Wright, I., Vargha-Khadem, F., & Baldeweg, T. (2007). The development of intellectual abilities in pediatric temporal lobe epilepsy. Epilepsia, 48(1), 201-204.

Covanis, A. (2012). Epileptic encephalopathies (including severe epilepsy syndromes). Epilepsia, 53((suppl.4)), 114-126.

Crawford, J. R., & Allan, K. M. (1996). WAIS-R subtest scatter: base-rate data from a healthy UK sample. British Journal of Clinical Psychology, 35 ( Pt 2), 235-247.

Cysique, L. A., Franklin, D., Jr., Abramson, I., Ellis, R. J., Letendre, S., Collier, A., . . . Group, H. (2011). Normative data and validation of a regression based summary


192

score for assessing meaningful neuropsychological change. Journal of Clinical and Experimental Neuropsychology, 33(5), 505-522.

de Bruin, E. I., Verheij, F., & Ferdinand, R. F. (2006). WISC-R subtest but no overall VIQ-PIQ difference in Dutch children with PDD-NOS. Journal of Abnormal Child Psychology, 34(2), 263-271.

de Bruyn, E. E. J., Vandersteene, G., & van Haasen, P. P. (1986). WISC-R. Wechsler Intelligence Scale for Children-Revised. Nederlandstalige uitgave. Verantwoording. Lisse, The Netherlands: Swets Test Services.

Deonna, T., & Roulet-Perez, E. (2005). Cognitive and Behavioural Disorders of Epileptic Origin in Children (Vol. 168). Cambridge: Mac Keith Press.

Desoete, A. (2008). Co-Morbidity in Mathematical Learning Disabilities: Rule or Exception? The Open Rehabilitation Journal, 1, 15-26.

Devena, S. E., & Watkins, M. W. (2012). Diagnostic Utility of WISC-IV General Abilities Index and Cognitive Proficiency Index. Difference Scores Among Children With ADHD. Journal of Applied School Psychology, 28, 133-154.

Devinsky, O., & Tarulli, A. (2002). Progressive cognitive and behavioral changes in epilepsy. In O. Devinsky & L. E. Westbrook (Eds.), Epilepsy and developmental disabilities (pp. 133-149). Woburn: Butterworth-Heinemann.

Dodrill, C. B. (2004). Neuropsychological effects of seizures. Epilepsy and Behavior, 5 Suppl 1, S21-24.

Donders, J., Elzinga, B., Kuipers, D., Helder, E., & Crawford, J. R. (2012). Development of an eight-subtest short form of the WISC-IV and evaluation of its clinical utility in children with traumatic brain injury. Child Neuropsychology.

Dunn, D. W., Buelow, J. M., Austin, J. K., Shinnar, S., & Perkins, S. M. (2004). Development of syndrome severity scores for pediatric epilepsy. Epilepsia, 45(6), 661-666.

Dunn, D. W., Johnson, C. S., Perkins, S. M., Fastenau, P. S., Byars, A. W., deGrauw, T. J., & Austin, J. K. (2010). Academic problems in children with seizures: relationships with neuropsychological functioning and family variables during the 3 years after onset. Epilepsy and Behavior, 19(3), 455-461.

Elger, C. E., Helmstaedter, C., & Kurthen, M. (2004). Chronic epilepsy and cognition. Lancet Neurology, 3(11), 663-672.

Ellenberg, J. H., Hirtz, D. G., & Nelson, K. B. (1986). Do seizures in children cause intellectual deterioration? New England Journal of Medicine, 314(17), 1085-1088.

Engel, J., Jr. (2006). ILAE classification of epilepsy syndromes. Epilepsy Research, 70 Suppl 1, S5-10.

Engel, J. J. (2006). ILAE classification of epilepsy syndromes. Epilepsy Research, 70, 5 - 10.

Fastenau, P. S., Johnson, C. S., Perkins, S. M., Byars, A. W., deGrauw, T. J., Austin, J. K., & Dunn, D. W. (2009). Neuropsychological status at seizure onset in children: risk factors for early cognitive deficits. Neurology, 73(7), 526-534.

Fastenau, P. S., Shen, J., Dunn, D. W., & Austin, J. K. (2008). Academic underachievement among children with epilepsy: proportion exceeding psychometric criteria for learning disability and associated risk factors. Journal of Learning Disabilities, 41, 195-207.

Fejerman, N., Caraballo, R., & Tenembaum, S. N. (2000). Atypical evolutions of benign localization-related epilepsies in children: are they predictable? Epilepsia, 41(4), 380-390.

Field, A. (2005). Discovering Statistics using SPSS (2nd ed.). London: Sage.

REFERENCES

193

Fisher, R. S., van Emde Boas, W., Blume, W., Elger, C., Genton, P., Lee, P., & Engel, J. (2005). Epileptic seizures and epilepsy: definitions proposed by the International League Against Epilepsy (ILAE) and the International Bureau for Epilepsy (IBE). Epilepsia, 46(4), 470-472.

Flanagan, D. P., & Kaufman, A. S. (2009). Essentials of WISC-IV assessment. (2nd ed.). Hoboken: Wiley.

Flynn, J. R. (2007). What is intelligence? Cambridge: Cambridge University Press. Funke, M. E., Moore, K., Orrison, W. W., Jr., & Lewine, J. D. (2011). The role of

magnetoencephalography in "nonlesional" epilepsy. Epilepsia, 52 Suppl 4, 10-14. Geerts, A., Arts, W. F., Stroink, H., Peeters, E., Brouwer, O., Peters, B., . . . van

Donselaar, C. (2010). Course and outcome of childhood epilepsy: a 15-year follow-up of the Dutch Study of Epilepsy in Childhood. Epilepsia, 51(7), 1189-1197.

Geurts, H. M., Grasman, R. P., Verte, S., Oosterlaan, J., Roeyers, H., van Kammen, S. M., & Sergeant, J. A. (2008). Intra-individual variability in ADHD, autism spectrum disorders and Tourette's syndrome. Neuropsychologia, 46(13), 3030-3041.

Glutting, J. J., Watkins, M. W., Konold, T. R., & McDermott, P. A. (2006). Distinctions without a difference: The Utility of Observed Versus Latent Factors From the WISC-IV in Estimating Reading and Math Achievement on the WIAT-II. The Journal of Special Education, 40(2), 103-114.

Gottlieb, L., Zelko, F. A., Kim, D. S., & Nordli, D. R. (2012). Cognitive proficiency in pediatric epilepsy. Epilepsy and Behavior, 23, 146-151.

Grégoire, J. (2013). Measuring components of intelligence: Mission Impossible? Journal of Psychoeducational Assessment, 31(2), 138-147.

Gülgönen, S., Demirbilek, V., Korkmaz, B., Dervent, A., & Townes, B. T. (2000). Neuropsychological functions in idiopathic occipital lobe epilepsy. Epilepsia, 41, 405-411.

Heaton, R. K., Temkin, N., Dikmen, S., Avitable, N., Taylor, M. J., Marcotte, T. D., & Grant, I. (2001). Detecting change: A comparison of three neuropsychological methods, using normal and clinical samples. Archives of clinical neuropsychology, 16(1), 75-91.

Helmstaedter, C., Aldenkamp, A. P., Baker, G. A., Mazarati, A., Ryvlin, P., & Sankar, R. (2014). Disentangling the relationship between epilepsy and its behavioral comorbidities - The need for prospective studies in new-onset epilepsies. Epilepsy and Behavior, 31C, 43-47.

Hermann, B., Jones, J., Jackson, D. C., & Seidenberg, M. (2012). Starting at the beginning: the neuropsychological status of children with new-onset epilepsies. Progress in Epileptic Disorders, 14(1), 12-21.

Hermann, B., Jones, J., Seth, R., Koehn, M., Becker, T., Fine, J., . . . Seidenberg, M. (2008). Growing up with epilepsy: a two-year investigation of cognitive development in children with new onset epilepsy. Epilepsia, 49(11), 1847-1858.

Hermann, B., Jones, J., Sheth, R., Dow, C., Koehn, M., & Seidenberg, M. (2006). Children with new-onset epilepsy: neuropsychological status and brain structure. Brain, 129(Pt 10), 2609-2619.

Hermann, B. P., Dabbs, K., Becker, T., Jones, J. E., Myers y Gutierrez, A., Wendt, G., . . . Seidenberg, M. (2010). Brain development in children with new onset epilepsy: a prospective controlled cohort investigation. Epilepsia, 51(10), 2038-2046.


194

Hermann, B. P., Jones, J. E., Jackson, D. C., & Seidenberg, M. (2012). Starting at the beginning: the neuropsychological status of children with new-onset epilepsies. Epileptic Disorders, 14(1), 12-21.

Hermann, B. P., Jones, J. E., Sheth, R., Koehn, M., Becker, T., Fine, J., . . . Seidenberg, M. (2008). Growing up with epilepsy: a two-year investigation of cognitive development in children with new onset epilepsy. Epilepsia, 49(11), 1847-1858.

Høie, B., Mykletun, A., Sommerfelt, K., Bjornaes, H., Skeidsvoll, H., & Waaler, P. E. (2005). Seizure-related factors and non-verbal intelligence in children with epilepsy. A population-based study from Western Norway. Seizure, 14(4), 223-231.

Høie, B., Mykletun, A., Waaler, P. E., Skeidsvoll, H., & Sommerfelt, K. (2006). Executive functions and seizure-related factors in children with epilepsy in Western Norway. Developmental Medicine and Child Neurology, 48(6), 519-525.

Hutchinson, E., Pulsipher, D., Dabbs, K., Myers y Gutierrez, A., Sheth, R., Jones, J., . . . Hermann, B. (2010). Children with new-onset epilepsy exhibit diffusion abnormalities in cerebral white matter in the absence of volumetric differences. Epilepsy Reseach, 88(2-3), 208-214.

I.L.A.E., (1981). Porposal for revised clinical and electrographical classification of epileptic seizures. Epilepsia, 22, 489-501.

I.L.A.E., (1989). Proposal for revised classification of epilepsies and epileptic syndromes. Epilepsia, 30, 389-399.

Jensen, F. E. (2011). Epilepsy as a spectrum disorder: Implications from novel clinical and basic neuroscience. Epilepsia, 52 Suppl 1, 1-6.

Jones, J. E., Siddarth, P., Gurbani, S., Shields, W. D., & Caplan, R. (2010). Cognition, academic achievement, language, and psychopathology in pediatric chronic epilepsy: Short-term outcomes. Epilepsy and Behavior, 18(3), 211-217.

Joseph, R. M., Tager-Flusberg, H., & Lord, C. (2002). Cognitive profiles and social-communicative functioning in children with autism spectrum disorder. Journal of Child Psychology and Psychiatry and Allied Disciplines, 43(6), 807-821.

Kanemura, H., Sano, F., Mizorogi, S., Aoyagi, K., Sugita, K., & Aihara, M. (2012). Duration of recognized fever in febrile seizure predicts later development of epilepsy. Pediatr Int, 54(4), 520-523.

Kanemura, H., Sano, F., Tando, T., Sugita, K., & Aihara, M. (2012). Repeated seizures induce prefrontal growth disturbance in frontal lobe epilepsy. Brain and Development, 34(3), 175-180.

Kass, R. E., & Raftery, A. E. (1995). Bayes Factors. Journal of the American Statistical Association, 90, 773-795.

Kaufman, A. S. (1976). A new approach to the interpretation of test scatter on the WISC-R. Journal of Learning Disabilities, 9, 160-168.

Kaufman, A. S. (1979). Intelligent testing with the WISC-R. New York: Wiley. Kaufman, A. S. (1994). Intelligent testing with the WISC-III. New York: Wiley. Kaufman, A. S. (2010). In What Way Are Apples and Oranges Alike?: A Critique of

Flynn’s Interpretation of the Flynn Effect. Journal of Psychoeducational Assessment, 28(5), 382-398.

Kaufman, A. S., & Lichtenberger, E. O. (2006). Assessing adolescent and adult intelligence (3rd ed.). New York: Wiley.

Klein, B., Levin, B. E., Duchowny, M. S., & Llabre, M. M. (2000). Cognitive outcome of children with epilepsy and malformations of cortical development. Neurology, 55(2), 230-235.

REFERENCES

195

Koepp, M. J., Thomas, R. H., Wandschneider, B., Berkovic, S. F., & Schmidt, D. (2014). Concepts and controversies of juvenile myoclonic epilepsy: still an enigmatic epilepsy. Expert Rev Neurother, 14(7), 819-831.

Kofler, M. J., Alderson, R. M., Raiker, J. S., Bolden, J., Sarver, D. E., & Rapport, M. D. (2014). Working memory and intraindividual variability as neurocognitive indicators in ADHD: examining competing model predictions. Neuropsychology, 28(3), 459-471.

Kwan, P., & Brodie, M. J. (2001). Neuropsychological effects of epilepsy and antiepileptic drugs. Lancet, 357(9251), 216-222.

Lange, N., Froimowitz, M. P., Bigler, E. D., Lainhart, J. E., & Brain Development Cooperative, G. (2010). Associations between IQ, total and regional brain volumes, and demography in a large normative sample of healthy children and adolescents. Developmental Neuropsychology, 35(3), 296-317.

Lee, E. S., Yeatman, J. D., Luna, B., & Feldman, H. M. (2011). Specific language and reading skills in school-aged children and adolescents are associated with prematurity after controlling for IQ. Neuropsychologia, 49(5), 906-913.

Lidzba, K., Staudt, M., Wilke, M., & Krageloh-Mann, I. (2006). Visuospatial deficits in patients with early left-hemispheric lesions and functional reorganization of language: consequence of lesion or reorganization? Neuropsychologia, 44(7), 1088-1094.

Liégeois, F., Connelly, A., Cross, J. H., Boyd, S. G., Gadian, D. G., Vargha-Khadem, F., & Baldeweg, T. (2004). Language reorganization in children with early-onset lesions of the left hemisphere: an fMRI study. Brain, 127(Pt 6), 1229-1236.

LIGA. (2013). Dutch League against Epilepsy. Vastgestelde Richtijnen Epilepsie. http://epilepsie.neurologie.nl/cmssite/index.php?pageid=7 Retrieved June 21st 2014

Lin, J. J., Mula, M., & Hermann, B. P. (2012). Uncovering the neurobehavioural comorbidities of epilepsy over the lifespan. Lancet, 380(9848), 1180-1192.

Lind, A., Korkman, M., Lehtonen, L., Lapinleimu, H., Parkkola, R., & Matomäki, L. (2011). Cognitive and neuropsychological outcomes at 5 years of age in preterm children born in the 2000s. Developmental Medicine & Child Neurology, 53(3), 256-262.

Lopes, A. F., Simoes, M. R., Monteiro, J. P., Fonseca, M. J., Martins, C., Ventosa, L., . . . Robalo, C. (2013). Intellectual functioning in children with epilepsy: Frontal lobe epilepsy, childhood absence epilepsy and benign epilepsy with centro-temporal spikes. Seizure, 22, 886-893

Lopes, R., Simoes, M. R., & Leal, A. J. (2014). Neuropsychological abnormalities in children with the Panayiotopoulos syndrome point to parietal lobe dysfunction. Epilepsy and Behavior, 31, 50-55.

Loring, D. W., Strauss, E., Hermann, B. P., Perrine, K., Trenerry, M. R., Barr, W. B., . . . Meador, K. J. (1999). Effects of anomalous language representation on neuropsychological performance in temporal lobe epilepsy. Neurology, 53(2), 260-264.

Maassen, G. H., Bossema, E., & Brand, N. (2009). Reliable change and practice effects: outcomes of various indices compared. Journal of Clinical and Experimental Neuropsychology, 31(3), 339-352.

Martin, R., Sawrie, S., Gilliam, F., Mackey, M., Faught, E., Knowlton, R., & Kuzniekcy, R. (2002). Determining reliable cognitive change after epilepsy surgery: development of reliable change indices and standardized regression-based change norms for the WMS-III and WAIS-III. Epilepsia, 43(12), 1551-1558.


196

Masur, D., Shinnar, S., Cnaan, A., Shinnar, R. C., Clark, P., Wang, J., . . . Glauser, T. A. (2013). Pretreatment cognitive deficits and treatment effects on attention in childhood absence epilepsy. Neurology, 81(18), 1572-1580.

Matarazzo, J. D., Daniel, M. H., Prifitera, A., & Herman, D. O. (1988). Inter-subtest scatter in the WAIS-R standardization sample. J Clin Psychol, 44, 940 - 950.

Matsuo, M., Maeda, T., Sasaki, K., Ishii, K., & Hamasaki, Y. (2010). Frequent association of autism spectrum disorder in patients with childhood onset epilepsy. Brain and Development, 32(9), 759-763.

Mayes, S. D., Calhoun, S. L., & Crowell, E. (1998). WISC-III profiles for children with and without learning disabilities. Psychology in the Schools, 35, 309 - 316.

McLean, J. E., Reynolds, C. R., & Kaufman, A. S. (1990). WAIS-R subtest scatter using the Profile Variability Index. Psychological Assessessment, 2, 289 - 292.

Meador, K. J., Baker, G. A., Browning, N., Cohen, M. J., Bromley, R. L., Clayton-Smith, J., . . . Loring, D. W. (2012). Effects of fetal antiepileptic drug exposure: outcomes at age 4.5 years. Neurology, 78(16), 1207-1214.

Meinardi, H., Aldenkamp, A. P., & Nunes, B. (1992). Mental deterioration at epilepsy onset: a hypothesis. Acta Neurochirurgica Suppl (Wien), 55, 68-71.

Miranda, C., & Smith, M. L. (2001). Predictors of Intelligence after Temporal Lobectomy in Children with Epilepsy. Epilepsy and Behavior, 2(1), 13-19.

Mitchell, W. G., Scheier, L. M., & Baker, S. A. (1994). Psychosocial, behavioral, and medical outcomes in children with epilepsy: a developmental risk factor model using longitudinal data. Pediatrics, 94(4 Pt 1), 471-477.

Moffitt, T. E., & Silva, P. A. (1987). WISC-R Verbal and Performance IQ Discrepancy in an unselected cohort: clinical significance and longitudinal stability. Journal of Consulting and Clinical Psychology, 55(5), 768-774.

Moreno Manso, J. M., M.E., G.-B., Blázquez Alonso, M., & Guerrero Barona, E. (2010). Pragmatic language development and educational style in neglected children. Children and Youth Services Review, 32, 1028-1034.

Neyens, L. G., Aldenkamp, A. P., & Meinardi, H. M. (1999). Prospective follow-up of intellectual development in children with a recent onset of epilepsy. Epilepsy Research, 34(2-3), 85-90.

Neyens, L. G. J., & Aldenkamp, A. P. (1996). Stability of cognitive measures in children of average ability. Child Neuropsychology, 2(3), 161-170.

Nicolai, J., Ebus, S., Biemans, D. P., Arends, J., Hendriksen, J., Vles, J. S., & Aldenkamp, A. P. (2012). The cognitive effects of interictal epileptiform EEG discharges and short nonconvulsive epileptic seizures. Epilepsia, 53(6), 1051-1059.

Noens, I. L., & van Berckelaer-Onnes, I. A. (2005). Captured by details: sense-making, language and communication in autism. Journal of Communcation Disorders, 38(2), 123-141.

Nolan, M. A., Redoblado, M. A., Lah, S., Sabaz, M., Lawson, J. A., Cunningham, A. M., . . . Bye, A. M. (2003). Intelligence in childhood epilepsy syndromes. Epilepsy Research, 53(1-2), 139-150.

Northcott, E., Connolly, A. M., Berroya, A., McIntyre, J., Christie, J., Taylor, A., . . . Bye, A. M. (2007). Memory and phonological awareness in children with Benign Rolandic Epilepsy compared to a matched control group. Epilepsy Research, 75(1), 57-62.

REFERENCES

197

O'Leary, S. D., Burns, T. G., & Borden, K. A. (2006). Performance of children with epilepsy and normal age-matched controls on the WISC-III. Child Neuropsychology, 12(3), 173-180.

Olson, H. E., Poduri, A., & Pearl, P. L. (2014). Genetic forms of epilepsies and other paroxysmal disorders. Seminars in Neurology, 34(3), 266-279.

Oostrom, K. J., van Teeseling, H., Smeets-Schouten, A., Peters, A. C., & Jennekens-Schinkel, A. (2005). Three to four years after diagnosis: cognition and behaviour in children with 'epilepsy only'. A prospective, controlled study. Brain, 128(Pt 7), 1546-1555.

Overvliet, G. M., Besseling, R. M., Jansen, J. F., van der Kruijs, S. J., Vles, J. S., Hofman, P. A., . . . Backes, W. H. (2013). Early onset of cortical thinning in children with rolandic epilepsy. Neuroimage Clinical, 2, 434-439.

Overvliet, G. M., Besseling, R. M., Vles, J. S., Hofman, P. A., van Hall, M. H., Backes, W. H., & Aldenkamp, A. P. (2011). Association between frequency of nocturnal epilepsy and language disturbance in children. Pediatric Neurology, 44(5), 333-339.

Owen, D. B. (1962). Handbook of Statistical Tables. Palo Alto: Addison-Wesley Pal, D. K. (2011). Epilepsy and neurodevelopmental disorders of language. Current

Opinion in Neurology, 24(2), 126-131. Pelletier, P. M., Ahmad, S. A., & Rourke, B. P. (2001). Classification rules for basic

phonological processing disabilities and nonverbal learning disabilities: formulation and external validity. Child Neuropsychology, 7(2), 84-98.

Pennington, B. F. (2006). From single to multiple deficit models of developmental disorders. Cognition, 101(2), 385-413.

Pesch, W., & Ponsioen, A. (2004). Flinterdunne en flagrante Flynn-effecten bij licht verstandelijk gehandicapte kinderen (Tiny and flagrant Flynn effects in mildly mentally retarded children). De Psycholoog, 29(2), 64 - 68.

Piccinelli, P., Beghi, E., Borgatti, R., Ferri, M., Giordano, L., Romeo, A., . . . Balottin, U. (2010). Neuropsychological and behavioural aspects in children and adolescents with idiopathic epilepsy at diagnosis and after 12 months of treatment. Seizure, 19(9), 540-546.

Pijl, I. J., & Pijl, S. J. (1998). Are pupils in special education too "special" for regular education? International Review of Education, 44(1), 5-20.

Rakhade, S. N., & Jensen, F. E. (2009). Epileptogenesis in the immature brain: emerging mechanisms. Nature Reviews Neurology, 5(7), 380-391.

Rathouz, P. J., Zhao, Q., Jones, J. E., Jackson, D. C., Hsu, D. A., Stafstrom, C. E., . . . Hermann, B. P. (2014a). Cognitive development in children with new onset epilepsy. Dev Med Child Neurol, 56(7), 635-641.

Rathouz, P. J., Zhao, Q., Jones, J. E., Jackson, D. C., Hsu, D. A., Stafstrom, C. E., . . . Hermann, B. P. (2014b). Cognitive development in children with new onset epilepsy. Developmental Medicine and Child Neurology.

Reckess, G. Z., Varvaris, M., Gordon, B., & Schretlen, D. J. (2014). Within-person distributions of neuropsychological test scores as a function of dementia severity. Neuropsychology, 28(2), 254-260.

Reijs, R. P., van Mil, S. G., Arends, J. B., van Hall, M. H., Weber, J. W., Renier, W. O., & Aldenkamp, A. P. (2007). Cryptogenic localization related epilepsy in children from a tertiary outpatient clinic: is neurological and neuropsychological outcome predictable? Clinical Neurology and Neurosurgery, 109(5), 422-430.

Reijs, R. P., van Mil, S. G., van Hall, M. H., Arends, J. B., Weber, J. W., Renier, W. O., & Aldenkamp, A. P. (2006). Cryptogenic localization-related epilepsy with


198

childhood onset: The problem of definition and prognosis. Epilepsy and Behavior, 8(4), 693-702.

Reilly, C., & Neville, B. G. (2011). Academic achievement in children with epilepsy: a review. Epilepsy Research, 97(1-2), 112-123.

Reilly, C. J. (2011). Attention deficit hyperactivity disorder (ADHD) in childhood epilepsy. Reseach in Developmental Disabilities, 32(3), 883-893.

Resing, W. C. M., Evers, A., Koomen, H. M. Y., Pameijer, N. K., Bleichrodt, N., & van Boxtel, H. (2002). Indicatiestelling: condities en instrumentarium in het kader van de leerlinggebonden financiering. Amsterdam: Boom.

Riccio, C. A., Pliego, J. A., Cohen, M. J., & Park, Y. (2014). Executive Function Performance for Children With Epilepsy Localized to the Frontal or Temporal Lobes. Applied Neuropsychology Children, 1-8.

Rodin, E. A., Schmaltz, S., & Twitty, G. (1986). Intellectual functions of patients with childhood-onset epilepsy. Developmental Medicine & Child Neurology, 28(1), 25-33.

Rudolf, G., Valenti, M. P., Hirsch, E., & Szepetowski, P. (2009). From rolandic epilepsy to continuous spike-and-waves during sleep and Landau-Kleffner syndromes: insights into possible genetic factors. Epilepsia, 50 Suppl 7, 25-28.

Russ, S. A., Larson, K., & Halfon, N. (2012). A national profile of childhood epilepsy and seizure disorder. Pediatrics, 129, 256-264.

Ryan, J. J., Tree, H. A., Morris, J., & Gontkovsky, S. T. (2006). Wechsler Adult Intelligence Scale-III inter-subtest scatter: a comparison of brain-damaged patients and normal controls. J Clin Psychol, 62(10), 1319-1326.

Rzezak, P., Fuentes, D., Guimaraes, C. A., Thome-Souza, S., Kuczynski, E., Li, L. M., . . . Valente, K. D. (2007). Frontal lobe dysfunction in children with temporal lobe epilepsy. Pediatr Neurol, 37(3), 176-185.

Saling, M. M. (2009). Verbal memory in mesial temporal lobe epilepsy: beyond material specificity. Brain, 132(570 - 582).

Sattler, J. M. (1990). Assessment of Children (3rd ed.). San Diego: author. Sattler, J. M. (2001). Assessment of Children. Cognitive Applications (4th ed.). San

Diego: author. Satz, P., Strauss, E., Wada, J., & Orsini, D. L. (1988). Some correlates of intra- and

interhemispheric speech organization after left focal brain injury. Neuropsychologia, 26(2), 345-350.

Scheirs, J. G. M., & Timmers, E. A. (2009). Differentiating among children with PDD-NOS, ADHD and those with a combined diagnosis on the basis of WISC-III profiles. Journal of Autism and Developmental Disorders, 39, 549-556.

Schittekatte, M. (2005). Overzicht van onderzoek met de WISC-III (Overview of research on the WISC-III). Tijdschrift Klinische Psychologie, 35(4), 214-230.

Schmidt, D., & Sillanpää, M. (2012). Evidence-based review on the natural history of the epilepsies. Current Opinion Neurology, 25(2), 159-163.

Schouten, A., Oostrom, K., Jennekens-Schinkel, A., & Peters, A. C. (2001). School career of children is at risk before diagnosis of epilepsy only. Developmental Medicine & Child Neurology, 43(8), 575-576.

Seidenberg, M., Pulsipher, D. T., & Hermann, B. (2007). Cognitive progression in epilepsy. Neuropsychological Review, 17(4), 445-454.

Sillanpää, M. (1992). Epilepsy in children: prevalence, disability, and handicap. Epilepsia, 33(3), 444-449.

REFERENCES

199

Sillanpää, M., & Shinnar, S. (2010). Long-term mortality in childhood-onset epilepsy. N Engl J Med, 363(26), 2522-2529.

Silverstein, A. B. (1987). Two indices of subtest scatter on Wechsler's Intelligence Scales: estimated vs. empirical values. Journal of Clinical Psychology, 43(4), 409-414.

Silverstein, A. B. (1989). Reliability and abnormality of scaled-score ranges. J Clin Psychol, 45(6), 926-929.

Skirrow, C., Cross, J. H., Cormack, F., Harkness, W., Vargha-Khadem, F., & Baldeweg, T. (2011). Long-term intellectual outcome after temporal lobe surgery in childhood. Neurology, 76(15), 1330-1337.

Smith, M. L., Elliott, I. M., & Lach, L. (2002). Cognitive skills in children with intractable epilepsy: comparison of surgical and nonsurgical candidates. Epilepsia, 43(6), 631-637.

Snijders, T. A. B., & Bosker, R. J. (1999). Multilevel Analysis: An introduction to Basic and Advanced Multilevel Modeling. London: SAGE Publications Inc.

Spence, S. J., & Schneider, M. T. (2009). The role of epilepsy and epileptiform EEGs in Autism Spectrum Disorders. Pediatric Research,65, 599-606.

Staudt, M., Lidzba, K., Grodd, W., Wildgruber, D., Erb, M., & Krageloh-Mann, I. (2002). Right-hemispheric organization of language following early left-sided brain lesions: functional MRI topography. Neuroimage, 16(4), 954-967.

Strauss, E., Loring, D., Chelune, G., Hunter, M., Hermann, B., Perrine, K., . . . Barr, W. (1995). Predicting cognitive impairment in epilepsy: findings from the Bozeman epilepsy consortium. Journal of Clinical and Experimental Neuropsychology, 17(6), 909-917.

Tedrus, G. M. A. S., Fonseca, L. C., Melo, E. M. V., & Ximenes, V. L. (2009). Educational problems related to quantitative EEG changes in benign childhood epilepsy with centrotemporal spikes. Epilepsy and Behavior, 15, 486-490.

Thomas, R. H., & Berkovic, S. F. (2014). The hidden genetics of epilepsy-a clinically important new paradigm. Nature Review Neurology, 10(5), 283-292.

Towgood, K. J., Meuwese, J. D., Gilbert, S. J., Turner, M. S., & Burgess, P. W. (2009). Advantages of the multiple case series approach to the study of cognitive deficits in autism spectrum disorder. Neuropsychologia, 47(13), 2981-2988.

Tuchman, R., Alessandri, M., & Cuccaro, M. (2010). Autism spectrum disorders and epilepsy: moving towards a comprehensive approach to treatment. Brain and Development, 32, 719-730.

Vaessen, M. J., Jansen, J. F., Vlooswijk, M. C. G., Hofman, P. A., Majoie, H. J. M., Aldenkamp, A. P., & Backes, W. H. (2012). White matter network abnormalities are associated with ognitive decline in chronic epilepsy. Cerebral Cortex, 22(9), 2139-2147.

Vago, C., Bulgheroni, S., Franceschetti, S., Usilla, A., & Riva, D. (2008). Memory performance on the California Verbal Learning Test of children with benign childhood epilepsy with centrotemporal spikes. Epilepsy and Behavior, 13, 600-606.

van Haasen, P., de Bruyn, E., Pijl, Y., Poortinga, Y. H., lutje Spelberg, H. C., Vandersteene, G., . . . Stinissen, J. (1986). WISC-R. Wechsler Intelligence Scale for Children Revised. Nederlandstalige uitgave. Lisse, The Netherlands: Swets Test Services.

van Iterson, L. (2010). Failing to progress in secondary school and Reliable Changes in cognitive functioning in adolescents with epilepsy (Falta de progreso escolar en ensenanza secundaria y Cambios Confiables (RC) de nivel cognitivo en


200

adolescentes con epilepsia). Poster presented at the 6 congreso latinoamericano de epilepsia (6th Latin American congress on epilepsy), Cartagena, Colombia, Aug 1 - 4 2010.

van Iterson, L., & Augustijn, P. B. (2006). Unilateral hemisphere involvement associated predominantly to VIQ > PIQ discrepancies in children with (refractory) epilepsy. Epilepsia, 47, 158-159.

van Iterson, L., & Augustijn, P. B. (2013, July 10-12). Rates of Reliable Cognitive Change (RC) after a change of WISC test version in children with refractory epilepsy: adjustment for Flynn effects. Poster presented at the 2013 Mid year meeting of the International Neuropsychological Society, Amsterdam.

van Iterson, L., & Augustijn, P. B. (2014). Cognitive patterns in children with epilepsy in relation to number of anti-epileptic drugs taken at time of neuropsychological testing. Perfil cognitivo en niños con epilepsia y el efecto de número de antiepilépticos durante la evaluación neuropsicológica, 8th LACE, Buenos Aires (Argentina), 17th-20th September 2014.

van Iterson, L., Augustijn, P. B., de Jong, P. F., & van der Leij, A. (2013). Establishing reliable cognitive change in children with epilepsy: The procedures and results for a sample with epilepsy. Journal of Psychoeducational Assessment, 31(5), 448-458.

van Iterson, L., & Kaufman, A. S. (2009). Intra-individual subtest variability on the Dutch Wechsler Intelligence Scales for Children-Revised (WISC-R(NL)) for children with learning disabilities, psychiatric disorders, and epilepsy. Psychological Reports, 105(3 Pt 2), 995-1008.

van Iterson, L., Zijlstra, B. J., Augustijn, P. B., van der Leij, A., & de Jong, P. F. (2014). Duration of epilepsy and cognitive development in children: a longitudinal study. Neuropsychology, 28(2), 212-221.

van Luit, J. E. H., Bloemert, J., Ganzinga, E. G., & Mönch, M. E. (2012). Protocol Dyscalculie: Diagnostiek voor gedragsdeskundigen. (Dyscalculia protocol for psychologists). Doetichem, The Netherlands: Graviant Educatieve Uitgaven.

van Mil, S. G., Reijs, R. P., van Hall, M. H., Snoeijen, S. M., de la Parra, N. M., & Aldenkamp, A. P. (2010). Behavior in children with cryptogenic localization related epilepsy: a follow-up study. Pediatric Neurology, 42(2).

van Mil, S. G. M., Reijs, R. P., Van Hall, M. H. J. A., Snoeijen, S. M., de la Parra, N. M., & Aldenkamp, A. P. (2009). Behavioral status of children with cryptogenic localization-related epilepsy. Journal of Child Neurology, 24(4), 449-453.

van Rijckevorsel, K. (2006). Cognitive problems related to epilepsy syndromes, especially malignant epilepsies. Seizure, 15, 227-234.

Vander Steene, G., & Bos, A. (1997). WPPSI-R Wechsler Preschool and Primary Scale of Intelligence. Vlaams-Nederlandse Aanpassing. Handleiding.

Vermeulen, J., & Aldenkamp, A. P. (1995). Cognitive side effects of chronic antiepileptic drug treatment: a review of 25 years of research. Epilepsy Research, 21, 65-95.

Vermeulen, J., Kortstee, S. W. A. T., Alpherts, W. C. J., & Aldenkamp, A. P. (1994). Cognitive performance in learning disabled children with and without epilepsy. Seizure, 3, 13-21.

Verrotti, A., D'Egidio, C., Agostinelli, S., Parisi, P., Chiarelli, F., & Coppola, G. (2011). Cognitive and linguistic abnormalities in benign childhood epilepsy with centrotemporal spikes. Acta Paediatrica, 100, 768-772.

REFERENCES

201

Vicari, S., Albertoni, A., Chilosi, A. M., Cipriani, P., Cioni, G., & Bates, E. (2000). Plasticity and reorganization during language development in children with early brain injury. Cortex, 36(1), 31-46.

Walsh, P., Elsabbagh, M., Bolton, P., & Singh, I. (2011). In search of biomarkers for autism: scientific, social and ethical challenges. Nature Reviews Neuroscience, 12(10), 603-612.

Watkins, M. W., & Glutting, J. J. (2000). Incremental validity of WISC-III profile elevation, scatter, and shape information for predicting reading and math achievement. Psychological Assessment, 12(4), 402-408.

Watkins, M. W., Glutting, J. J., & Youngstrom, E. A. (2005). Issues in Subtest Profile Analysis. In D. P. Flanagan & P. L. Harrison (Eds.), Contemporary Intellectual Assessment: Theories, tests and issues (2nd ed., pp. 251-268). New York: Guilford.

Wechsler, D. (1992). WISC-IIIUK. Wechsler Intelligence Scale for Children. UK- Manual (3rd ed.). London: The Psychological Corporation. Harcourt.

Wechsler, D. (2004a). WAIS-III. Nederlandstalige bewerking. Wechsler Adult Intelligence Scale - Derde editie. London: Harcourt.

Wechsler, D. (2004b). WISC-IVUK Wechsler Intelligence Scale for Children (4th UK ed.). London: Harcourt Assessment.

Wechsler, D. (2005). WISC-IIINL. Wechsler Intelligence Scale for Children. Derde editie NL. Handleiding en Verantwoording. London / Amsterdam: The Psychological Corporation / Nederlands Dienstencentrum NIP.

Weiss, L. G., Saklofske, D. H., Schwartz, D. M., Prifitera, A., & Courville, T. (2006). Advanced clinical interpretation of WISC-IV index scores. In L. G. Weiss, D. H. Saklofske, A. Prifitera & J. Holdnack (Eds.), WISC-IV advanced clinical interpretation (pp. 99-138). San Diego, CA: Elsevier.

Westerveld, M., Sass, K. J., Chelune, G. J., Hermann, B. P., Barr, W. B., Loring, D. W., . . . Spencer, D. D. (2000). Temporal lobectomy in children: cognitive outcome. Journal of Neurosurgery, 92(1), 24-30.

Wirrell, E. C., Grossardt, B. R., So, E. L., & Nickels, K. C. (2011). A population-based study of long-term outcomes of cryptogenic focal epilepsy in childhood: cryptogenic epilepsy is probably not symptomatic epilepsy. Epilepsia, 52(4), 738-745.

Wirrell, E. C., Grossardt, B. R., Wong-Kisiel, L. C., & Nickels, K. C. (2011). Incidence and classification of new-onset epilepsy and epilepsy syndromes in children in Olmsted County, Minnesota from 1980 to 2004: a population-based study. Epilepsy Reseach, 95(1-2), 110-118.

Witt, J. A., & Helmstaedter, C. (2013). Monitoring the cognitive effects of antiepileptic pharmacotherapy--approaching the individual patient. Epilepsy and Behavior, 26(3), 450-456.

Woods, S. P., Childers, M., Ellis, R. J., Guaman, S., Grant, I., Heaton, R. K., & Group, H. I. V. N. R. C. (2006). A battery approach for measuring neuropsychological change. Archives of clinical neuropsychology, 21(1), 83-89.

Woods, S. P., Weinborn, M., & Lovejoy, D. W. (2003). Are classification accuracy statistics underused in neuropsychological research? Journal of Experimental and Clinical Neuropsychology, 25(3), 431-439.

Zimet, G., Goodman Zimet, S., Farley, G. K., Shapiro Adler, S., & Zimmerman, T. (1994). Intellectual competence of children who are beginning inpatient and day psychiatric treatment. Journal of Clinical Psychology, 50, 866-877.


202

Publications and Posters (related to the topics)

van Iterson, L. (2006). Different rates of directional VIQ – PIQ discrepancies in special education samples: below average IQ, learning disabilities, childhood psychiatry, and epilepsy. Poster presented at the INS/ SVNP / GNP conference: From plasticity to rehabilitation, Zürich, July 26 - 29.

van Iterson, L. (2007). Kaufman factors in three groups of special school children: learning disabilities. childhood psychiatry, and epilepsy. Poster presented at the International Neuropsychological Society, Federation of Spanish Societies of Neuropsychology, Spanish Neuropsychological Society, Spanish Psychiatry Society. Joint Mid-year meeting., Bilbao, July 4 - 7.

van Iterson, L. (2010). Failing to progress in secondary school and Reliable Changes in cognitive functioning in adolescents with epilepsy (Falta de progreso escolar en ensenanza secundaria y Cambios Confiables (RC) de nivel cognitivo en adolescentes con epilepsia). Poster presented at the 6 congreso latinoamericano de epilepsia (6th Latin American congress on epilepsy), Cartagena, Colombia, August 1 - 4.

van Iterson, L., & Augustijn, P. (2010). Cognitive outcome of childhood epilepsy in adolescence: utility of the Epilepsy Syndrome Severity Scale for Children (ESSS-C). Epilepsia, 51 (Suppl 4), 104.

van Iterson, L., & Augustijn, P. B. (2013). Rates of Reliable Cognitive Change (RC) after a change of WISC test version in children with refractory epilepsy: adjustment for Flynn effects. Poster presented at the 2013 Mid year meeting of the International Neuropsychological Society, Amsterdam, July 10 - 12.

van Iterson, L., & Augustijn, P. B. (2014). Cognitive patterns in children with epilepsy in relation to number of anti-epileptic drugs taken at time of neuropsychological testing. Perfil cognitivo en nin˜ os con epilepsia y el efecto de número de antiepilépticos durante la evaluación neuropsicológica, 8th LACE, Buenos Aires (Argentina), September 17 - 20.

van Iterson, L., San Miguel-Montes, L., & Rios, M. (2009). The factor analytic structure of the WISC-R and WISC-III in children with||epilepsy: a study across countries and cultures. Epilepsia, 50 (S10), 145.

van Iterson, L., San Miguel Montes, L., Rios-del Pozo, J., & Rios, M. (2009). Classificatory utility of large intra-individual WISC-R variability in focal vs generalizes seizures: Data across countries and cultures. Puerto Rico & The Netherlands.

Additional publications and posters (not related to the topics)

Augustijn, P. B., & van Iterson, L. (2011). Vitamin D status in an out clinic patient population of a tertiary referral epilepsy center. Epilepsia, 25 (suppl 6), 25.

van Iterson, L. (2002). Buidelnesten. Wandelingen door Colombia en Venezuela. Baarn: Zipatá.

PUBLICATIONS

203

van Iterson, L. (2009). Cued learning trials for children with epilepsy in story retelling: gains at group level, but not all individuals thrive. Epilepsia, 50(4 (S)), 165 - 166.

van Iterson, L. (2010). Nidos de Oropéndola. A pie por los Andes de Colombia y Venezuela. Bogotá: La Serpiente Emplumada.

van Iterson, L., Augustijn, P., Mora, E., & Parra, P. (2005). Neuropsychological functioning during non convulsive status epilepticus in ring chromosome 20. Selective impact on cognitive measures. Epilepsia, 46, supplement 6, 3-415

van Iterson, L., Augustijn, P., & Neijens, L. (2008). Learning and forgetting of AVLT word lists in children with focal epilepsy. Journal of the International Neuropsychological Society, 14(2), 114.

van Iterson, L., & Broekhuis, J. B. C. (2007). De hellingshoek van het vergeten (The slope of forgetting). Epilepsie, 5, 15 - 16.

van Iterson, L., Davelaar, S., & Augustijn, P. B. (2014). Story retelling in 14 and 15-year old youngsters with epilepsy compared to control children matched for initial learning score: accelerated long term forgetting. Poster presented at the 11th European Congress of Epilepsy, Stockholm, June 29 – July 3.

Acknowledgements

ACKNOWLEDGEMENTS

207

Acknowledgements

The children who participated in the testing are the most important in this work. Compacting

data renders individuals unrecognizable. Undertaking the testing, however, has been an effort

to identify each child’s individuality. Working with individual children and their parents has

infused my conviction, that the child tested today is the only one who counts. While

seemingly an iterative process, to engage in a new assessment is, in fact, a new adventure.

Beyond the most significant – the children – there are many I wish to mention and

thank in these pages. First, Professor Ley Sander, research director of Stichting Epilepsie

Instelligen Nederland, SEIN. Ley advised in search for promotores in an early stage,

commented critically on the work. He encouraged the thesis with words which got their

reassuring strength from their simplicity: “if you want to write a thesis, you can,” sometimes

adding, “the more, the merrier”. Obrigada, Ley, por sua confiança. By the same token, I wish

to thank SEIN and De Waterlelie , for the opportunity to work on this project.

Second, I am most indebted to Professor Peter F. de Jong at the University of

Amsterdam for his efforts to guide me as a clinician and external student. With his

background as a trained psychologist and methodologist, Peter imparted his overwhelmingly

razor-sharp thoughts in every discussion. He literally “promoted” the work to a new stage,

always improving it. It was humbling and instructive to experience how Peter’s own

professional and personal endeavours would express themselves on the thesis. Thank you for

making the project increasingly inspiring and challenging – hartelijk dank, Peter.

Third, this work owes much to Professor Alan S. Kaufman from Yale School of Medicine. It

does so in diverse ways, both related to his own work and as well as to him as a person and

supervisor. To start with his work: Alan Kaufman’s seminal work, Intelligent testing with the

WISC-R (Kaufman, 1979), provided the inspiration for the first building block of the present

work. I read his book in 1979 – 1980, when it had just appeared on the market, I had just

started my first job and the director of the institution on school advice gave me the book to

review and to advice whether it should be considered for the institution’s library.

Before detailing the other ways I am indebted to Professor Kaufman, I start with a

diversion which will lead to tell how I got to know him and his wife personally. This –

atypical – detour takes me to the Pacific coast of Colombia: In 2000, I spent some pleasant

days in a primitive hotel. The owner had been a famous local cook in the village in long


208

forgotten years. When I knocked at her door, the clocks of the hotel had stopped ticking.

Time had moulded off the signs from the front, no longer identifying the house as a hotel or

restaurant. The lady received me – and later other guests as well – only hesitantly. She no

longer had the keys for the rooms. No wrinkle of surprise would show on her face when I told

her that other guests, returning drunk in the middle of the night, mistook my room for theirs.

With the devotion of a famous cook, my landlady sat down to accompany me at every

meal. While watching me as I ate, one evening she contended that she could have long been

dead. In the early sixties, she said, she had “given away her seat” to a school teacher who

urgently needed a place for a flight. She gave him the ticket she had booked for herself. As

she was no longer traveling, the morning of the flight she went to church. Meanwhile, the

teacher took her place in the airplane and left for the capital. At exiting the church, an

enormous turbulence startled my landlady. The airplane had crashed against the Andes. All

passengers died. “The teacher was replaced,” she added laconically (van Iterson, 2002, 2010,

unrelated publications).

So I could have been warned when, years later, I took my colleague’s place and went

to a conference on the Kaufman Adolescence and Adult Intelligence Test. I could have

known that taking someone’s seat could change your fortune. Fortunately, “being given

someone else’s seat” is not always for the worst – quite the contrary.

My colleague neuropsychologist from SEIN, Willem Alpherts, had registered for the

day conference on intelligence testing in Amsterdam but readily gave me his “seat” when he

noticed that I also was eager to go. Alan and Nadeen Kaufman gave a presentation on their

test which had just been published in German and Dutch, as well as a brilliant lecture on

stability and decline of cognitive abilities during the course of life – not unlike what we

describe in epilepsy at a much younger age.

After the conference, I approached Professor Kaufman with hesitant timidity. I was

thrilled to meet a man who was not exhausted after a daylong talk, by remarkably alert. “Send

your paper to me…” he said and scribbled his e-mail address on a piece of paper.

It was the beginning of intensive correspondence, joint research on the Dutch WISC –

it was the start of the dissertation. Alan provided examples of his own-work-in-progress to

help me get acquainted with the process of writing and editing. He asked me to comment on

them to obtain experience in reviewing. He guided the search of and dealing with journal

editors. Throughout the process of thinking and writing, Alan critically read every single

word. In Alan Kaufman I met a gentle person who swiftly would reply every letter, would

answer every question and would keep every promise. Together with Nadeen we shared an

ACKNOWLEDGEMENTS

209

evening with a private tour at the Louvre, concerts in Paris and in their home town San

Diego. Over a meal in a Japanese restaurant we discussed a thesis on the Wechsler Scales.

And yes, we shared the publication of what would be the first article. At my various visits, I

would leave San Diego with a suitcase full of books, the Spanish translation of his famous

seminal book being one of them.

Alan Kaufman did more. He gave me the case with the WISC-V test which was far

from being launched. He asked me to study it overnight, and let me comment on it the next

day. Nadeen and Alan, beyond your guidance, inspiration and scientific contribution, I want

to thank you for your friendship.

Fourth, I thank Bonne Zijlstra from the University of Amsterdam for sharing his expertise in

statistics and methodology which greatly improved the fourth and fifth paper. Also, I am

grateful for his reassuring words when most needed. I very much admired Bonne’s calm,

friendly and youthful wisdom.

I much appreciated Professor Aryan van der Leij’s enthusiasm when we first talked on the

idea of working on Wechsler scales in clinical samples. In every word I sensed the shared

clinical background and the shared interest of turning an “internal database” into an empirical

research project.

I wish to thank Professor Hilde Geurts and Professor Maurits van der Molen (University of

Amsterdam), Professor Wilma Resing (University of Leiden), Professor Pol Ghesquière

(Catholic University Leuven) and dr. Geert Thoonen (Radboud University Nijmegen) for

agreeing to take place in the “promotiecommissie”, for reading and rethinking the thesis.

To my co-author Paul Augustijn, child neurologist and epileptologist of SEIN, I am indebted

the insight in the classification of the epilepsies and the scoring of the syndrome severity, as

well as his solidarity throughout the years.

I thank Jacob Aten for his support and general indications about how the preparation

of a thesis and the finding of promoters take place and his keen interest during the course of

the project.

I thank Roos Rodenburg for her actual help in the process of finding the Dutch thesis

directors. She was the colleague from SEIN to connect me to the University of Amsterdam,

contending that is was worth analysing “een bak met data”, as she called the vast data


210

collection. Together with Roos, we extended the data with joint work with her students at the

University of Amsterdam. I thank her especially for her friendship. Roos Rodenburg and Paul

Augustijn, I am very pleased that both of you are willing to act as a paranymphs for the

ceremony.

I thank Ben Schomakers for his encouraging sentence: “you need not write a

proefboek, but if you want to write it, you can”, consistent over the years we shared. His

superb eloquence allowed him to configure a new word in order to make it sound an easy job.

I want to mention my colleagues at the various work places. Form earlier times, Han

Meliëzer and Dienke Evenhuis. More recently, through Willem Alpherts, Mieke Siffels,

Agnes Dommerholt, Erna Haverkort, Huibert Geesink and Dimitri Velis, I want to thank all

at SEIN’s psychology, neurology and neurophysiology departments. Through Lidwien

Neijens, Rien Moerland, Karin Luymes, Annemieke Smit, Marije van Wijngaarden, Jan van

der Kuip, Teus van den Brink, Piet Klein, José van Veen, Marian Fenenga, Eberdien

Karsdorp, Richard Donkers, Carola Bol, I want to thank all from De Waterlelie.

A Fernando Zamora le agradezco la revisión del Resumen.

It was Bernhard Sleumer, my brother-in-law, who long time ago, put forward: “why

not write a PhD thesis?” Sadly, he died the day after I sent the manuscript to the committee.

I also wish to say köszönöm to György Katalin, dankjewel to Marie and Mette

Lindhout, dankjewel to Bernhard and to my sister Foyita, and gracias to my brothers Victor

and Marnix for their explicit or tacit consent to spend my spare time at the computer, my

holidays at the library of Budapest, in hotel rooms of seldom visited Portuguese-speaking

islands, in down-town Cartagena or in a tiny study in Spain – instead of sharing my hours

with them.

Finally, I want to thank Dick for returning me my smile. Dick, it is simply wonderful to be

with the tender lovable person you are.

I dedicate this work to my mother for teaching her children that it is worthwhile to pursue

writing down their thoughts. I dedicate this work to the memory of my father for wordlessly

understanding that the instant a study is concluded, is the moment one has the least

understanding of the topic.

Baarn, Bogotá, Budapest, Buenos Aires, Cartagena, Cómpeta, Ilha do Mel, Leiden. May 2015.

About the author

ABOUT THE AUTHOR

213

About the author

Loretta van Iterson was born January 3rd 1956 in The Hague, The Netherlands, but was

raised in Bogotá, Colombia, where she moved with her family just after turning two years

of age. She had her kindergarten, primary and secondary

schooling at the Colegio Andino, the German school in Bogotá.

After completing her bachillerato, at age 17 she returned to The

Netherlands to study psychology at the Radboud University

(then called Katholieke Universiteit Nijmegen). By the time of

her graduation in developmental psychology in January 1980,

she was working as a child psychologist at the SABD, the school advice centre in Drenthe

that covered all schools for regular and special education in the province. This work gave

her the opportunity to get acquainted with all kinds of children with developmental needs.

Over time she changed her work premises to the school for children with psychiatric

disorders (Van der Reeschool, Smilde) and in 2002 she added to it the School De

Waterlelie (Cruquius) for children with epilepsy. From 2003, does the neuropsychological

assessment of children referred by the child neurologists in the clinics of Stichting

Epilepsy Instellingen Nederland, SEIN Heemstede and takes part of the Child Epilepsy

Centre (KEC).

In 1984, she did an informal neuropsychology internship with Barbara C. Wilson at North

Shore University Hospital in the state of New York. In 1992 – 1993, she specialized as a

child neuropsychologist at the European Graduate School of Child Neuropsychology

(EGSCN, PI Duivendrecht, Free University of Amsterdam). Being a native speaker of

Spanish and Dutch, in between, in 1987, she graduated as a sworn translator and

interpreter for the Dutch and Spanish languages in Utrecht.

Between 1979 and today, with refreshing changes in content over the years, most

of her career has been devoted to neuropsychological assessment of children and

youngsters with special needs, and the guiding of parents and teachers. The vast majority

of children are Dutch, but sometimes South American as well. Her field has been the

exploration of cognitive development in children and youngsters with mild intellectual

disabilities, specific learning disorders, psychiatric and behavioural disorders, and

children with epilepsy, in regular or special educational settings. Nowadays, together with


214

the school for epilepsy, the tertiary clinic for children with epilepsy is her major focus of

activities. Besides her interest in intelligence testing, she has particular interest in

improving test norms. As such, with her students she has collected experimental norm

data for a variety of neuropsychological tests. She took the initiative for the design and

development of a Dutch test on story learning and retelling, intended especially for

children with epilepsy, and is currently working on the manual.

Between 2000 and 2003, she took part of the pilot study of the Indication

Committees for children with psychiatric disorders to aid in the development and try out

of diagnostic criteria, which later led to the “CvI cluster IV”. Between 2005 and 2013, she

was a member of the RSG, the registration committee for specialisms in the mental

health, an organ of the FGzP (now called Federatie gezondheidszorg psychologen en

psychotherapeuten). From 2007 onward, she has been a member of the editorial board of

“Epilepsie”, a Netherlands journal for professionals in the field of epilepsy. Since 2008

she participates in “EURAP extension protocol”-project on the development of children

born to mothers with epilepsy. Since 2014 she is a member of the board of the Dutch

chapter of the League against Epilepsy (LIGA).

Loretta van Iterson is registered as a specialist on Children and Youngsters

(specialist K&J, NIP), as a mental health psychologist (Gz, BIG), as well as specialist

Clinical Psychologist (KP, BIG, FGzP), and as a sworn translator Spanish (Utrecht).

She interrupted her work career twice to spend a sabbatical year in South America. The

first time, in the mid-1980s, she spent in Colombia, the second time, near the turning of

the century, in Venezuela and partly in Hungary.

She is a keen traveller, mostly to South America but to many other – preferably

remote – places as well. She wanders around these far places on her own, sometimes with

her partner, sometimes aiming at visiting her brothers and sister scattered around the

world. She has authored four Dutch books on travels in Latin America, one of which has

also been published in Spanish, in Bogotá in 2010.

As a partner to Dick, Loretta rejoices in the lovability, undertakings and accomplishments

of his children.

Documents

UvA-DARE (Digital Academic Repository) Cognitive patterns ... · Resumen (Spanish summary) 153 Appendices 161 Appendix A Subtest scaled score range (subtest scatter) 166 Table A.1