Biological Psychiatry (Trimble/Biological Psychiatry) || The Schizophrenias

7The Schizophrenias

INTRODUCTION

Definition and Classification

Many of the illnesses referred to in the past as‘neuroses’ have melted away into more clearlydefined neurological conditions with a recog-nizable structural basis. So it is with Kraepelin’sconcept of dementia praecox. Although for himit represented a disease, whose origins mayoften be found in early life, its pathogenesis andphysiological accompaniments largely eludedhim. Bleuler, recognizing the heterogeneity ofthe condition, used the term ‘disease group . . .

about analogous with the group of the organicdementias’ (Bleuler, 1924, p. 373), while Slaterand Roth (1960) refer to schizophrenia as a term‘for a group of mental illnesses characterizedby specific psychological symptoms’ (p. 237).Gradually over time some aetiological factorshave become clarified.

Schizophrenia therefore should be consid-ered not as a single disease, but, like epilepsy(or the epilepsies), as a group of syndromes,recognized by a collection of signs and symp-toms, which have diverse pathogeneses. It issuggested that the symptoms represent theoutcome of abnormal cerebral functioning,itself provoked by disease processes, some ofwhich we readily recognize. However, it isimportant to understand that, as time passes,and as the cerebral basis of schizophrenicsymptoms becomes clearer, the clinical diag-nostic category will get whittled down, and

preferred pathological diagnoses will begiven. This is one process whereby medicineadvances, and the principle is as important inpsychiatry as in other areas. Thus, Parkinson’sdisease, dementia (the dementias) and epilepsy(the epilepsies) are no longer thought of asclearly defined entities with a common cerebralpathology, but as recognizable conditionson clinical grounds, perhaps with commonunderlying functional changes that helpexplain the symptom pattern, but with severalpathological antecedents.

The concept that schizophrenia (or more aptlythe schizophrenias) is a single disease entity hascaused much confusion in psychiatry, even lead-ing some authors to the fanciful conclusion thatit does not exist (Szasz, 1976). There remainarguments over the precise definitions, if onlybecause they are operational and clinical, butthis is common with many conditions. How todefine epilepsy is likewise an elusive problem(see Chapter 10). The use of the plural in thechapter title and intermittently in the text how-ever emphasizes that the schizophrenias shouldproperly be viewed as a collection of disorderswith differing pathogeneses.

This view is not that developed by Kraepelin,who, while acknowledging the differing clin-ical pictures that may arise, noted the samebasic disturbances suggestive of a commonunderlying process. Bleuler (1911), who intro-duced the term schizophrenia ‘to show that thesplit of the several psychic functions is one of its

Biological Psychiatry, Third Edition Michael R. Trimble and Mark S. George© 2010 John Wiley & Sons Ltd. ISBN: 978-0-470-68894-6

148 Biological Psychiatry

most important characteristics’ (p. 5), extendedthe range of the observed psychopathology,referred to the group of schizoprenias andintroduced Freudian psychodynamics into hisideas. He suggested that many of the typicalsymptoms were determined by ‘psychic’ causesand distinguished various types, includingprimary, secondary and basic. The symp-toms delineated by Kraepelin were mainlysecondary, derived by a reaction of the sickmind to the illness, primary ones being directlycaused by the disease process. Basic symptoms,present at all times, were alteration of affect andvolition, ambivalence and autism. The theorywas elaborated further, particularly by Jung, toinclude the idea that the psychic cause couldset an organic process in motion, thus beingaetiological. In addition, Bleuler considerablywidened the boundaries of the disorder toinclude personality changes and other subtlechanges of mental function, atypical maniasand melancholias, anergia, obstinacy andmoodiness. Simple schizophrenia was addedto the classification, which included many ofthese marginal cases, and the idea of latentschizophrenia was accepted. Bleuler’s influencewas profound, not only in central Europe,but also in the USA through the writings ofAdolf Meyer.

An important step was taken by KurtSchneider. Influenced by both Kraepelin andJaspers, he was especially concerned withsymptoms, and was dissatisfied with the waythat Bleuler defined such concepts as primaryand basic ones. He introduced the terms ‘firstrank’ and ‘second rank’, quite different fromthe divisions of other, previous writers. Thefirst-rank symptoms, shown in Table 7.1, wereintroduced for pragmatic diagnostic use, and ‘ifthis symptom is present in a non-organic psy-chosis, then we call that psychosis schizophrenia. . . the presence of first rank symptoms alwayssignifies schizophrenia, but first rank symptomsneed not always be present in schizophrenia’(Schneider, 1957, p. 44).

Schneider had no presumption of a commonstructure for these phenomena, although hediscussed them in terms of the lowering of a‘barrier’ between the self and the surrounding

Table 7.1 The first-rank symptoms of Schneider(after Schneider, 1957, p. 43)

The hearing of one’s thoughts spoken aloud inone’s head

The hearing of voices commenting on what one isdoing at the time

Voices arguing in the third personExperiences of bodily influenceThought withdrawal and other forms of thought

interferenceThought diffusionDelusional perceptiona

Everything in the spheres of feeling, drive andvolition which the patient experiences asimposed on them or influenced by others

aAn abnormal significance attached to a real perception without anycause that is understandable in rational or emotional terms.

world. Other symptoms were termed second-rank, and were considered much less importantfor diagnosis.

The first-rank symptoms assumed impor-tance for the diagnosis of schizophrenia, espe-cially in the UK, and were employed widely inresearch, for example the Present State Exam-ination (PSE) being biased to detect them forthat diagnostic category. Their influence on theDSM III and its successors is clear.

The International Pilot Study of Schizophre-nia was a transcultural investigation of1202 patients in nine countries: Colombia,Czechoslovakia, Denmark, India, Nigeria,China, the USSR, the USA and the UK (WorldHealth Organization, 1973). The main ratinginstrument was the PSE. The symptom profilesand the rank order of the most frequentsymptoms were similar across countries, andwhile no single symptom was present in allpatients from all cultures, the conclusion wasthat when psychiatrists diagnose schizophreniathey have the same condition in mind. Lack ofinsight, delusional mood, ideas of reference,flatness of affect, auditory hallucinations andpassivity experiences were commonly recordedand although first-rank symptoms were presentin only one-third, when recorded they almostinevitably led to a diagnosis of schizophrenia.The Americans and Russians differed the most,having a broader concept of the conditionthan the others. This difference in part led to

The Schizophrenias 149

the tightening up of the diagnostic criteria inthe DSM III.

The ICD 10 refers to schizophrenic disordersas a group in which there is a disturbance of thepersonality, distortion of thinking, delusions, asense of outside influence, disturbed percep-tion, an abnormal affect out of keeping withreality, and autism. Thought, perception, mood,conduct and personality are all affected by thesame illness, but the diagnosis is not restricted toone with a deteriorating course. The main typesrecognized are undifferentiated, hebephrenic,catatonic and paranoid, but residual and simpletypes were also included.

It is necessary for symptoms to have beenpresent for a month or more for a diagno-sis of schizophrenia to be given. For a moreacute episode, the term acute schizophrenia-likepsychotic disorder is used.

DSM IV-TR includes characteristic symp-toms, such as hallucinations, delusions andthought disorder, but emphasizes also socialand occupational dysfunction. The problemsmust have been present for at least six months.Illnesses lasting for a shorter time are referredto as schizophreniform disorder. Such Schneide-rian symptoms as voices keeping a runningcommentary on the patient’s behaviour, or twoor more voices conversing with each other,are specifically associated with a diagnosisof schizophrenia.

DSM IV-TR gives the following subtypes: dis-organized, catatonic, paranoid, undifferentiatedand residual. The disorganized type is character-ized by incoherence, an inappropriate bluntedaffect associated with mannerisms, social with-drawal and odd behaviour, and by an absenceof well-systematized delusions. This is equatedwith the hebephrenia of the ICD 10. In the cata-tonic type there is marked motor disturbance,sometimes fluctuating between the extremesof excitement and stupor. The paranoid typeemphasizes preoccupation with delusions orhallucinations. The undifferentiated type is forthose not classified in these groups, and theresidual is for patients who have a history of atleast one past episode but have minimal psy-chotic symptoms in the presence of continuedevidence of the illness.

Missing from the DSM IV-TR criteria are themotor disorders seen in schizophrenia (otherthan the catatonia), which form an integral partof the syndrome. These include stereotypies,negativisms, automatic behaviours, posturingand catalepsy. In addition, tremors, tics andchoreoathetosis have all been identified – notas a consequence of treatment but as a partof the disorder (Trimble, 1981a). Further, theemphasis on current or recent symptomsdetracts from the essential nature of schizophre-nia, which, in the majority of cases, pursues achronic course.

The history, typically, is of a patient, who maywell have demonstrated early developmentallanguage and behaviour problems, who mani-fests bizarre behaviour in late teens and earlyadulthood with a loss of academic potential.Initially, a bewildering variety of psychopatho-logical states, including anxiety states, maniasand depressions, may be recorded, but eventu-ally the longitudinal course becomes apparentand the cognitive and psychotic symptoms out-lined above manifest.

DSM IV-TR notes that psychotic disorder canbe due to a wide variety of general medicalconditions. This excludes disorders that onlypresent as delirium, and is subdivided depen-dent on whether the predominant symptomsare delusions or hallucinations.

Both the DSM IV-TR and the ICD 10 have thecategory schizoaffective, a term that is a sourceof considerable controversy. It was first intro-duced to refer to a psychosis with an admixtureof affective and schizophrenic symptoms that isprecipitated by emotional stress. Sitting betweenthe two Kraepelinian pillars of psychosis, thisdiagnostic category is not accepted by many, itsvery existence suggesting a return to a unitarypsychosis theory (see below). However, fam-ily and genetic data support the concept thatsome form of schizoaffective subtype exists,though its prognosis is related more to theaffective disorders (Procci, 1976). Some use theterms ‘schizodepressive’ and ‘schizomanic’ toreflect this; DSM IV-TR refers to the bipo-lar and the depressive types. It is diagnosedonly when definite schizophrenic and definiteaffective symptoms are present simultaneously.


It must be distinguished from post-treatmentschizophrenic depression.

The schizoaffective subtype may describethe same patients referred to as having cycloidpsychosis (Leonhard, 1957). The features arepsychotic episodes which resolve completelyand are associated with confusion and per-plexity, mood and motor changes, pan-anxietyand delusions.

Kraepelin’s original concept, of a dementiapraecox, embedded the idea of a progression.There have now been several follow-up studies,particularly with brain imaging (see below), thatrather support his views. Thus it is estimatedthat some 50–70% of cases take a progressivecourse, which includes not only the psychoticsymptoms but also the cognitive changes andloss of the ability to function independently at ahigh level in society.

Paranoid Disorders

The word ‘paranoid’ has been used in psy-chiatry for many years, but its meaning hasshifted from a general term for madness to amore precise technical expression (Lewis, 1970).Kraepelin initially classified paranoia asa separate entity from dementia praecox;it represented the insidious development of adelusional system in the presence of an intactpersonality. He used the term ‘paraphrenia’to refer to those who have a paranoid illnessdeveloping later than dementia praecox.Lewis (1970), after a careful consideration of theworld literature, gave the following definition:‘A paranoid syndrome is one in which thereare delusions of self reference which maybe concerned with persecution, grandeur,litigation, jealousy, love, envy, hate, honour,or the supernatural, and which cannot beimmediately derived from a prevailing morbidmood such as mania or depression’ (p. 11).He felt that the adjective ‘paranoid’ could beapplied to a personality disorder with similarfeatures, except that dominant ideas shouldreplace delusions.

ICD 10 uses the term ‘persistent delusionaldisorders’, and DSM IV-TR ‘delusional disor-der’ for these syndromes. Patients develop a

single delusion or a related set of delusions,not necessarily so bizarre in content, in theabsence of criteria for schizophrenia. Such delu-sions often arise in middle age, can contain ideasabout the individual’s body and should last atleast one month. In many patients, these disor-ders are persistent, but psychosocial functioningmay remain adaptable. DSM IV-TR recognizesseveral types, namely erotomanic, grandiose,jealous, persecutory, somatic and mixed. Thelitigious variant is also well known to lawyers.

An important addition to this group is thepsychotic disorder due to a general medical con-dition. Prominent symptoms are hallucinationsand delusions not due to delirium or dementia.

Endophenotypes

The term ‘endophenotypes’ refers to a marker ofan illness that might be closer to the underlyinggenetic mechanisms than the actual phenotype,as for example specified by DSM IV-TR andICD 10. It was introduced in 1972 by Gottesmanand Shields and represents a neurobiologicaltrait that is related to the illness in question. Theconcept has assumed particular prominence inthe genetics of schizophrenia and is one wayto deconstruct the phenotype into componentparts for a potentially more powerful geneticanalysis. Most studies have highlighted cogni-tive and brain-imaging variables.

GENETICS

The contribution of genetics to schizophreniahas been long recognized, but is often curiouslyneglected in some pathogenic explanations ofthe disorder. It has been reviewed in detailby many authors (McGuffin, 1984; Slater andRoth, 1960; and more recently Gur et al., 2007;Norton et al., 2006; Porteous, 2008; Riley andKendler, 2006; Tandon et al., 2008; van Winkelet al., 2008). In brief, the twin studies showimpressive differences in rates of concordancebetween monozygous (MZ) and dizygous (DZ)pairs, and the adoption data indicate that thereis a greater incidence of schizophrenia andschizophrenic spectrum disorders in those who,


Table 7.2 Classic twin studies in schizophrenia

Study Concordance

Monozygous (%) Dizygous (%)

Kringlen (1967) 45 15Pollin et al. (1969) 43 9Tienari (1971) 35 13Fischer (1973) 56 26Gottesman and Shields (1972) 58 12Kendler and Robinette (1983) 40 6Mean 46.1 13.5

adopted to normal parents, have a biologicalparent with the condition. Table 7.2 summarizesthe data, indicating concordance rates from 35to 58% in MZ and from 6 to 26% in DZ pairs.

While there were several criticisms ofthe early data, for example the method ofascertainment of the zygosity and the clinicaldiagnostic criteria for the schizophrenia, laterstudies have attempted to overcome these.Some authors (Gottesman and Shields, 1972;Kendler and Robinette, 1983) noted that themore severe the illness, the more likely theconcordance in MZ pairs, and Kallmann (1946),whose own series noted a 91.5% concordancefor MZ twins living together, reported arising familial incidence of the condition assubjects approach a genetic relationship withindex cases. Thus, compared to a percentageexpectation of 0.9% for the general population,the equivalent figure for half-sibs was 7.1%, forparents 9.2%, for full-sibs 14.2%, for children16.4%, for DZ co-twins 14.5% and for childrenof two schizophrenic parents 39.2%. Aftera review of the data, Slater concluded ‘theevidence is very strong that the geneticalconstitution of an individual contributes alarge part of his total potentiality of becomingschizophrenic’ (Slater and Roth, 1969, p. 246).McGuffin (1984) calculated that the heritabilityof the condition was 0.66, genetic factors thusaccounting for two-thirds of the variance invulnerability to the disorder. A similar figure(68%) has been given by Kendler (1983), whonotes that this is similar to that for diabetesand hypertension, and rather in excess ofthat for epilepsy, peptic ulcer and coronaryartery disease.

There is some suggestion that schizophrenicprobands resemble their affected relatives withregards to subtype, although there is muchoverlap, and twin and adoption studies whichmight clarify this issue have not been conducted.Leonhard (1980) has suggested, on the basis offamily studies, that periodic catatonia has a highgenetic loading, and Perris (1974) noted the samefor cycloid psychosis. Winokur (1977) reportedthat paranoia may also be genetically separate.

McGuffin et al. (1984) examined the heritabil-ities of a twin population using six different setsof operational criteria for schizophrenia. Thoseof Feighner and the RDC criteria gave the high-est heritability, while the Schneiderian criteriahad a heritability value of zero. Dworkin andLenzenweger (1984) examined the case histo-ries of MZ twins in whom at least one hadschizophrenia and assessed concordance forpositive and negative symptoms. Their resultssuggested that negative symptoms may have thegreater genetic component. Similar conclusionshave emerged from the studies of Tsaung (1993).

With regards to interpretation, the contentionthat environmental factors are so important,in that the environments of MZ twins are saidto be more alike than those of DZ pairs, hasbeen discounted by the adoption studies.Heston (1966) showed that the adopted childrenof schizophrenic mothers, removed from themshortly after birth, had a higher incidence ofthe illness than the adopted children of non-schizophrenic mothers. These findings havebeen replicated and extended by a seriesof studies on Danish patients, including inves-tigation of the relatives of adopted childrenwho later became schizophrenic (Kety, 1983;


Kety et al., 1994). These studies have notonly shown the higher incidence of schizophre-nia among the biological relatives but alsoan overrepresentation of the schizophreniaspectrum of disorders. Further, there wasan overrepresentation of schizophrenia orschizophrenia-spectrum disorders in thebiological relatives of adopted schizophrenicpatients, significantly greater than amongthe relatives of nonschizophrenic adoptedcontrols. In addition, in investigations of pater-nal half-siblings of schizophrenic probands,the incidence of schizophrenia was higherthan in control cases, ruling out intrauterinecontributions to the congenital effects.

Replication studies (Kendler et al., 1981),particularly improving the criteria forschizophrenia-spectrum disorders, have beenconducted. The prevalence of schizotypal per-sonality disorder was significantly higher in thebiological relatives of schizophrenic adopteesthan controls, but this was not the case for delu-sional disorder of the nonschizophrenic typeor for anxiety disorder. These adoption studies,particularly the careful analysis of the Danishpatients, have confirmed the genetic contribu-tion to schizophrenia and identified the conceptof the schizophrenia spectrum more clearly.

Gottesman and Bertelsen (1989) observed thatthe risk for schizophrenia and related disorderswas similar in the offspring of schizophrenicMZ twins and those of their nonschizophrenicco-twins, suggesting that the discordance ratein MZ twins may in part be explained by anunexpressed genetic diathesis.

With regards to the mode of inheritance,several authors favour polygenic as opposed tomonogenic transmission (Kendler, 1983), math-ematical models failing to provide evidence forthe latter. To date, no consistent traditionalgenetic markers linking to schizophrenia havebeen identified. However, the identification ofsuch markers will only be successful if thereis a major gene effect (either single or a smallnumber) or if there is a subgroup with such acomponent. HLA markers have been claimed,but the results of different studies are variable.McGuffin et al. (1981), after review of the

available literature, noted that the mostpersistent genetic marker to date was HLA A9with paranoid schizophrenia, and the moreconsistent associations (found by more thanone group) relate to HLA A9 and B5.

There are several important negative-linkagestudies to regions of chromosomes 2, 5, 9, 11, 12and 22 and to the D2, D3, D4, 5-HT2 and GABAasubunits (Macciardi et al., 1994; Waddingtonet al., 1992; Su et al., 1993). No mutation of theD1 receptor gene has been found.

Crow (1988) suggested that the pseudoau-tosomal locus of the X or Y chromosomes isimportant. This is a region of the distal seg-ment of the short arm of the Y chromosome thatin meiosis exchanges genetic material with theshort arm of the X chromosome. Genes thereare therefore transmitted in an autosomal ratherthan sex-linked way. Further, genes of the pseu-doautosomal locus do not seem subject to Xinactivation. His theory is based on a numberof observations, notably that the age of onset ofschizophrenia is earlier in males, that pairs offirst-degree relatives with psychosis are morelikely to be the same sex, that in parent–childpairs with psychosis the mother is twice aslikely to be affected than the father and thatconcordance by sex is paternally rather thanmaternally inherited. Further, sex-chromosomeaneuploides such as XXY, XX, but not XO,are associated with an increased incidence ofschizophrenia. No direct evidence of such link-age has yet been found (Crow et al., 1994), butCrow’s theory has developed to include a searchfor the gene responsible for handedness andcerebral dominance, which may be located atthe pseudoautosomal region (Crow, 1991).

The substantial evidence for genetic influ-ences on the development of schizophreniahas led to several studies of high-risk people.These include the Edinburgh group (Lawrieet al., 2008a, 2008b; Owens and Johnstone, 2006),NAPLS (the North America Prodrome Longi-tudinal Study) (Addington et al., 2007) and theEPOS (the European Prediction of PsychosisStudy). Each of these studies has taken subjectswith at least one parent with schizophrenia andfollowed them longitudinally, gathering data


at inception before there was any diagnosisof the disorder. They are assessed at variousintervals, and those who develop the disorderare then compared with those who do noton many different variables. The studieshave included psychological and behaviouralvariables, neurocognitive profiles and imaging,although each study has varied with regards toentry criteria and measurements selected. Somehave included healthy normal controls andsome alternative high-risk groups such as thosewith learning disability (or mental handicap).Assessments have included such ratings as theStructured Interview for Schizophrenia (SIS),the Structured Interview for Prodromal Syn-dromes (SIPS) and tasks of verbal abilities andworking memory. The results of these studiesare awaited with interest. These massive projectsoffer some promise, but caution is needed inlight of the amount of genetic work done todate that has not shown real progress. Oneproblem is that as sample sizes have increased,the signals heralded from earlier studies seemto have weakened rather than getting stronger(Weinberger, 2009). The genome-wide studiesare in fact revealing overlaps in the revealedgenetics of schizophrenia and bipolar disorder.At present, if differences can be suggested, it isbetween susceptibility for bipolar disorder andgenes involved in calcium ion-channel function,and the zinc-finger transcription factor inschizophrenia. Two promising position candi-dates in schizophrenia are NRG1 (neuregulin)and DTNBP1 (protein-coding dystrobrevin-binding protein 1). The deletions of the neurexingene (involved in the development of synapses)may represent a discontinuity with bipolardisorder and the potential genetic links betweenthe latter and the genes coding for the GABAareceptor and calcium channels may turn out tobe significant (Kirov et al., 2009).

It needs to be emphasized however that thegenetic variations noted so far in these disordersexplain only a small amount of the variabilityof the phenotype, and the limitations of thetechnique with regards to the statistics involvedand the definition of phenotypes as discussed inChapter 4 need to be taken into account.

Landis and Insel (2008) described only cau-tious optimism toward the results in this areaat present.

SOMATIC VARIABLES

Abnormalities in muscle endplates in schizo-phrenic patients have been reported. Usingmuscle biopsies stained for motor neuronesand terminals, Crayton and Meltzer (1976)reported the size and dispersion of terminalbulbs to be abnormal in psychotic patients, withlarger and more variable endplate arborizationand low density of endplate neural structures.Although similar findings were seen in somemanic–depressive patients and no associationwas noted with type or length of illness, theauthors felt the findings might indicate regener-ative processes of previously denervated fibres,reflecting perhaps altered CNS physiology.These data are in keeping with other evidenceof disturbed muscle pathology in psychoticpatients. A number of authors have noted anincrease in serum creatine phosphokinase inacute psychoses. This does not simply reflectthe use of intramuscular injections or methodsof restraint since elevations are found in aproportion of first-degree relatives and there issome correlation with the alpha-motor neu-rone abnormalities noted on biopsy. Further,mean creatine kinase levels during a hospitaladmission are also elevated (Meltzer 1980).

Schizophrenics have been shown in severalstudies to present with an excess of majorphysical abnormalities, but the contribution ofmental handicap to the ensuing clinical pictureis unclear. Minor physical and dermatoglyphicabnormalities are overrepresented and a num-ber of these abnormalities are linked to the firsttrimester of foetal development (Guy et al., 1983;Weinberg et al., 2007). These seem unrelated tosymptomatology or to the presence or absenceof neurological soft signs. Dysmorphic craniofa-cial features are also increased in schizophrenia,including brachycephaly (McGrath et al., 2002).

The velo-cardiofacial syndrome (DiGeorgesyndrome or chromosome 22q11.2 deletion


syndrome) is associated with an increasedrisk for schizophrenia and with alterations ofcerebral structure, especially temporal-lobesize (Debbane et al., 2009; Eliez, 2007; Greenet al., 2009; Schaer and Eliez, 2007).

METABOLIC AND BIOCHEMICALFINDINGS

One early finding that has been replicated wasof abnormal nitrogen balance in patients withperiodic catatonia (Gjessing, 1947). Thus, phasesof stupor and excitement were associated withnitrogen retention, with later compensatoryincreased excretion. In such patients it wasclaimed that alteration of nitrogen input, byreducing protein intake, altered the course ofthe illness.

The concept that endogenous neurotoxinsexist, derived from the abnormal metabolismof various chemicals, was encouraged by obser-vations of psychoses resembling schizophreniathat could be provoked by the ingestion of vari-ous psychotropic drugs, such as LSD, mescalineand amphetamine. Attempts were made to iden-tify abnormal compounds in the blood and urineof schizophrenics, with some authors reportingsuccess, although much of the early data wasnot replicated. From these ideas arose thetransmethylation hypothesis (Smythies, 1976).Several hallucinogenic compounds were seen tobe closely related chemically to the endogenouscatecholamines and indolamines, which werenotably methylated derivates. It was suggestedthat abnormal transmethylation of thesecompounds might produce hallucinogenicmethylated monoamines related to the patho-genesis of at least some forms of schizophrenia.Administration of methyl donors such asmethionine to schizophrenics gave variableresults but, especially when combined with amonoamine oxidase inhibitor, led to the exac-erbation of symptoms in some 40% of patients.These studies did not, however, distinguishbetween an acute organic brain syndrome andtrue exacerbation of schizophrenic symptoms(Wyatt et al., 1971).

A related theme is the possibility that 5-HTis involved in schizophrenia. Thus, methionine

may inhibit the uptake of tryptophan into thebrain by competitive blockade (Smythies, 1976)and schizophrenia may in part reflect defectivetryptophan metabolism. However, supplemen-tary tryptophan loading in patients has pro-vided only equivocal results, and to date therole of indolamines in the psychoses remains tobe clarified.

Administration of the 5-HT agonistmetachlorophenylpiperazine (MCPP) toschizophrenic patients increases positive symp-toms (Krystal et al., 1993). In controls there is noeffect, suggesting a propsychotic rather than apsychotogenic effect. These data suggest a rolefor serotonergic mechanisms in schizophrenia.This has been supported by observations thatpipamperone, a selective 5-HT2 antagonist, isantipsychotic, and some newer antipsychoticssuch as risperidone are selective 5-HT2 receptorantagonists (see Chapter 12).

Platelet MAO activity in schizophrenia hasbeen studied by many authors. In general, thevalues are low, although, since similar dataare noted in some other psychiatric conditions,it is not specific for schizophrenia (Delisiet al., 1982; Siever and Coursey, 1985). Patientswith paranoid symptoms probably have thelower levels, and in studies of normal adultsan inverse correlation between the paranoiascore on the MMPI and MAO levels has beennoted. Several authors link low values to hallu-cinations and delusions, others to poor out-come. Baron et al. (1980) identified studentswith abnormally high or low levels of MAOand found that 70% of those with low levelsand none with high levels had a diagnosis ofborderline schizophrenia. This same group alsoreported that schizophrenic patients with a highgenetic load have reduced levels in contrastto those with nonfamilial schizophrenia, andrelatives with low MAO are more often affectedwith schizophrenia-spectrum disorders thanhigh-MAO relatives (Baron et al., 1984).

Reveley et al. (1983) examined MAO in MZtwins discordant for schizophrenia, normal MZand DZ twins, and controls. The MZ pair witha schizophrenic co-twin had significantly lowervalues than did the healthy MZ twins and theother controls, the MZ schizophrenics having


only slightly lower values than their healthyco-twins.

Although some authors have failed to findlow MAO in schizophrenic patients, and fur-ther show that neuroleptic treatment may itselflower MAO (Owen et al., 1981), the consistencyof the data and the findings in twins, rela-tives and volunteers suggests some association,genetically determined, between MAO and vul-nerabilities to these behaviour patterns, and arein agreement with the adoption studies that bothschizophrenia and related spectrum disordersneed to be considered together.

Low-platelet MAO has been implicated inboth the transmethylation and the dopamine(see below) hypotheses of schizophrenia in thesense that reduced MAO activity may reflect animpaired capacity to degrade monoamines ortheir abnormal metabolites, thus increasing theload of these substances in the brain.

The Dopamine Hypothesis

The predominant biochemical hypothesis ofschizophrenia has been the dopamine hypothe-sis, and a considerable amount of research workhas been carried out investigating dopamineand its metabolites in patients. The hypothesisemerged in the wake of the introduction of thephenothiazines in the 1960s, stemming fromrecognition of the similarity of a psychosisinduced by amphetamine to schizophrenia, andthat in small doses they can activate symptomsin patients, and the discovery that the principalmode of action of the phenothiazines is to blockdopamine receptors. The ability of neurolepticdrugs to block dopamine receptors correlatedbetter than other biochemical effects with theirefficacy in the control of psychotic symptoms(Snyder et al., 1974). The dopamine hypothesiswhich emerged thus suggested that schizophre-nia was the result of overactivity of dopamineneurotransmission in the CNS and that neu-roleptics acted by reducing this. However, theevidence was largely circumstantial, and overtime modifications have had to be made.

In animal models, behavioural distur-bances provoked by dopamine agonists givensystemically or by injection into the limbic

forebrain were reversed by neuroleptic drugs(Anden, 1975), and the ability of these com-pounds to inhibit dopamine-sensitive adenylcy-clase stimulation is proportional to their clinicalpotency, with notable exceptions, namely thebutyrophenones. However, the recognitionof different classes of dopamine receptorsand the fact that the main receptor linked toadenylcyclase activity is the D1 have helpedexplain this apparent discrepancy.

Generally the evidence that neuroleptics actat dopamine receptors in the limbic system isstrong (Scatton and Zivkovic, 1984). Electro-physiologically their iontophoretic applicationantagonizes the depression of cellular activityinduced by the application of dopamine ago-nists, and binding studies show that a varietyof neuroleptics specifically bind to dopaminesites within the limbic system. Further, theyincrease the metabolism and turnover ofdopamine, an effect related to the feedback acti-vation of neurones following the postsynapticreceptor blockade.

Early on it was noted that some neurolepticdrugs were less frequently associated withextrapyramidal side effects. The latter werethought to occur with the blockade of dopaminereceptors in the striatum, and the reason whysome drugs, notably thioridazine, clozapine andsulpiride, did not do this required explanation.Some evidence pointed to differential effectsat limbic and nonlimbic sites (ventral anddorsal striatum, respectively), either due toanticholinergic properties (strong with thior-idazine and clozapine, anticholinergic effectsbeing greater in nonlimbic striatal areas) ordue to preferential action on different receptorsubpopulations at the two sites (Scatton andZivkovic, 1984; see also Chapter 12), drugssuch as sulpiride acting mainly at the limbicreceptors. It seems that the effects of neurolepticdrugs on psychotic symptoms relate to changesof dopamine turnover in the limbic forebrain atleast, while the development of extrapyramidalsymptoms is related to the dorsal striatum(Crow et al., 1979a). Further, the antipsychoticeffect most likely relates to blockade of selec-tive dopamine receptors, especially D2, D3and D4.


An effective demonstration of the antipsy-chotic effect of dopamine antagonism has beenprovided by Crow et al. (1979a). Patients withacute schizophrenia took part in a double-blindtrial of treatment with flupenthixol, comparingthe alpha isomer, which blocks the dopaminereceptor, to the beta isomer, which does not.The former had a significant effect comparedto placebo, although this was confined toonly certain symptoms, referred to by theauthors as ‘positive’. Thus, using the Krawieckascale, the improvements were seen on scoresfor delusions, hallucinations and speechincoherence, but not on poverty of speech oraffective flattening.

One problem for the dopamine theoryrelates to the time course of the clinicaleffect. Prolactin, which is elevated by blockingdopamine receptors (the prolactin inhibitoryfactor probably being dopamine), rises muchfaster than the noted clinical improvements,which do not become manifest for two or threeweeks after commencing treatment. In contrast,the extrapyramidal effects often arise within 48hours. This paradox has some explanation interms of adaptive effects to the acute blockadein dopamine systems. Thus, cerebrospinalfluid (CSF) studies indicate that the initialrise of dopamine turnover diminishes aftertime (van Praag, 1977b), as does the plasmalevel of HVA (Pickar et al., 1984), and similartolerance of the biochemical effects has beenobserved in animals. However, it seems that thetolerance relates more to dopamine receptorsin the caudate areas, less to those in the limbicforebrain and not at all to those in the frontalcortex (Scatton and Zivkovic, 1984). The latterfinding may relate to the lack of autoreceptorsin frontal areas. Reuptake of dopamine infrontal cortex is via the NA transporter.

Another line of evidence has been to observethe effects of drugs known to increase dopamineactivity on the symptoms of schizophrenia.Paradoxically, administration of the dopamineagonists L-dopa and apomorphine seems toimprove symptoms in some patients (Buchananet al., 1975; Tamminga et al., 1977). In somereports the effect is on such symptoms as apa-thy, isolation and blunted affect (Gerlach and

Luhdorf, 1975); in others it is on hallucinationsand delusions (Tamminga et al., 1977). How-ever, several carefully conducted trials havefailed to replicate these findings (Ferrieret al., 1984b; Syvalahti et al., 1986). It is sug-gested that any observed effect is due to inhi-bition of dopamine release by interaction withpresynaptic receptors.

More direct evidence of abnormal dopamineactivity in schizophrenia has been hard todemonstrate. If dopamine is overactive in theCNS then low basal levels of prolactin maybe expected, but studies have generally beennegative. In contrast, provocation tests usingapomorphine and measurements of eithergrowth hormone (raised by dopamine ago-nists) or prolactin have been more successful.Growth-hormone responses have been reportedto be greater in patients with first-rank symp-toms (Whalley et al., 1984), with delusions,hallucinations and thought disorder (Meltzeret al., 1984), and to be both increased (Meltzeret al., 1984) and blunted in association with neg-ative symptoms and in chronic schizophrenia(Ferrier et al., 1986). Prolactin baseline levels andapomorphine-induced prolactin suppressionhave been shown to correlate inversely withthe presence of positive symptoms (Johnstoneet al., 1977; Ferrier et al., 1984b), especiallyin patients with normal ventricular size onCT scans (Kleinman et al., 1982b). Althoughthe interpretation of such findings may beeasily blurred by the effects of medicationon hormonal release, these studies were ondrug-free patients, and the results tend tosupport the hypothesis that at least somesymptoms in schizophrenia are related todopaminergic mechanisms, notably hallucina-tions, delusions and thought disorder, similarto the ones shown to respond preferentially todopamine-antagonist drugs.

Other estimates of dopamine activity haverelated to peripheral markers. Plasma dopaminehas been reported to be elevated in some studies(Bondy et al., 1984). Measurement of dopaminebeta-hydrolase, the enzyme that convertsdopamine to noradrenaline, has providedinconsistent results (Castellani et al., 1982),as have measurements of plasma HVA.


Other markers have included the reporting ofreduced platelet cyclic-AMP production (Kafkaet al., 1979) and elevated spiperone binding tolymphocytes apparently unrelated to treatment(Bondy et al., 1984). Further evidence that linksdopamine to schizophrenia comes from CSFand post-mortem studies, and is discussedbelow (see Table 7.3).

PET ligand-binding studies have mainlyused tracers that bind to one or more of thedopamine receptors. There is discrepancyof results, partly based on technical factorssuch as the ligand used. There are morenegative findings with normal D2 densities inschizophrenia (Martinot, 1990; Farde et al., 1990;Pilowsky et al., 1995) than positive ones (Tuneet al., 1992), failing to provide direct evidencefor increased dopamine-receptor activity inschizophrenia. Fletcher et al. (1996) have usedapomorphine as a dopamine agonist to showthat in schizophrenia, using a verbal fluencyactivation task, there is impaired activationin the cingulate cortex, which is enhanced byapomorphine. Their data suggest a regionalfunctional deficit in schizophrenia, which canbe modulated by dopamine. Research continueswith dopamine PET imaging, with some studies

Table 7.3 Main arguments for and against thedopamine hypothesis

For:

1. Amphetamine-induced psychosis produces aschizophrenia-like illness

2. Amphetamine exacerbates schizophrenia3. Drugs that block dopamine receptors are antipsy-

chotic in relationship to their blocking potential4. Post-mortem and endocrine studies5. Animal models of dopamine agonism resem-

ble psychosis, and are reversed by dopamine-blocking agents

Against:

1. Time course between administration of adopamine blocker and clinical response

2. Dopamine agonists have an inconsistent effecton psychosis

3. Direct evidence of increased dopamine activityis poor

suggesting an important role for dopamine inschizophrenia (Howes et al., 2009; Howes andKapur, 2009) and others not (Valli et al., 2008).

Catechol-O-methyl-Transferase (COMT) isone of several enzymes that degrade thecatecholamines, such as dopamine or nora-drenaline. A series of studies have now fairlyconsistently found that individuals with certainCOMT gene variants (specifically the valvariant) carry an increased risk of developingschizophrenia, and the val variant of theCOMT gene is associated with both structuraland functional brain changes within patientswith schizophrenia (van Haren et al., 2008;Lawrie et al., 2008a; Ross et al., 2006; Roffmanet al., 2006). These findings raise suspicion thatan abnormality in catecholamine metabolismis fundamental to schizophrenia pathogenesis,at least in some subjects (Gur et al., 2007;Benes, 2007; Keshavan et al., 2007; van Harenet al., 2008).

The Noradrenergic Hypothesis

An alternative to the dopamine hypothesis isthe noradrenergic hypothesis, suggested byStein and Wise (1971) and revived recently byYamamoto and Hornykiewicz (2004). This wasbased on the idea that the central noradrenergicpathways, then shown to be related to a centralreward system, might degenerate and lead to atleast some of the symptoms of schizophrenia.Generally the data are less favourable for thishypothesis, although several studies indicatea role of noradrenergic mechanisms in somesymptoms of schizophrenia. This has beenreviewed in detail (Kleinman et al., 1985).No noradrenergic abnormalities have beenreported in urine of schizophrenic patients,although there is one report of increased levelsin the plasma of drug-free patients (Ackenheilet al., 1979). Neither plasma nor urine 3-methoxy-4-hydroxyphenylglycol (MHPG) orvanillylmandelic acid (VMA) levels have beenconsistently abnormal in schizophrenia, moststudies showing no difference from controls.CSF and neurochemical data are discussedbelow. Attempts to assess the functions ofalpha-2 adrenergic receptors have been done


either by direct measurement of plateletbinding or by changes of metabolites afterchallenge with clonidine. Alpha-2 receptorshave been reported as increased in plateletsin only a minority of subjects, probably thosewho are schizoaffective (Kafka et al., 1980),while others have noted decreased binding ina group with negative symptoms and poorresponse to treatment (Rosen et al., 1985), withsubsensitivity of MHPG release to a clonidinechallenge (Sternberg et al., 1982).

The Glutamate Hypothesis

It has been known for some time that phen-cyclidine and ketamine, drugs which areantagonists at the NMDA receptor, increasesymptoms in schizophrenia. Glutamate isubiquitous in cerebral cortex, and is the mainexcitatory neurotransmitter regulating activityin the cortical-basal ganglia thalamic re-entrantcircuits. This relates not only to motor controlbut also to sensory gating. The NMDA receptorsare therefore hypothesized to be hypofunctionalin schizophrenia. Serum glutamate levels arelow at the onset of the disorder (Palominoet al., 2007), as are serine levels in serum andCSF, serine being modulatory at the NMDAreceptor. Attempts to use this hypothesis todevelop new treatment strategies include trialsof the glutamate agonist D-serine, which isreleased by glutamate from astrocytes. Patho-logical studies in support of this hypothesis arediscussed below.

Other Studies

Early investigations of endocrine function inschizophrenia yielded inconsistent results andwere aimed in particular at examination of thethyroid and adrenal glands. Schizophrenia-likepsychoses have been described in associationwith several endocrinopathies, including hypo-and hyperthyroidism and hypoparathyroidism(Lishman, 1987). The prolactin and growth-hormone data, utilized more as a methodof challenging hypothalamic–pituitary (HPAaxis) function than to implicate the hormonal

changes themselves as pathogenic, have beendiscussed. Few other studies of endocrinefunction in schizophrenia have been carried outrecently, although there are reports of lower LHand FSH levels in chronic schizophrenia, withan inverse relationship to positive symptoms,while negative symptoms have been inverselyrelated to testosterone levels (Alias, 2008).There seems to be loss of normal episodicsecretion of LH, possibly reflecting disturbedhypothalamic or limbic-system function (Ferrieret al., 1982). Further evidence for this is thefinding of blunted FSH and prolactin responsesto thyrotropin-releasing hormone in similarpatients (Ferrier et al., 1983).

In contrast to what is seen in the affectivedisorders, the number of schizophrenic patientsshowing abnormal dexamethasone suppressionis low, unless there is coexistent depression(Munro et al., 1984).

Finally, following the discovery of enkeph-alins, attempts have been made to examinepeptide function in schizophrenia, largely bygiving peptides to patients and estimating clin-ical changes. Table 7.4 lists some of the sub-stances tried to date, but the results have beendisappointing.

Of growing importance has been thecannabinoid group of compounds, especiallythe social use of cannabis. It is acknowledgedthat the strength and the availability of cannabispreparations has increased, and it is knownthat patients with psychotic disorders are morelikely to abuse illicit drugs. Schizophrenicpatients are often heavy cannabis users, and

Table 7.4 Some peptides used as treatments inschizophrenia

Peptide Author

CCK Lostra et al. (1984)Albus et al. (1984)

Des-tyrosinegamma-endorphin

van Praag et al. (1982)

Des-enkephalingamma-endorphin

van Praag et al. (1982)

Beta-endorphin Berger et al. (1980b)Thyroid-releasing

hormonePrange et al. (1979)


the administration of tetrahydrocannabinolincreases the reporting of psychotic symp-toms in healthy volunteers and schizophrenicpatients in remission (Henquet et al., 2008;Tandon et al., 2008; d’Souza et al., 2005).

Several long-term follow-up studies havebeen reported in which populations have beenexamined over several years and their drugusage noticed or monitored. A Swedish cohortof over 50 000 conscripts noted that cannabisuse by the age of 18 was associated with a 2.4-fold increase in a diagnosis of schizophrenia(Andreasson et al., 1987). A longer follow-up ofthe same population (27 years) reported a doserelationship between the amount of cannabisuse at entrance and the later risk of developingthe disorder (Zammit et al., 2002). A Netherlandsstudy (three-year follow-up of nearly 5000 peo-ple) confirmed these findings and noted thatthose who had reported any psychotic symp-toms at baseline were more likely to go onto develop schizophrenia if they were cannabisusers (van Os et al., 2002). Arseneault et al. (2002)followed 759 people in a cohort from birthto young adulthood. Cannabis use by age 15increased the risk of developing psychosis. In ameta-analysis of the data, the increased risk ofdeveloping a psychosis with cannabis use wasgiven as 1.4 (odds ratio), and the dose responsewas confirmed (odds ratio 2.84). Most stud-ies in this area had adjusted for confoundingfactors such as the use of other drugs and psy-chotic symptoms at baseline (Moore et al., 2007).One problem with interpreting these results ina causal way is that evidence that the inci-dence of schizophrenia is increasing pari passuis not yet reliably available. However, elevatedanandamide (a cannabinoid agonist) has beenreported in CSF (Leweke et al., 2007), as hasincreased density of CB1 receptors in prefrontalcortex (Dean et al., 2001).

NEUROCHEMICAL INVESTIGATIONS

In order to explore further the dopamineand related neurochemical hypotheses inschizophrenia, many neurochemical investiga-tions have been conducted. The CSF data will

be discussed first, followed by neurochemicalpathological data.

CSF

The main metabolite assessed in CSF to test thedopamine hypothesis is HVA, although it is sug-gested that the main source of this is nigrostriataldopamine and not mesolimbic and mesocorticaldopamine. Nonetheless, the anatomical prox-imity and the overlap between these systemssuggests that at least some of it reflects limbic-system dopamine function. In general, eitherthe baseline CSF values or the accumulationof metabolite after probenecid administrationhave been sampled. The latter technique inhibitsthe exit of HVA and related compounds fromthe CSF, diminishing the gradient that existsbetween ventricular and spinal CSF values.Theoretically this should provide a more accu-rate determination of metabolite turnover andconcentration, but it too has methodological dif-ficulties. Not the least is that the results arerelated to the CSF probenecid levels.

One of the early baseline studies of HVA wasthat of Rimon et al. (1971), who reported normalvalues, except that patients with paranoidsymptoms had elevated levels. Generally, how-ever, the baseline data have failed to show dif-ferences between schizophrenics and controls,although Sedvall and Wode-Helgodt (1980)reported that high HVA levels were associatedwith a family history of schizophrenia. Theprobenecid studies likewise have tended tobe negative with singular exceptions. Thereare reports of decreased accumulation inpatients with Schneiderian first-rank symptoms(Bowers, 1973; Post et al., 1975) and a suggestionthat those with low values have poorer progno-sis (Bowers, 1974). These data are compatiblewith increased dopamine-receptor sensitivity,which subsequently decreases dopamineturnover. Some support for this comes froma study in which the relationship betweenapomorphine-induced growth-hormone releaseand CSF–HVA levels was assessed in drug-freeschizophrenic patients. An inverse relationshipwas found, those patients with low HVAhaving a higher release of hormone (Zemlan


et al., 1985). Thus low HVA was associatedwith higher dopamine-receptor activity, and inaddition with first-rank symptoms and thoughtdisorder. Pickar et al. (1990) also reportednegative correlations between HVA ratings ofpsychosis and positive symptoms.

Both CSF levels of DOPAC and dopamineitself are reported as normal in schizophrenia(Berger et al., 1980a; Gattaz et al., 1983b), as areprolactin levels (Rimon et al., 1981).

With regard to the noradrenergic hypothe-sis, MHPG levels are consistently reported asnormal (Kleinman et al., 1985), although thereare reports of increased levels of noradrenaline(Gomes et al., 1980; Lake et al., 1980), especiallyin paranoid subtypes, and of MHPG correlatingwith negative symptoms (Pickar et al., 1990).

A variety of other substances have been mea-sured in the CSF of schizophrenic patients, someof the more important ones being shown inTable 7.5.

Brain Neurochemistry

Most studies in this area relate to testing thedopamine hypothesis. Studies on post-mortembrains are difficult to interpret. Not only arethey derived from patients that have usuallybeen on medication or have used a large dose of

psychotropic drugs to kill themselves, but alsorapid post-mortem biochemical changes occur.Several groups have set up centres designedspecifically for collection of brains and appropri-ate storage after death, control specimens beingtreated in the same manner. It seems likely thatpeptides and receptors are much more stablethan monoamines and their metabolites. Gen-erally the areas of brain examined have beenthose with a high level of the neurotransmitterunder investigation, and studies have mainlyconcentrated on the limbic system and relatedstructures. For all of these reasons, ligand PETimaging is somewhat better, although it also hasproblems, primarily that not all substances, par-ticularly peptides, can be measured with PET.

The majority of post-mortem neurochem-istry studies show normal values for bothdopamine concentrations and HVA, notablyin the caudate nucleus, putamen and nucleusaccumbens. However, Bird et al. (1979) reportedincreased concentrations of dopamine in thenucleus accumbens and the anterior perfo-rated substance. In an extended study onover 50 schizophrenic brains, with controls,Mackay et al. (1982) confirmed the significantelevations, especially in patients with earlyonset of illness. Owen et al. (1978) reportedincreases in dopamine in the caudate nucleus,but not in the putamen or nucleus accumbens,

Table 7.5 Some CSF studies in schizophrenia

Compound Result Reference

Dopamine beta-hydrolase Normal Lerner et al. (1978)Dopamine sulphate Increased: negative symptoms Risby et al. (1993)5-HIAA Normal Post et al. (1975)

Gomes et al. (1980)Roy et al. (1985)

Cyclic-AMP Decreased Biedermann et al. (1977)Increased Gomes et al. (1980)

Cyclic-GMP Decreased Gattaz et al. (1983a)GABA Normal (↓ young females) van Kammen et al. (1982)Angiotensin-converting enzyme Decreased Beckmann et al. (1984)Neurotensin Decreased (subgroup only) Widerlov et al. (1982)Glutathione Decreased: also in prefrontal

cortex (MRS)Do et al. (2000)

Kynurenic acid Decreased, correlated withgrey-matter densities

Schroeder et al. (2008)

Copper Decreased Tyrer et al. (1979)D-serine Decreased Bebdikov et al. (2007)


and a significant decrease in HVA in thecaudate. Farley et al. (1977) reported increasesin dopamine only in the septal region of chronicparanoid patients. The cingulate has beenrarely examined, but Wyatt et al. (1995) havereported decreased DOPAC levels in anteriorcingulate cortex.

Reynolds (1983, 1987) has attempted to assesslaterality effects in relation to neurochemistry inschizophrenia. Comparing patient brains to con-trols, dopamine was increased in the amygdalaof the schizophrenics, selective for the left side.In one study, the increase of dopamine in theamygdala was found to correlate with decreas-ing levels of GABA binding in the hippocampus(Reynolds et al., 1990).

Dopamine-receptor binding, with suchcompounds as haloperidol, spiperone, flu-penthixol and apomorphine, has also beeninvestigated. Studies of the dopamine receptorin vivo with PET imaging have shown high inter-individual variability, which links to person-ality variables (Farde et al., 1990; Karlssonet al., 1995; Karlsson et al., 2002) and to cognition(Cervenka et al., 2008). Most of the studies inschizophrenia show increased binding in someareas, although the interpretation of this has ledto much debate. In summary, most studies of D2receptors show increases of caudate binding (13of 16) (Clardy et al., 1994), others show increasesin the putamen and/or nucleus accumbens.The contribution of prescribed neurolepticdrugs to these findings continues to be animportant controversy, although some authorsexamined groups of drug-naıve patients (Owenet al., 1978) or patients drug-free for a long time(Lee and Seeman, 1980). However, these verysame changes can be observed in rat brainsgiven long-term neuroleptics (Clow et al., 1979).

Other dopamine receptors have been lessstudied. D1 receptors seem normal (Seemanet al., 1987), similar to data from animals receiv-ing neuroleptic drugs. However, D4 bindingin the striatum has been reported as increased(Seeman et al., 1993).

It remains unclear what it is that drives theincrease in dopamine activity. Early work hadsuggested a link through GABA. Decreased ter-minals of GABA in the hippocampus correlate

with increased dopamine levels and are linkedto the observation of increased GABAa-receptoractivity. Attention has now centred on asubgroup of GABA neurones, the parvalbumin-containing cells. Parvalbumin, along withcalretinin and calbindin, is a calcium-bindingprotein, which develops at between three andsix months of foetal development. Decreasedparvalbumin cells are found in the frontalcortex and hippocampus in schizophrenia. Thisall suggests GABAergic deficits, which it ispostulated lead to disturbed neuronal rhythmic(oscillatory) expression in key structures (thehippocampus, but also areas of frontal cortex),and to altered dopamine release.

Glutamate is also linked with this story. Glu-tamate input to dopamine-releasing cells (from,for example, the hippocampus and the pedun-culopontine areas) normally induces bursts ofspikes within the VTA, which are disrupted bya deficit of glutamate. Glutamate can be identi-fied within neurones by looking at transporters(EAAT3 or VGluT1 – the neuronal and vesicu-lar transporter respectively), and recent studieshave shown decreases of these in post-mortemmaterial, especially in the left-temporal area(Nudmamud-Thanoi et al., 2007; Reynolds andHarte, 2007). These are associated behaviourallyto novelty detection, which under normal cir-cumstances activates the VTA and accumbens,interlinked with behavioural reward, and whichin schizophrenia is thought to become disrupted(Lodge et al., 2009; Sohal et al., 2009).

Glutamic acid decarboxylase levels arereported as decreased in prefrontal cortx(Volk et al., 2000). The changes in GABAare supported by decreases in reelin mRNA,reelin being a protein released by GABAergicneurones, and decreased GABA transportersin DLPFC neurone (Costa et al., 2001; Akbarianet al., 1995).

Several studies of glutamate binding havebeen carried out. Measurement of glutamatelevels in CSF had earlier yielded conflictingresults, but the psychotomimetic agent phen-cyclidine (PCP) was known to be an NMDA-receptor antagonist, leading to the hypothesisthat glutamate transmission may be reducedin schizophrenia. Decreased kainate and


glutamate binding have been reported (Kerwinet al., 1988; Deakin et al., 1989), especiallyin the left temporal lobes, as has reducedrelease of both glutamate and GABA fromsynaptosomes from temporal cortex followingapplication of the agonists kainic acid or NMDA(Sherman et al., 1991). Data from frontal areasare equivocal, especially for GABAa receptors,although they suggest, if anything, increasedglutamate binding in orbitofrontal areas. Thesedata argue for decreased cortical excitatorytransmission in some areas of the limbic systemin schizophrenia.

Noradrenaline and its metabolites have al-so been examined. Decreased levels of nora-drenaline have been reported in the putamen(Crow et al., 1979b); others have reportedno differences (Bird et al., 1979; Winbladet al., 1979) and others selective increases (Farleyet al., 1977). MHPG is reported to be elevated inthe hypothalamus and nucleus accumbens byone group (Kleinman et al., 1982a). Dopaminebeta-hydrolase and monoamine oxidase appearnormal.

Binding of 5-HT receptors to various brainregions has been variable. Low 5-HT2 bindingto frontal cortex has been noted by some groups,especially if patients who commit suicide areexcluded from the sample (Laurelle et al., 1993),

although the contribution of chronic neuroleptictreatment to these effects remains unclear.

Peptides have also been studied. Levels ofangiotensin-converting enzyme were reducedin the substantia nigra and globus pallidus(Arregui et al., 1980). Opioid and naloxonebinding is probably normal (Owen et al., 1985).Ferrier et al. (1984a), in an extensive studyof peptide distribution in limbic structuresin schizophrenic and control brains, reportedthat CCK was reduced in temporal cortex,especially in the hippocampus and amygdalaof patients with negative symptoms (type 2 ofCrow); somatostatin was likewise decreased inhippocampus in the same group, while VIP wasincreased in the amygdala of those with positivesymptoms (type 1 of Crow). No changes wereseen for neurotensin, while substance P wasincreased in the hippocampus. A fuller review ofthe post-mortem findings of peptides in schizo-phrenia is given by Caceda et al. (2006, 2007).

Several other compounds have been assessedin various brain regions, and many of these aresummarized in Table 7.6.

Neuropathological Data

A related area of research has been neuropatho-logical studies of the brains of patients who had

Table 7.6 Some post-mortem biochemical findings

Chemical Result Author

Dopamine Increased D2, D3, D4/transporter normal

Review, see Guillin et al. (2007)

GABA Probably normal Cross et al. (1979)Glutamic acid decarboxylase Normal Bird et al. (1979)5-HT Increased: Putamen Crow et al. (1979a)

5-HT1a frontal cortex Gurevich and Joyce (1997)Abi-dargham (2007)

5-HT2 probably normalLower:Hypothalamus Winblad et al. (1979)Medulla oblongataHippocampusFrontal cortex Bennett et al. (1979)

Tryptophan Normal Crow et al. (1979b)Cathechol-O-methyl transferase Normal Crow et al. (1979b)Tyrosine hydrolase Normal Crow et al. (1979b)Choline acetyl transferase Normal Bird et al. (1979)Muscarinic receptors Decreased Dean et al. (2002)Neurotrophins BDNF Variable results Buckley et al. (2007)


‘the schizophrenias’. Although many studiesover the years have been carried out, all ofthem detecting changes in some brains exam-ined, such data are often criticized on account ofthe lack of consistency of the findings. However,similar criticisms would apply to pathologicalstudies of such disorders as the dementias orthe epilepsies if it were supposed that they rep-resented the outcome of only a single process.

Earlier studies had shown abnormalitiesin thalamic, hypothalamic, periventricular,periaqueductal and basal forebrain areas(Nieto and Escobar, 1972; Stevens, 1982). Sincethat time other investigators have examinedpost-mortem collections of brains and whilemost detect some abnormalities, there aredifferences between types of pathology andareas affected. The most consistent abnormal-ities have been noted in hippocampal andrelated structures.

Overall, the volumes of schizophrenic brainsare reduced (Pakkenberg, 1987). This is asso-ciated with negative symptoms (Johnstoneet al., 1994), and an excess of nonspecific focalpathology can be seen, often in basal-gangliastructures (Bruton et al., 1990). The length ofthe sylvian fissure on the left is decreased(Falkai, 1992).

There are reports of abnormal gyral pat-terns, especially in the frontotemporal areas,with associated cytoarchitectural abnormalities(Jacob and Beckman, 1989) confirmed with MRIstructural imaging.

Structural disorganization of hippocampalpyramidal cells (Kovelman and Scheibel, 1984;Bogerts et al., 1985), reduced hippocampalmossy cell-fibre staining (Goldsmith andJoyce, 1995), decreased hippocampal volumesand cell loss (Falkai and Bogerts, 1986) andabnormalities or reduced volumes of thesubiculum and hippocampal CA regions haveall been reported (Arnold et al., 1995). In theentorhinal cortex, smaller neurones, displacedpre-alpha and pre-beta cells, and reducedneuronal number and density have been noted(Arnold et al., 1991). The changes reported arein both the left (Kovelman and Scheibel, 1984;Bogerts et al., 1985) and the right hemispheres(Conrad et al., 1991).

Crow and colleagues (Brown et al., 1986),in a comparison of brains from patientsmeeting strict criteria for affective disorderand schizophrenia, noted that the latter hadlarger temporal horns of the lateral ventricleand thinner parahippocampal gyri. The greaterdifferences were noted in the left hemisphere.

Some of these reports suggested thatgliosis was a typical pathological reaction inschizophrenic brains. However, most studiesfail to find gliosis, supporting the sugges-tions of Kovelman and Scheibel (1984) thatthe pathological changes instead representdevelopmental anomalies.

Other neuropathological findings includediminished neuronal density in some layersof the frontal, cingulate and motor cortices(Benes et al., 1986), higher numbers of longvertical axones in cingulate cortex, possiblyrepresenting increased input of associativeaxons (Benes et al., 1987), and abnormalarrangement of cells and decreases of smallinterneurones (possibly inhibitory) in cingulatecortex (Benes and Bird, 1987; Benes et al., 1991).Increased neuronal density has been reportedin hippocampus and prefrontal cortex (Selemonet al., 1995), while reduced dendritic spinedensity of pyramidal neurone in the DLPFC hasbeen found. These findings add evidence to thegrowing appreciation that the schizophreniasmay be neurodevelopmental disorders, andthat abnormalities in cortical cytoarchitextureand connectivity may be key facets of theunderlying pathophysiology.

Akbarian et al. (1993) reported reducednumbers of NADPH-diaphorase-stained cellsin superficial and increased numbers in deeplayers of the dorsolateral prefrontal cortex.The stain specifically identifies cells of theembryological subplate area, the findingssuggesting anomalous migration of cells. Inanother study, decreased gene expression forglutamic acid decarboxylase in prefrontal cor-tical neurones was found. This occurred in theabsence of any cell loss, suggesting a functionaldown-regulation of neurotransmitter geneexpression (Akbarian et al., 1995). This mayrelate to the reported decrease in glutamic aciddecarboxylase levels in prefrontal cortex and


decreased GABA axon terminals. The changesin GABA are supported by decreases in reelinmRNA. As with many studies, the contributionof medication to these effects is unknown.

There are reported decreases of neuronesin medial dorsal nucleus of the thalamus andnucleus accumbens (Pakkenberg, 1990), sub-stantia innominata (Stevens, 1982) and cerebel-lum (Jeste et al., 1985). The thalamic findingsare of interest in view of the increasing evi-dence from imaging studies of thalamic abnor-malities in schizophrenia. Some evidence hasemerged for abnormalities of glutamate recep-tors (NMDA) and transporters in this structure,but also in other areas such as the caudate, hip-pocampus, DLPFC and cingulate gyrus (Watiset al., 2008).

A selective reduction in glial cells in cingulatecortex (Brodmann’s area 24) has been shown inschizophrenia patients relative to comparisonsubjects, calling into question whether cingulatechanges are specific only for affective disorders(Stark et al., 2004).

Recently, direct measurement of proteinexpression in brain areas has been carriedout (proteomic studies) using electrophoresisor specialized mass-spectrometry techniques.These can identify the profile of many hundredsof proteins, but to date have not yieldeduseful clinical information. Likewise there area growing number of studies in which geneexpression in various cortical areas has beenlooked at, again with heterogenous results.The frontal cortex and genes that link withneurotransmitter function have been moststudied. Genes related to myelin-formingoligodendrocytes and metabolic enzymes,as well as synaptic marker proteins such assynaptophysin, spinophilin (markers at pre-and post-synapyic terminals, respectively) anddysbindin, have all been reported to be reducedin some areas, notably hippocampus (Weickertet al., 2007). The mRNA that encodes GAD67 (glutamic acid decarboxylase), an enzymethat synthesizes GABA, is reduced in theDLPFC. This is reported particularly for theparvalbumin-expressing neurones, which havebeen shown in several studies to be reduced

in schizophrenia in the hippocampus and pre-frontal cortex.

While the pathological studies achieve somedegree of agreement, there is also quite a bit ofconfusion. This relates to several factors. Agonalchanges, prior drug prescriptions and statisticalanalysis of such databases are problematic. Itis also necessary to take into account the meth-ods selected for analysis and the heterogenouspopulation. There seems to be evidence for cellloss or disarray, or both, in frontal and tempo-ral areas, but also for alterations of cell density,which may better reflect cell size rather thanloss. However, the abnormalities are relatedto the structures implicated in the anatomy ofschizophrenia, especially the prefrontal areas,medial temporal (especially hippocampal struc-tures) and thalamus. Glutamate and GABAsystems are becoming very relevant as attentionhas been directed to underlying anatomical cir-cuitry. The enthusiasm for post-mortem studieshas waned in recent years, overtaken partic-ularly by the advances in brain imaging, butnewer methods of protein and gene analysis arealso ascendant and may lead to important find-ings, especially if they can be coordinated withimaging and genetic data.

NEUROPHYSIOLOGICALAND NEUROLOGICAL DATA

EEG

Soon after the introduction of the electroen-cephalogram, reports of abnormalities inschizophrenia appeared. Hill (1950) reviewedthese as follows: EEGs suggestive of epilepsywere most often seen in catatonia, althoughgeneralized nonparoxysmal dysrhythmias andother paroxysmal phenomena were also found.Since then, there have been many reports ofEEG abnormalities in schizophrenic patients.In a comparison with patients suffering fromaffective disorder, Abrahams and Taylor (1979)noted that the proportion of EEG abnormalitieswas twice as great in the schizophrenics, whohad more temporal abnormalities. Slow-waveasymmetries, slow bursts, spikes and sharp


waves were the recorded changes, and theytended to be more frequent on the left side. Inpatients with temporal-lobe EEG changes thislaterality was associated with formal thoughtdisorder and emotional blunting, but notfirst-rank symptoms. Although patients wereon medications, this did not correlate with theEEG findings.

Investigations of patients using electrodesimplanted within the brain have clearlydemonstrated abnormalities in schizophrenics,again notably within the limbic system. In aseries of studies, Heath (1982) investigated 63patients with psychosis, 38 being diagnosed asschizophrenic. Spiking was seen in the septalregion. This is defined by Heath as an area bor-dered dorsally by the base of the anterior horn ofthe lateral ventricle and the head of the caudatenucleus, and ventrally by the free surface of thegyrus rectus of the frontal lobe. It includes septalnuclei, the accumbens, olfactory tubercle andthe diagonal band, as well as parts of the gyrusrectus. It has extensive frontal and temporal con-nections. The abnormalities occurred when, andonly when, the patients were actively psychotic.In many cases changes were not observed onsurface recordings, and neither were similarfindings noted in chronic-pain control patients.Violence and aggression were associated withhippocampal and amygdala discharges. Somepatients showed activation in sensory-relaynuclei in association with hallucinations (forexample, medial geniculate recordings) andoften cerebellar discharges. Similar findings,especially with regards to the septal region,have been reported by others (Rickles, 1969).

Examination of the spectral power of EEG fre-quencies in schizophrenia reveals more fast betaactivity, more theta and delta, and less alphathan normals (Itil, 1975). Psychotic childrenand those with a high risk for schizophreniashow a similar pattern of changes. Generallythese patterns are ‘normalized’ by antipsychoticmedication.

Fenton et al. (1980) noted the changes to bemaximal over temporal derivations in acuteschizophrenic patients, while chronic patientsshowed more diffuse slow-wave power, notably

theta and delta, diffusely or maximally fron-toparietally.

Stevens and Livermore (1982) used teleme-tered EEG recordings during psychoticbehaviour of schizophrenic patients. They iden-tified so-called ‘ramp patterns’, characterizedby a monotonic decline in power from lowestto highest frequencies, which can be seenin epileptic patients during subcortical spikeactivity. These were seen only in recordingsfrom schizophrenic patients in their sample,and not in controls. Patients with catatonicepisodes showed them in the right temporalregion, while 50% of paranoid patients withauditory hallucinations had left-sided ramps,with increased slow activity. Psychotic eventsrecorded clinically were associated withsuppression of left temporal alpha frequencies.

Evoked-potential studies in schizophreniaare numerous, and the results complicatedto interpret. Several reviews are available(Shagass, 1972; Flor-Henry, 1983). Generally,evoked potentials show greater variabilityand amplitudes are decreased; a subgroup ofpatients show a ‘reducing’ pattern of decreasingamplitude with increasing stimulus intensity.Flor-Henry (1983) summarized a review bystating ‘the evoked potential characteristics inschizophrenia consistently implicate the lefthemisphere, and particularly the left temporalregion’ (p. 215). The contingent negative varia-tion (CNV) also tends to be reduced, maximallyover frontal areas.

Studies using brain electrical activity map-ping (BEAM) EEG have shown increasedfrontal delta and diminished left temporalP300 amplitudes (Morihisa et al., 1983; Morstynet al., 1983).

More recent studies have investigatedrhythmic oscillations and synchronizationsbetween neural networks. Such oscillationsundergo changes during development and areclosely related to GABA and glutamate activity.As noted, there are suggestions of decreasedGABA neurones in cortical structures inschizophrenia, especially those neurone whichexpress parvalbumin (chandelier neurones)with thalamic afferents. Alteration of this input


can disrupt inhibition and hence oscillatorypatterns (Lodge et al., 2009; Sohal et al., 2009).

Using MEG it has been shown that responsesrecorded from auditory cortex in schizophreniato rhythmic auditory trains or to visual stimuliare different from controls, with decreases ofhigher frequencies (40 Hz, gamma) and excessof lower ones (20 Hz, beta). Gamma and betaoscillations have different source generators,the latter coming from deeper cerebral areas.This disruption of gamma oscillations inschizophrenia is in keeping with theoriesof disruption of GABA/glutamate activity,and with disturbed cognition, especially withworking memory (Rotarka-Jagiela et al., 2009;Haenschel et al., 2009; Haenschel et al., 2007).

Radiological Studies

A common saying throughout the 1900s wasthat ‘schizophrenia was the graveyard ofneuropathologists’. That is, numerous neu-ropathologists had worked for many years tofind neuropathological evidence correlatingwith the behavioural changes observed by pa-tients and clinicians, all to no avail. A modernupdate of that saying might be that ‘schizophre-nia is the graveyard of neuroimagers’. However,what is different with the modern methods isthat many, in fact most, published imagingstudies do in fact find differences betweenpatients with schizophrenia and controls.Nonetheless, often the studies are methodologi-cally flawed or are of small sample size and theresults do not replicate.

MRI

The earlier findings with CT are reviewed inprevious editions of this book, but remain bothinteresting and valid. MRI has become the pre-dominant research method. There are well over1000 studies, and they have served to confirmmuch of the earlier CT data, but in addition haveprovided new findings. A summary is given inTable 7.8.

With MRI it has been possible to confirm thatthe cerebral cortex is smaller than in controls

Table 7.7 Some features of ventricular enlargementof schizophrenia

Early brain damage or birth complicationPoorer premorbid adjustmentMore negative symptomsLess positive symptomsCognitive declinePoorer response to treatmentMore minor neurological signsDecreased CSF–HVADecreased frontal CBF

(Zipursky et al., 1992), even at the onset of theillness (Degreef et al., 1991). The degree of losscorrelates with the severity (Young et al., 1991)and duration of the schizophrenic symptoms(Waddington et al., 1991). Ventricular enlarge-ment has been confirmed in many reports, againat the onset of the illness (Degreef et al., 1991).The relationship of this to positive and negativesymptoms is unclear (Table 7.7).

One of the significant advantages of MRIis the production of high-resolution images oftemporal-lobe structures. Reduced temporal-lobe volume and increased size of thetemporal horn of the lateral ventricle is a well-replicated finding (Suddath et al., 1989;Kawasaki et al., 1993). Smaller superiortemporal-lobe gyri on the left have been notedby several authors, with correlations to clinicalphenomena, namely auditory hallucinations(Barta et al., 1990), thought disorder (Shentonet al., 1992) and reduced P300 amplitude(McCarley et al., 1993). These findings are ofparticular interest since the superior temporalgyrus contains the planum temporale andHeschl’s gyrus (primary auditory cortex) andmerges into Wernicke’s area, and is thereforeintimately involved with language and themanipulation of symbolic knowledge.

Suddath et al. (1990) studied 15 MZ twinsdiscordant for schizophrenia, and found theaffected twin to show larger lateral and thirdventricles, and smaller hippocampi with de-creased grey-matter volume, an effect greater onthe left. No such differences were found in nor-mal twin pairs. In a further study of twin pairs,they noted a significant relationship between theleft hippocampal volume, and rCBF activation


Table 7.8 Summary of MRI findings in schizophrenia

Region of interest Findings

Cerebrum Volume loss in cerebral cortex and abnormal relaxation timesCerebellum InconclusiveVentricles Enlarged lateral and third ventriclesTemporal lobes Abnormal relaxation times and volume loss in grey matterSuperior temporal gyrus Smaller – especially on the leftMesial temporal structures Abnormal relaxation times and smaller

amygdala–hippocampus–parahippocampus in particular the left side;associations between hippocampal size and prefrontal cortex activity;reduced NAA in left hippocampus with MRS

Frontal lobes Smaller and abnormal relaxation times; decreased phospholipid turnoverCaudate nucleus Various reports, smaller and largerThalamus SmallerSeptum pellucidum Enlarged and higher incidence of cavum septum pellucidumCorpus callosum Thinner/thicker/longer/shape distortion or normalCingulate gyrus Smaller on both sidesGender effects Brain abnormalities more in malesProgression Ventricular enlargement – nonprogressive > progressive

Brain volume loss – cortical thinningClinical correlation Positive symptoms: mesial temporal structures and superior temporal gyrus

Negative symptoms: ventricular enlargement and frontal lobe abnormalities

in the prefrontal cortex with the Wisconsin card-sorting test with differences between healthyand affected twins. In the affected twins, pre-frontal activation was correlated to both rightand left hippocampal volumes.

Individual structures in the temporal lobehave also been examined. The hippocampi havebeen reported smaller, a finding that is morerobust on the left (Breier et al., 1992; Weinbergeret al., 1992; Bogerts et al., 1993), and in onestudy correlated with the severity of positivesymptoms (Bogerts et al., 1993). The amygdala,a much more difficult structure to measurethan the hippocampus, is also probably reducedin size on the left (Barta et al., 1990; Shentonet al., 1992) or bilaterally (Breier et al., 1992). Theparahippocampal gyrus is also smaller on theleft (Shenton et al., 1992; Kawasaki et al., 1993)or bilaterally (McCarley et al., 1993).

The frontal cortex has also been investigated.Abnormal relaxation times of both T1 and T2have been reported, in some studies correlatedwith the severity of negative symptoms (Bessonet al., 1987; Williamson et al., 1992). The pre-frontal cortex is also smaller (Breier et al., 1992),the size correlating with reduction of prefrontalactivation during the Wisconsin card-sorting

test (Weinberger et al., 1992), but also with non-deficit symptoms (Buchanan et al., 1993). How-ever, unlike the findings with the temporallobes, many of the investigations of frontal areasare negative (Suddath et al., 1990).

Other MRI findings include increased thirdventricular enlargement, which is correlatedwith the degree of cognitive deficit (Bornsteinet al., 1992); variable abnormalities of thecorpus callosum in thickness, length or shape(Casanova et al., 1990), which findings can-not yet be resolved; smaller cingulate gyri(Blackwood et al., 1991); and less grey-mattervolume in areas of heteromodal associationcortex such as the dorsolateral prefrontal cortex,the inferior parietal area and the superiortemporal gyrus (Schlaepfer et al., 1994).

Several of the MRI studies have noted greatereffects in male patients (Waddington et al., 1991;O’Callaghan et al., 1992). Most now agree thatthe bulk of ventricular enlargement is earlyin the disease course and reaches a plateau(Borgwardt et al., 2009; Delisi, 2008).

A particularly interesting MRI methodinvolves diffusion tensor imaging (DTI), whichallows one to test white-matter fibre integrityand potential anatomic connectiveness. DTI


studies in schizophrenia have repeatedly foundabnormalities in white-matter fibre tracts.Included are the uncinate fasciculus. As withmost findings in schizophrenia, it is not clearif these changes are primary, or simply reflectproblems elsewhere in the brain and are ‘Wal-lerian degeneration-like’ (Delisi et al., 2006a)

A summary of the morphological MRIfindings is that nearly all reported studieshave found abnormalities of morphology inschizophrenia (Tables 7.7 and 7.8). Smaller brainsize, third and lateral ventricular atrophy andincreased size of the temporal horns are robustfindings. Abnormalities of the left temporallobe seem to be reported more than findings onthe right, especially with decreases in size ofthe superior temporal gyrus and hippocampus,which seem inversely correlated with the pres-ence of positive symptoms. The frontal lobes arealso abnormal. These data substantially supportthe neuropathological data presented above,and emphasize the relevance of abnormalitiesof the limbic system, especially on the left, inthe pathology of schizophrenia.

Unfortunately, contemporary meetings of theAmerican Psychiatric Association or the USSociety of Biological Psychiatry, or other forumsrelated to schizophrenia research, are filled withposter after poster of imaging studies, but oftenwithout an ability to extrapolate the findingsinto a comprehensive understanding of the ill-ness. The vast and ever-growing literature how-ever can be distilled down into a few keyfindings that appear to be replicated and tendto transcend the different methods (CT, MRI):

1. Enlargement of the lateral ventricles, partic-ularly the left, and the temporal horn.

2. Diffuse grey-matter reductions and thinningbilaterally.

3. Reduced white-matter integrity (particularlynow as seen with DTI).

4. Reduced size of the entire temporal cortex,superior temporal gyrus and hippocampus.

There are also multiple reports of increasedbrain asymmetries, abnormalities of the cavumseptum pelllucidum and corpus callosum, andenlargement of the caudates with medications,particularly the antipsychotic medications.

Studies of individuals at first onset of psy-chosis and then following them through thefirst phases of the illness have shown that withtime there is a decrease in size of the left parahip-pocampal gyrus, bilateral cingulate gyri and leftfusiform and orbitofrontal gyri.

Recent longitudinal studies have shown thatmany of the abnormalities are progressive, withleft ventricular expansion, bilateral hemisphericvolume decreases, changes in the cerebellumand corpus-callosum thinning. Relatives fromhigh-risk families who may be exhibiting pro-dromal symptoms also have measurable brainchanges as a group. The high-risk individu-als had reduced apparent diffusion coefficient(ACD) DTI values in the left parahippocampalgyrus, lingual gyrus and superior and mid-dle frontal gyri compared to controls (Delisiet al., 2006a).

The interested reader is referred to severalrecent reviews of this massively expanding andcomplex area (Abi-Dargham and Guillin, 2007;Delisi et al., 2004, 2006a, 2006b). A recentsummit of the experts in this area concluded‘(1) that progressive brain change associatedwith schizophrenia is a real phenomenon, (2)that it is widespread in the brain, but that (3) itvaries among individuals in how widespread orsevere it is and it varies in the timing. (4) It wasgenerally agreed that there is much evidence tosuggest that in the majority of patients it occursearly in the illness and levels off. (5) Althoughthe underlying cause is not known, it could bepartly genetic and cannot wholly be accountedfor by pharmacological treatment. (6) It clearlyhas clinical implications; and (7) its unknowncause is a likely future target for pharmaceuticalintervention’ (Borgwardt et al., 2009).

CBF and PET Studies

Ingvar and Franzen (1974) initially reported thatschizophrenic patients had normal mean hemi-sphere blood flows, but demonstrated a shiftin distribution such that higher levels were lesscommon in frontal structures and more com-mon in postcentral structures when comparedwith controls. This ‘hypofrontality’ has becomea well-replicated finding using several different


imaging techniques. However, like most areaswith functional brain imaging, there is a prob-lem with causality. Does the hypofrontalityoccur because the patients with schizophre-nia are not activating the prefrontal cortex aswell – that is, they are not doing the task aswell – as controls? Or is it primary? We still donot understand this basic problem, which con-tinues to plaque the early PET studies and themore recent fMRI BOLD-activation studies.

Several early studies used xenon inhalationto measure CBF and used brain-activation par-adigms. For example, Gur et al. (1983a, 1983b),in medicated patients, assessed cerebral activityduring verbal and spatial tasks. No hypo-frontality was seen, but schizophrenic patients,unlike controls, showed no changes in hemi-sphere activation during the verbal task, anda greater increase than expected in left hemi-sphere activity when carrying out the spatialtask. Further investigations on unmedicatedpatients (Gur et al., 1985) showed higherresting left hemisphere flow. Further, moreseverely affected patients showed decreasesof anterior left hemisphere activity duringspatial tasks, a pattern rarely found in normals.Their data, they suggested, supported ahypothesis of left-hemisphere overactivationin schizophrenia, and the differences betweenmedicated and unmedicated patients suggestedthat medication helps restore symmetricalblood flow, the main effect being on theleft hemisphere.

Weinberger et al. (1986) have specificallytested frontal-lobe function by administeringpatients the Wisconsin card-sorting test duringxenon-133 inhalation. Controlled cognitivetasks, such as a number-matching test, werealso used, patients being drug-free for aminimum of four weeks. During the number-matching task no differences were notedbetween patients and controls, but with thefrontal-lobe task the clear increases seen incontrols were not seen in frontal regions inthe schizophrenic patients. The changes wereregionally specific, in particular involving thedorsolateral aspect of the prefrontal associationcortex. These data, they suggested, indicatethat patients with schizophrenia have a specific

physiological dysfunction of prefrontal cortex.In later studies (Berman et al., 1986), similarresults have been obtained in patients offmedication and in twin studies, only theaffected MZ twin shows hypofrontality onactivation (Berman et al., 1992). Further, sinceschizophrenic patients and controls did notdiffer on a continuous-performance task withregards to CBF, the abnormalities recordedwere thought to be independent of medicationstatus and of state factors such as attention andmental effort.

There have been many SPECT- and PET-activation studies in schizophrenia. The earlierPET investigations were largely with glucose,but more recently oxygen and ligand studieshave yielded the more interesting results. Acaveat to most of these studies is that theyinvariably use patients that have been on neu-roleptic medications, even if they were free frommedication at the time of testing. Generally,hypofrontality has been reported (Buchsbaumet al., 1990; Siegel et al., 1993), although thereis considerable overlap with control values,and there are several negative studies, espe-cially in acute-onset medication-free patients(Sheppard et al., 1983). Hypofrontality does notseem to correlate with ventricular dilatation andis increased by neuroleptic medication (Delisiand Buchsbaum, 1986). The hypofrontality is inpart a reflection of hyperoccipitality, and is moreprominent in medial rather than lateral frontalcortex (Siegel et al., 1993).

Increased anterior temporal-lobe activity hasbeen shown with deoxyglucose PET (Delisiet al., 1989) and oxygen (Liddle et al., 1992).

Hallucinations have been associated withincreased blood flow in Broca’s area withSPECT (McGuire et al., 1993) and PET (Cleghornet al., 1992), and with lower metabolism in theregion of Wernicke’s area and increased activityin the anterior cingulate regions and stria-tum with fluorodeoxyglucose PET (Cleghornet al., 1992).

Based on factor analysis of patients’ symp-toms in schizophrenia, Liddle (1987) proposedthree clusters, respectively referred to as psy-chomotor poverty, reality distortion and dis-organization syndromes. The former involves


poverty of speech, affect and movement; realitydistortion comprises hallucinations and delu-sions; and the disorganization syndrome refersto formal thought disorder and inappropri-ate affect. While each syndrome can coexist inthe same individual, they were postulated torepresent independent functional systems, cor-responding to limbic and subcortical distributedbrain systems.

Using 15O2-PET and SPM, Liddle et al. (1992)identified separate neuroanatomical profilesfor the three clusters shown in Table 7.9(Liddle et al., 1992). These data, especially forthe reality-distortion factor, have receivedpartial replication (Kaplan et al., 1993). Fristonet al. (1992) have identified the left parahip-pocampal region as being the centralneuroanatomical substrate of schizophrenia,abnormalities here being associated withdysfunction in distributed limbic-associationareas, the pattern of distribution dependenton the presenting behavioural syndrome. Theprimacy of the hippocampus and cingulateregion has also been discussed by Tammingaet al. (1992).

Friston et al. (1992) point out a resolu-tion to the conflicting data with regardsto hypofrontality. Thus, a patient can haveregional hypofrontality (associated with psy-chomotor poverty in the psychomotor-povertysyndrome) and hyperfrontality (in the anteriorcingulate in the disorganisation syndrome).

There is little conformity of data with regardsto basal ganglia findings in psychotic patients,

high, normal and low values being noted(Volkow et al., 1985; Gur, 1986; Buchsbaumet al., 1992). Increased values in basal gangliaregions following neuroleptic treatment arereported from several PET studies (DeLisiand Buchsbaum, 1986; Gur, 1986; Cleghornet al., 1992). These data are in keeping with theknown high dopamine content of these regionsof the brain and their possible abnormalities inpatients with psychosis.

Following an analysis of 50 male schizo-phrenics and controls using SPM, Andreasen(1995) has highlighted the thalamus and whitematter connecting to the frontal lobes as possi-ble key areas of abnormality. More recently therole of the thalamus has been emphasized byBuchsbaum and colleagues (Hazlett et al., 2008;Byne et al., 2009).

Examination of the dopamine hypothesishas been further enhanced by the developmentof selective radioactive dopamine ligands(e.g. fallypride and 11c NPA for the D2/D3receptor), although none are selective for justthe D2 receptor. In the past much attentionwas given to striatal activity, but with theseligands linked to higher-resolution scanners(as low as 2 mm), extrastiatal dopamine hasbeen examined. The thalamus has becomea centrepiece, especially the dorso-medialnucleus. There are several studies suggesting adecreased size and/or activity of this nucleus(Byne et al., 2006) and decreased dopaminebinding (Buchsbaum et al.,2006). Interestingly,the majority of studies, but not all, report

Table 7.9 PET findings in the syndromes of schizophrenia (Liddle et al., 1992)

Syndrome cluster Clinical features PET

Psychomotor poverty Poverty of speech, affect, movement Decreased: prefrontal L parietalIncreased: caudate

Reality distortion Hallucinations and delusions Increased: L – parahippocampal; ventralstriatum

Decreased: R – post-cingulate;post-temporal

Disorganization Thought disorder; inappropriate affect Increased: ant cingulate, mediodorsalthalamus

Decreased: R frontotemporal, Broca’sarea; R, L angular gyrus

Increased/decreased refers to cerebral blood flow compared with controls.


that the left side is the more affected (Kessleret al., 2009). The thalamus is looped with thebasal ganglia and with frontal cortex, themedio-dorsal nucleus being a relay betweensensory and cognitive systems representing anadditional way in which dopamine may beinvolved in the pathogenesis of schizophrenia.There is in normals usually a tight correlationbetween frontal and medio-dorsal thalamicactivity, which is lost in schizophrenia (Hazlettet al., 2008).

The Viral Hypothesis

As noted in earlier chapters, the possibility thatviral infections of the limbic system are asso-ciated with some cases of psychosis has hada long history. There are several reasons tosearch for viruses, including the known excessof winter births in schizophrenia (Fuller Torreyet al., 1977), the schizophrenia-like presentationsof some patients with known encephalitis, thediscovery of slow viruses which affect the ner-vous system, but whose effects are delayed formany years (Fuller Torrey and Peterson, 1973),and the reporting by some authors of evidencefor viral activity in schizophrenic patients.

A second line of evidence has been to exam-ine serum and CSF for abnormal antibody titres,or to seek impaired immunological function inschizophrenic patients. Scattered early reportsof decreased delayed hypersensitivity, priorto the introduction of neuroleptic drugs, areof interest, implying possible immunologicaldysfunction which cannot be attributed to theseagents (Delisi, 1984). Estimations of lympho-cyte characteristics, including the percent-ages of B- and T-cells, have produced variabledata, as has the quantitative assessment ofvarious immunoglobulin classes. OligoclonalIgG bands appear normal (Roos et al., 1985).Reports of raised cytomegalovirus antibodylevels in CSF in up to 70% of patients (Albrechtet al., 1980) have not consistently been replicated(Shrikhande et al., 1985; Alexander et al., 1992)and post-mortem brain studies using stainingfor these viruses have also been negative(Stevens et al., 1984).

There is thus minimal evidence that viralillness is aetiologically related to schizophrenia,and the variable findings, in particular theabnormal features of the immunological system,may reflect immunosuppression related tomedications.

The search for a viral aetiology has turnedinto a quest to implicate prenatal infection, andthe influenza virus has attracted most attention.As noted above, there is good evidence that thereis an excess of winter births in schizophrenia.From over 50 studies, the number that confirmthis effect outways negative results. The effectmay be more robust in females and those with-out a family history, and exposure to the A2influenza virus has been most strongly impli-cated (Mednick et al., 1988; O’Callaghan, 1991;Adams et al., 1993; Takei et al., 1994). Infantsin their second trimester of foetal developmentwere said to be vulnerable to later developingschizophrenia, timing which fits well with someof the neuronal development abnormalities dis-cussed above (Jacob and Beckman, 1989) and theobserved excess of minor physical abnormalitiesnoted in schizophrenics.

This view has its detractors, however. Theyclaim that the largest cohorts do not revealthe effect, that there are statistical objectionsto the handling of the data and that prospec-tive studies are negative (Crow, 1994). In theNational Child Development Study in the UK,in which all children born in a specified weekin 1958 were followed at regular intervals, noexcess of schizophrenia was noted in children ofmothers who suffered influenza in the 1957 epi-demic (Crow, 1994). A further problem seemsto be that the influenza virus fails to cross theplacental barrier.

Another way to examine this theory, orat least the theory that chronic infection hassomething to do with the slow progression ofschizophrenia, has been to examine the inflam-matory cytokine profiles. The implication is thatthere is a mild but chronic immune process thatalters certain brain structures and may alterneurochemical expression. Thus the proinflam-matory cytokines (INF, TNF) shift the balanceof the metabolism of tryptophan from 5-HT


towards kynurenine, a compound which hasbeen shown elevated in CSF in schizophreniaand is a NMDA-receptor antagonist.

Associations with Neurological Disease

The most comprehensive survey of the rela-tionship of schizophrenia-like psychoses toneurological conditions is that of Davison andBagley (1969). It has already been pointed out(in Chapter 3) that a variety of pathologiesthat affect the limbic system may lead to a

disorder with a phenomenological appearanceof a schizophreniform psychosis, the mostconvincing data so far being collected forepilepsy, in particular temporal-lobe epilepsy(see Chapter 10).

Table 7.10 lists the neurological conditionsthat have been associated with a schizophreni-form psychosis, giving the expected number andthe actual number in the early published litera-ture. Taken from Davison and Bagley’s review,it indicates a higher than expected frequency, inparticular for Huntington’s chorea, narcolepsy,

Table 7.10 Comparison of the relative proportions of CNS disorders with psychosis in the1958–1967 literature and their prevalence in the general population (from Davison and Bagley,1969, p. 149)

CNS disorder Prevalence in Distribution of psychotic casesgeneral population in 1958–1967 literature

(per 100 000)a

Expected number Actual number

Trauma – – 532Epilepsy 548 211 276Huntington’s chorea 4 2 86b

Cerebrovascular disease 450 173 42b

Parkinsonism 114 44 20b

Narcolepsy – – 19b

Choreoathetosis – – 19b

Cerebral glioma 8.3 3 19b

Benign cerebral tumour 30 12 17Pituitary adenoma 5.4 2 17b

Postmeningitis/encephalitis – – 16Multiple sclerosis 80 30 11b

Congenital disorders – – 8General paresis 24 9 5Hepatic encephalopathy – – 5Wilson’s disease – – 4Cerebellar degeneration 7 3 3Other cerebral degeneration – – 3Cerebral lipoidosis – – 3Hypoglycaemia – – 3Motor neurone disease 11 4 2Cerebral reticulosis – – 2Torsion spasm – – 2Leber’s optic atrophy – – 1Phenylketonuria – – 1Schilder’s disease – – 1Friedreich’s ataxia – – 1Myotonia congenital – – 1

aPrevalence figures taken from Brewis et al. (1966).bIndicates probable significant difference.


cerebrovascular disease, cerebral gliomas andpituitary adenomata. A lower than expectedincidence occurs with Parkinson’s disease andmultiple sclerosis.

The conclusion with regards to tumoursin particular emphasizes temporal-lobe anddiencephalic locations. A survey of theworld literature on psychiatric symptomatol-ogy in Huntington’s chorea reveals an excessof schizophrenia-like and paranoid condi-tions, the psychopathology often emergingbefore the onset of the movement disorder(Trimble, 1981a). It is of interest that thiscondition is associated with relative overac-tivity of dopamine function, in contrast toParkinson’s disease, in which the incidence ofpsychosis is said to be rare, unless provokedby dopaminergic medications such as L-dopaor bromocriptine. Wilson’s disease, in whichcondition the main pathological findings aremultilobular cirrhosis of the liver and cerebralchanges maximal in the putamen, caudatenucleus and globus pallidus, is also notassociated with an excess of schizophrenia-likesymptoms. Sydenham’s chorea and idiopathicbasal ganglia calcification (Cummings, 1985)are sporadically reported to be associated witha schizophreniform psychosis.

In contrast to some of the other conditionsmentioned, multiple sclerosis is a white-matterdisease affecting in particular periventricularstructures, and its psychopathology, whichincludes emotional lability, affective disorderand a characteristic euphoria (Trimble, 1981a),is markedly different. On the other hand,metachromatic leucodystrophy is an autosomalrecessive disorder due to a deficiency of aryl-sulphatase A which leads to an accumulationof sulphated sphingolipids. In the juvenile andadult-onset forms metachromatic leucodys-trophy presents with behavioural symptoms,which include personality changes, demen-tia and a schizophrenia-like illness (Hydeet al., 1992).

Davison and Bagley (1969) particularlyemphasized diencephalic, brainstem and tem-poral lobe locations of pathology. Midline cere-bral malformations, especially cavum septum

pellucidum, agenesis of the corpus callosumand aqueduct stenosis, are also overrepresentedin patients presenting with schizophrenia (Scottet al., 1993).

There is additional literature emphasizingneurological impairments and signs which maybe found in patients at risk for schizophreniaor those with the condition. Offspring ofschizophrenic parents, thus representing ahigh-risk group for the later development of thedisorder, have an increased number of preg-nancy and birth complications. The high-risksubjects who have a psychiatric breakdownhave the greatest number of such complications,and show abnormal electrodermal galvanicskin responses with shorter latency, slowerhabituation and greater resistance to extinction(Parnas et al., 1981). Further, infants at riskfor schizophrenia show increased evidence ofneurological dysfunction, with neurological softsigns, poor motor coordination and perceptualdeficits (Marcus et al., 1985). Schizophrenicpatients have lower birth weights than expected(Parnas et al., 1981), and cerebral dysfunction,as shown by neurological signs, EEG abnor-malities or a history of seizures, is overreportedin patients with infantile psychosis (Kolvinet al., 1971).

Data from the National Child DevelopmentStudy showed children who later developedschizophrenia to have early language, readingand motoric disturbances, and these abnormal-ities are seen earlier and more frequently inmales (Crow, 1994). Similar findings are beingreported in more recent studies of high-risk peo-ple which emphasize neuromotor developmentand cognition, especially memory/learning.Interestingly, such studies suggest that familyand environmental variables are largely sec-ondary or indirect associations. Nonspecificaffective symptomatology prior to the onset ofthe disorder seems to be of more importancethan most psychotic phenomena (Owens andJohnstone, 2006).

Patients diagnosed as schizophrenic fre-quently show neurological abnormalities onroutine testing. These affect up to 60% ofpatients and cannot be solely ascribed to


medication or undiagnosed neurological illness(Woods and Short, 1985); they may also bedetected in a high proportion of relatives(Kinney et al., 1986). Potential frontal-lobe signs,such as the grasp reflex, and perseveration areespecially reported (Cox and Ludwig, 1979),while Fuller Torrey (1980) reported abnormalface–hand tests (particularly pronounced inthe right hand, suggesting left-hemisphereinvolvement) and graphesthesia. Minor motorand sensory disturbances, more recognizablechoreoathetotic, dystonic and tic-like presen-tations, gait disturbances, difficulties withcoordination and smooth performance ofmotor activities, reflex changes (increased ordecreased) and the presence of infantile reflexessuch as palmomental reflexes, mirror move-ments and variable Babinski responses, have allbeen reported (Claude and Bourguignon, 1927;Quitkin et al., 1976; Tucker and Silberfarb, 1978;Keshavan et al., 1979). These have an inverserelationship to IQ levels (Quitkin et al., 1976),but no apparent association to prognosis.

Abnormalities of eye blinking (Stevens, 1978),either decreased or increased, or episodic rapidparoxysms of blinking, have been noted inschizophrenia, the paroxysms being associ-ated with psychotic episodes. Disturbed eyetracking, including abnormal smooth-pursuiteye movements and saccadic eye movements,are also reported. These movements are madeas the fovea tracks a moving object. Saccadesare fast but ballistic movements that bringthe fovea and the target together, while thesmooth-pursuit eye movements are responsiblefor fixation of slower-moving targets, andthus are continuous with the fovea sitedon the target. Saccadic movements can bevoluntarily executed, whereas smooth-pursuitmovements require stimulus activation andare maintained by attention to the target. Upto 85% of schizophrenic patients are reportedto show abnormal smooth-pursuit tracking,similar abnormalities being noted in 34% ofrelatives (Holzman et al., 1984). This does notappear to be related to inattention, motivationor drug effects, and higher concordance forthe abnormality is shown for MZ as opposedto DZ twins. Saccadic abnormalities have also

been described in schizophrenia (Mialet andPichot, 1981).

Neuropsychological Disturbancesin Schizophrenia

That patients with schizophrenia have alteredneuropsychological performance on a varietyof tests has been known for many years and thefindings have been reviewed by several authors(Flor-Henry, 1983; Cutting, 1990). The CognitiveNeuroscience Treatment Research to ImproveCognition in Schizophrenia (CNTRICS) grouphas summarized the literature in specificdomains and has issued recommendationsabout standardized tests or measures of atten-tion, social cognition and working memory(Barch et al., 2009a, 2009b; Carter et al., 2009;Nuechterlein et al., 2009). In general, all whohave looked for it have noted intellectualdecline of some sort in schizophrenic patients,reflecting Kraepelin’s (1919) summary that pro-found dementia occurs in 75% of patients withhebephrenia and 59% of those with catatonia.The Halstead–Reitan test battery, designed toidentify patients with brain damage, fails todiscriminate between schizophrenic patientsand those with known neurological damage(Heaton and Crowley, 1981), and decrementsare more likely to be found in those with longerhospitalizations and chronic illness (Goldsteinand Halperin, 1977). The decline does notappear to be related to neuroleptic medication;in fact, cognitive testing is often seen to improvein patients receiving these drugs (Heaton andCrowley, 1981).

Patients with simple schizophrenia showthe most markedly dilapidated cognition,while those with a delusional form of thedisorder may be least impaired (Robertson andTaylor, 1985). Schizophrenic patients demon-strate average reading and spelling abilities inthe setting of lower IQ and memory scores,the former tasks reflecting premorbid ability;similar patterns are noted in the dementias(Dalby and Williams, 1986).

Other studies emphasize altered attention,psychomotor speed and problem-solving(Goldberg et al., 1993), and the concept of


deterioration of working memory has becomeimportant. This involves the memory capacityto hold information for short periods of time,usually seconds. It is required for guidingfuture behaviour from representations of theoutside world, rather than for responding toimmediate stimuli (Goldman-Rakik, 1994).It provides temporal and spatial continuitybetween past experience and present actions.In animal studies it is impaired by frontallesions, which produce decrements on delayedresponse tasks. During such delays, increasedneuronal firing is found in Brodmann’s area46 of primate brains. This is equivalent to partof the middle frontal gyrus in humans, and inprimates is anterior to area 8, lesions of whichproduce smooth-pursuit eye deficits, similar tothose reported in schizophrenia. Since patientswith schizophrenia can also be shown to beimpaired on delayed-response tasks, it has beensuggested that these frontal impairments arecentral to the psychology of schizophrenia.

Liddle and Morris (1991) presented cogni-tive data on their patients, separated into thethree syndromes discussed above. Psychomo-tor poverty was associated with impairment ofabstract thinking and long-term memory; disor-ganization syndrome with impaired concentra-tion and new learning; while reality distortionwas associated only with impairment of figureground perception. Frontal-lobe tasks were per-formed badly by those in the psychomotor-poverty and disorganization groups, but weredone better by patients in the reality-distortiongroup. The disorganization syndrome was asso-ciated with impaired verbal fluency.

Johnstone et al. (1978) have drawn attentionto age disorientation in chronic schizophrenicpatients, affecting approximately 25%, in whichthey usually underestimate their age. Thosewith this problem are more likely to have had ayounger first admission to hospital and a longerduration of stay. This same group has drawnattention to the ‘dementia’ of dementia praecox.They thus note that there appears to be a sub-group of patients with schizophrenia who notonly show intellectual impairment but on CTscan demonstrate evidence of structural braindamage. These patients often display ‘negative

features’ such as poverty of speech, retardationand affective changes, while patients with ‘pos-itive features’ such as hallucinations, delusionsand thought disorder are less likely to show acognitive decline.

Laterality effects have also been examined.Schizophrenic patients perform abnormallyon tests of aphasia. On a battery of neuropsy-chological tasks including subsets from theHalstead–Reitan and the Luria–Nebraskabatteries, schizophrenic patients (the majoritydrug-free) not only displayed marked impair-ments but had maximal abnormalities on testsrelated to frontotemporal functions bilater-ally (Taylor and Abrams, 1984). In a morecomprehensive evaluation, Flor-Henry (1983)compared large numbers of schizophrenicand affective-disorder patients and notedasymmetric frontotemporal dysfunction, theleft side being more impaired than the right inthe schizophrenic sample. When an analysis ofcovariance was carried out, looking at depres-sion, mania, schizophrenia and control patients,a continuum of increasing cerebral disorgani-zation was seen, depression showing minimalchanges, while schizophrenia showed maximal.

Recently much attention has been paid tothe potential for cognitive abnormalities to beendophenotypic variants, and deficits of execu-tive function and working memory have beenthe focus of attention (Gur et al., 2007). The lat-ter is capacity-limited; in other words, thereis an inverted U-shaped curve of efficiency innormals, closely linked to increased activationof the dorsolateral prefrontal cortex. Abnor-malities of working memory have been wellrecorded, schizophrenic patients operating at orbeyond efficiency being unable to maintain per-formance within their expected capacity. Duringworking-memory tasks and imaging they showeither hyperfrontality or hypofrontality in acomplex relationship with task difficulty andperformance (Karlsgodt et al., 2009). Unaffectedsibs of patients also show difficulties with work-ing memory, including unaffected MZ twins(Prata et al., 2008a, 2008b; Filbey et al., 2008;Toulopoulou et al., 2006) and genetic links havebeen probed. The latter have included an asso-ciation with the disrupted-in-schizophrenia 1


(DISC1) gene (Matsuzaki and Tohyama, 2007;Ishizuka et al., 2006; Callicott et al., 2005).

Another strategy has been to use high-risksubjects. Neuropsychological impairments havebeen shown in those who later develop the dis-order, more severe than that seen in those whogo on to develop bipolar disorders, and work-ing memory again is highlighted. Performanceabnormalities on verbal tasks, including verballearning and fluency, are most often reportedin those who later convert to schizophrenia.However, these early cognitive problems do notclearly correlate with later psychotic symptoms,except weakly with negative features. What ismore relevant is the identification of abnormalthinking, reflecting abnormal cognitive struc-ture in these high-risk people. In addition tothese cognitive variables, other strong predic-tors are high scores on the SIPS and SIS, bizarrethinking, schizotypal personality disorder, sus-piciousness, poor social functioning and poorsocial skills. These findings link with otherssuch as deficits of facial recognition and emo-tional discrimination of faces in schizophrenia,correlated with abnormal early evoked poten-tials, suggesting processing problems occur atan early stage of sensory processing.

Of particular interest is the potential link andoverlap with findings from patients with autism,who show many similarities to and differencesfrom schizophrenia. The largest similarities arein the problems with social cues, face processing,empathy and understanding the actions of oth-ers, called ‘theory of mind’ (Anckarsater, 2006;Rapoport et al., 2009; Vollm et al., 2006).

SOME OUTSTANDING ISSUES

The evidence reviewed clearly indicates a sig-nificant genetic contribution towards the devel-opment of schizophrenia. However, it is clearthat this contribution to schizophrenia doesnot explain all the variance, and environmentalvariables have been persistently sought. Indi-vidually, these environmental factors are ofunknown effect size and several have been sug-gested, although exactly how they interact withthe genetic liability is unclear.

Further, the consensus that what is inheritedis a schizophrenia spectrum, as opposed toschizophrenia per se, has gained ground, andcertain biological markers, for example plateletMAO activity or dysfunction of eye move-ments, may reflect upon underlying geneti-cally determined biochemical or neurocircuitrypropensities to develop psychosis under certainconditions. The search for the gene(s) responsi-ble for the schizophrenias continues. Some obvi-ous candidates, especially relating to dopamineregulation, seem to have been ruled out. Theconsensus has been to look for co-acting loci, andthere have been many reports of linkage find-ings or results from genome-wide scans, withquite heterogenous results; some 30–40 suscep-tibility genes have been identified at thresholdsignificance. In part this may reflect the factthat genetic markers are themselves only prox-ies for the risk variant. But there are otherconfounding factors, including the problem ofthe quality of genotype analysis, the statisticalrequirements of such potentially large data setsand the delineation of the phenotype. What arethe features of a subset of cases that show anenhanced genetic signal? The strategy of lookingat endophenotypes, and away from DSM IV-TR-defined categories, seems a promising wayforward, but to date no endophenotypes relatedto schizophrenia have yielded any secrets togenetics. However, recent attempts to combinegenetics with brain imaging are proving inter-esting, examining the genetic contributions tothe cerebral structural, functional changes andendophenotypic variance.

There was hope that the newer generation ofgenome-wide studies might prove more power-ful than linkage and candidate gene strategies.That promise has not been fulfilled as yet. Thelatest genome-wide studies if anything tendto suggest similarities, rather than differences,between bipolar affective disorders and theschizophrenias (Esslinger et al., 2009; Keshavanet al., 2007; Tan et al., 2008; van Haren et al., 2008;Lin and Mitchell, 2008; Thaker, 2008; van Oset al., 2008) (see Chapter 8).

To date, and so many years after Slater’s stud-ies, the picture remains that there may be severalor many genes of small effect contributing to the


schizophrenia spectrum. Essentially we may belearning that, with only a few exceptions suchas Huntington’s disease, there are no genes formental illness per se, but rather genetic effectsthat relate to information processing in thebrain and subsequent cognitive and behaviouralstrategies, the endophenotypes, which providea substrate for the later development of the phe-notype. No longer is a statistical relationship be-tween genes and a DSM IV-TR diagnosis beinglooked for, but rather the relationship betweenthe neurobiological effects of genes and the pro-cesses that lead to any particular phenotype.

With regards to environmental agents, thereis evidence that some schizophrenic patientsdemonstrate increased early brain damage,either through perinatal or uterine insults.Murray et al. (1985) advocate it would be use-ful to divide schizophrenia into familial and spo-radic cases. The sporadic cases are thusassociated with increased ventricular size(highly replicated) and environmental traumasreflective of perinatal damage or early headinjury. The genetic predisposition, if strong, willpromote the disorder, while, in the presence ofa lesser genetic vulnerability, the environmentalinsult becomes necessary for the full expressionof the illness. Where there is no geneticvulnerability, as in, for example, some cases ofepilepsy, the cerebral pathology itself directlyprovokes the schizophrenia-like symptoms.

More recent discussion of environmentaleffects has concerned epigenetic and pharma-cological influences. The former should includeuterine and neonatal events on cerebral devel-opment, synaptogenesis and synaptic pruning,as well as early psychosocial stressors (Parentand Meaney, 2008). Alternative environmentalinsults have been sought, in particular neu-rological conditions, head injury or infection,for example with viruses. However, recentresearch has centred on the use of psychoactivedrugs, especially cannabis. The emphasisfrom CNS disorders is on those which affecttemporal-lobe, basal-ganglia and diencephalicfunction. Temporal-lobe disorders, in particularsuch conditions as epilepsy affecting limbicstructures, seem to be more associated withhallucinations, while basal-ganglia disorders

lead more to disturbances of motoric activity,motivation and cognition via the fronto-striatalre-entrant networks.

The pathological and MRI studies reviewedreinforce a neuroanatomical approach, empha-sizing medial temporal structures in associationwith hallucinations, paranoia and thoughtdisorder, and the basal-ganglia frontal axiswith negativism and motor signs. Such findingsemphasize that the ‘anatomy of schizophrenia’(Stevens, 1973) is predominantly related toallocortical and subcortical structures, particu-larly those of the limbic lobe (entorhinal cortex:hippocampus), the thalamus and the relatedbasal ganglia.

In cerebral development, neuronal migrationis from the ventricular zone outwards to thecortical plate. Younger neurones thus migratethrough more mature ones, and the migration isdirected by previously formed radial glial fibres.In the human entorhinal cortex, these migra-tions are complete by six months of gestation;the hippocampus and temporal neocortex aresomewhat later, the hippocampus not reachingadult volume until about the age of two years,and continuing to myelinate well into adult life(Benes et al., 1994).

The upper layers of the entorhinal cortex,especially layer 2, containing pre-alpha cells,are sites of origin of the glutamate projectionsto the hippocampus, and receive a dopaminer-gic input. It is important to note that duringnormal development, the ventricular volumeis reduced, and so abnormal development willlead to increased volumes. Further, the develop-ment of the left hemisphere is later than the right,and any damage that influences early develop-ment of these structures is therefore more likelyto impact on the left.

The parts of the entorhinal cortex involvedin pathological studies of schizophrenia areintimately connected with the insula and theorbitofrontal cortex, and it is of interest thatthese entorhinal areas are not found in otheranimals, including other primates.

Some authors have drawn attention to thelikeness of some schizophrenic symptoms tofrontal-lobe disorders. In particular this in-volves the dorsolateral prefrontal cortex, and the


symptoms include those seen following frontalinjury, such as affective changes, impairedmotivation, poor insight and other ‘defect’symptoms frequently seen in schizophrenia.Indeed, the evidence for frontal-lobe dysfunc-tion in schizophrenic patients is clear, and hasbeen noted not only in the neuropathologicalstudies but also in EEG, evoked-potential data,neuropsychological studies (especially those ofworking memory), testing for soft neurologicalsigns, and more recently imaging techniquessuch as PET and MRI.

The pathological studies of frontal corticalareas are more limited, but do suggest somechanges, which either represent evidence ofmore generalized pathology in schizophrenicbrains or are a localized contribution to thedevelopment of the psychopathology.

The orbitofrontal and temporal pole areasare connected by the uncinate fasciculus, andlesions at such sites may impair higher-ordercognitive processes and lead to disorderedsocial communication. Recent data clearlysuggest that abnormal hippocampal structureis associated with functional impairment offrontal-lobe activity, supporting a suggestionthat the temporal/limbic changes may beprimary. Alternatively, disturbances of frontalprojections to the parietal and temporal associ-ation areas may impair sensory representationsin these areas, explaining positive symptoms,setting the frontal lobes as crucial for thedevelopment of both positive and negativesymptoms. More recent imaging studies haveconcerned changes in white-matter tracts andhave identified alterations in white matter,including the corpus callosum, the uncinatefasciculus and the anterior limb of the internalcapsule. It is quite unclear whether these arepart of the structural alterations that relateto schizophrenia or are secondary to otherneuronal loss. It does seem to be the case thatmany of the structural (and functional) changesoccur early on in the disorder, are not secondaryto antipsychotic drug prescriptions and may beseen in the high-risk group before diagnosis.

The evidence for two distinct syndromesin schizophrenia has been pursued by Crowand colleagues (Crow, 1980). The suggestion

is that there is both a neurochemical and aneuropathological contribution to schizophre-nia, with two syndromes. Thus, the type Isyndrome (equivalent to acute schizophrenia) ischaracterized by positive symptoms (delusions,hallucinations and thought disorder) and isrelated to change in dopaminergic transmission.The type II syndrome (equivalent to the defectstate) is characterized by negative symptoms(affective flattening and poverty of speech) andis associated with intellectual impairment andstructural changes in the brain. Current thoughtalso involves a continuum model and includesconsideration of the schizophrenia spectrum. Itis established that abnormalities of cognitionand thinking are present well before the onsetof the actual psychosis. The high-risk studiesall come to similar conclusions, namely thata prodromal state can be identified linked tounusual thinking and cognitive problems, espe-cially with working memory and verbal tasks,which, if identified early, may be more amenableto treatment strategies than the fully-developedsyndrome. However, the studies beg the ques-tion as to whether the identified pattern shouldbe seen as part of the schizophrenia, especiallyas it also includes those who achieve a diagno-sis of schizotypal personality disorder. In otherwords, should the syndrome schizophrenia(s)not include such preludes? The role of envi-ronmental factors in the actual descent into psy-chosis remains unclear, although neurochemicalinvasions, especially from psychoactive drugssuch as cannabis, are clearly implicated.

There is certainly considerable evidence thatstructural changes are likely to be found insome schizophrenic brains and that cerebralatrophy on the whole is related to more negativesymptoms. Positive symptoms, in particular thefirst-rank symptoms of Schneider, have beenrelated to abnormal dopaminergic transmission.The dopamine hypothesis has in part stoodthe test of time. Dopamine is closely linkedto disorders of motivation (see Chapter 9), aprominent part of the schizophrenia clinicalpicture, and the known interactions betweenbrain areas such as the frontal cortex, theaccumbens and the midbrain dopamine areas(VTA) as part of the underlying anatomy of


schizophrenia are supported by pathologicaland MRI findings. Patients with positivesymptoms show increased growth-hormoneresponses to apomorphine, decreased baselineprolactin, low CSF–HVA accumulation andminimal heritability (first-rank symptoms), andtend to have less cerebral atrophy. Further, thepathological and PET data reviewed suggestthat temporal-lobe, including parahippocampalgyrus, abnormalities may be more related tosuch symptoms. These data certainly supporta particular relationship between positivesymptoms, relatively normal brain structure,abnormalities of dopamine and possiblytemporal-lobe dysfunction.

In contrast, negative symptoms appear morerelated to periventricular abnormalities, cere-bral atrophy and environmental insults, and areperhaps related to dopamine dysfunction, but ina different way (as the disorder progresses anddisorder of motivation becomes more appar-ent). Experimental support for defining twosubtypes has derived from intercorrelations ofpatients’ symptoms, which indicate generallythat positive symptoms correlate positively, asdo negative symptoms, although the internalconsistency is much greater for negative symp-toms (Andreasen and Olsen, 1982).

The important contribution of abnormalitiesof dopamine to schizophrenia has been per-sistently argued. But it is well known that theantipsychotic effects of neuroleptic drugs are notconfined to schizophrenia, and they amelioratepsychotic symptoms in mania and acute organicbrain syndromes. Further, because such medica-tions alleviate some symptoms of schizophreniadoes not mean that dopamine itself has theprimary role in pathogenesis. While it has beennoted that the dopamine-blocking effect ofthese drugs is the most important for helpingthe psychosis, it should be remembered thatanticholinergic drugs alleviate the symptoms ofParkinson’s disease, even though the primarypathology in that condition relates to dopamineand not acetylcholine. Other neurochemicalhypotheses have obviously been entertained,and as noted involve serotonergic, other cate-cholaminergic, peptidergic and glutaminergicabnormalities. Mackay (1980) has argued

that a chronic dopamine deficit is relatedto schizophrenia, rather than overactivity,highlighting the reported low CSF–HVA levels.This reduced dopamine turnover leads not onlyto reduced presynaptic dopamine accumulationbut also to a compensatory postsynapticdopamine-receptor proliferation. In his theory,both positive and negative symptoms thusrelate to a defect of dopaminergic function,the positive symptoms emerging as a resultof a switch from a deficit to an increase inpresynaptic release of dopamine, the latterinteracting with the excessive number ofpostsynaptic dopamine receptors and leadingto the florid positive symptoms.

An alternative is that abnormally low pre-frontal dopamine activity is related to deficitsymptoms and excess dopamine activity in sub-cortical structures relates to positive symptoms.The mainly negative PET dopamine ligand-binding studies to date were disappointing forthose in favour of the dopamine hypothesis,although attention is now being directed toalternative dopamine receptors, especially D2,D3 and D4. The importance of frontal dopamineactivity is highlighted by the data that relateit to working memory, and the links to COMTalleles, genes associated with an enzyme closelylinked with dopamine metabolism and shownto be associated with cognitive performance(Lawrie et al., 2008a Roffman et al., 2006; vanHaren et al., 2008).

Decreased D2 and D3 receptor occupancy iswell replicated in several brain areas. Thesefindings have been enhanced by the develop-ment of selective radioactive dopamine ligands(e.g. fallypride and 11c NPA for the D2/D3receptor, although none are selective for just theD2 receptor). In the past much attention wasgiven to striatal activity, but with these ligands,extrastiatal dopamine can be estimated. The tha-lamus, especially the dorso-medial nucleus isimportant (Andreasen et al., 1994). Several stud-ies suggesting a decreased size and/or activityof this nucleus (Siegel et al., 1993; Shihabud-din et al., 2001), along with decreased dopaminebinding are reported. Lower prefrontal corti-cal dopamine has been shown to correlate withincreased striatal release. This may be mediated


by the reduced excitatory feedback from frontalcortex (glutamate) to the inhibitory GABA neu-rones in the VTA, thus increasing firing andhence output of dopamine from the VTA tostriatal structures such as the accumbens.

While dopamine remains a key transmit-ter in the schizophrenia story, it is obviouslyinvolved in a variety of psychopathologies, anda question remains as to its biological func-tion in regulating behaviour aside from its rolein motor programming. The findings relatedto disorders of motivation (Chapter 9) suggestit is related to novel reward detection in theenvironment and is interlinked with hedonia.It somehow attributes motivational salience toneural events, influencing attention and the sub-sequent behavioural repertoires. This in itselfdoes not lead to psychotic symptoms, but theabnormal attribution of salience to environmen-tal and somatic representations may lead overtime to cognitive restructuring (Kapur, 2006).This may vary with pre-existing or consequentstructural neuronal and circuitry changes, andsuch hypotheses need to be integrated with theaccumulating knowledge of brain morphologi-cal changes.

The current status of the dopamine hypothe-sis suggests that D2/D3 receptors are still centralto the theories, decreased receptor occupancybeing well replicated in several brain areas.Attention has moved to extrastriate regions andtheir involvement not only with the anatomyof the cortical–striatal–thalamic re-entrant cir-cuits, but with reciprocal relationships betweenGABA and dopamine, and the link throughglutamate.

It remains unclear what it is that drivesthe increase in dopamine activity. Early workhad suggested a link through GABA, espe-cially with a subgroup of GABA neurones, theparvalbumin-containing cells. Disturbed neu-ronal rhythmic expression in key structures is adeveloping story.

The known importance of the balancebetween GABA and glutamate within theCNS for coordinating neuronal firing andoscillations, and the importance of glutamatein the regulation of the cortical–subcorticalre-entrant circuits, has led to looking at GABA

as well as glutamate as possible substrates.Thus, glutamate input to dopamine-releasingcells (from, for example, the hippocampus andthe pedunculopontine areas) normally inducesbursts of spikes within the VTA, which aredisrupted by a deficit of glutamate. Thesetheories put more emphasis on the possiblerelevance of cortical changes as primary in thedisorder.

Crow (1993) has persistently argued againstthe Kraepelinian dichotomy, pointing out thatoverlaps of schizophrenia and affective disor-ders occur phenomenologically and genetically(the proportion of relatives of schizophrenicsthat have an affective disorder is increased).In his continuum (see Figure 7.1), whichextends from unipolar depression at one endto schizophrenia with a defect state at theother, variations of sociability and emotionalityare expressed.

Murray et al. (1992) on the other hand arguethat there is a congenital schizophrenia, witha primary genetic defect and/or a primaryenvironmental insult such as a viral infection,and abnormal neuronal maturation in crucialCNS sites. These patients show personality andsocial impairments in childhood, and developthe more severe cognitive impairments andnegative symptoms. In contrast, patients withadult-onset schizophrenia, which includes sev-eral syndromes, do not have the classic dementia

UP BP

elation

mood-incongruentdelusions

nuclear symptomsmood change

flat affect(+) (−)

SA S+A S+DS−A

Figure 7.1 The continuum concept of psychotic ill-ness. UP, unipolar affective illness; BP, bipolar; SA,schizoaffective psychosis; S, schizophrenia with andwithout significant mood disturbance (+ and – A),and with defect state (S+D) (reproduced with per-mission from Crow, 1993; Biol Psych, 2 Ed, p. 223)


Table 7.11 Evidence for laterality in schizophrenia

Neurological examination Increased abnormalities in right-side sensory testing; increasedleft-handedness

Neuropsychology Association with poor performance on language tasksEEG Dominant hemisphere abnormalities on routine testing, power spectra,

with telemetry, evoked potentials and BEAM studiesCT studies Density changes; new onset casesMRI studies Especially positive symptoms: left superior temporal gyrus changesPET studies Variable: recent thalamic findingsNeurochemistry Increased dopamine, left amygdala; decreased GABA and glutamate, left

hippocampusNeuropathology Parahippocampal gyrus abnormalities in some studies onlyNeurology Epilepsy and head-injury studies

praecox and are more likely to have relapsing,remitting psychosis with positive and affec-tive symptoms.

The search for virus infections has not beenrewarding and the evidence that intrauterineinfection with the influenza virus is important isat present quite unclear. However, one impor-tance of the infective theory relates to somehistorical considerations of schizophrenia: it hasbeen suggested that schizophrenia was uncom-mon in the medical literature prior to the 19thcentury, implying, rather like AIDS encephali-tis, that it may have been a new disease, andthat an environmental agent, such as a virus,must have become prevalent around that time.The alternative is that it is an old conditionwhich has been ubiquitous and always present.The problem then is to explain its persistencein view of the low fecundity of schizophrenicpatients (Hare, 1979). This paradox remainsunresolved.

Finally, substantial evidence has accumu-lated suggesting that at least in some formsof schizophrenia, the major disturbance is in thedominant hemisphere (see Table 7.11). Becauseone of the cardinal features of schizophrenia isa disturbance of thought and language and thedominant hemisphere is known to be involvedin modulating such activities, it is hardly sur-prising that the weight of evidence suggests thatthe dominant hemisphere is primarily involved,and this continues to be the case. It is clear thatto implicate the left hemisphere alone would

be nonsensical. In fact, Cutting (1990) made astrong case for the nondominant hemispherehaving a primary role in the symptomatology,a view recently supported by Crow and hiscolleagues (Crow et al., 2007, 2009; Crow 2009).Nonetheless, the role of the dominant hemi-sphere is of outstanding importance, perhapsearly on in the disorder, or in providing thekey harmonic that reverberates through thedeveloping phenomenology. The findings frompathology, and the progressive nature of theimaging data, have initiated a discussion asto whether schizophrenia should be viewed asa progressive disorder, as Kraepelin originallyenvisaged, or even as a degenerative disorder:a ‘dementia’ in our current-day terminology.Although at first sight the degenerative hypoth-esis seems possible (schizophrenia being viewedas a dementia praecox), this is an unlikelyhypothesis. The early age of onset vis-a-visthe dementias, the subtle nature of the patho-logical changes and the eventual defect stateas opposed to a continued progressive declineargue against it.

In summary, the current perspective onschizophrenia is to view it as a neurode-velopmental disorder, with a clear geneticcomponent. The final clinical picture howevermust be viewed as a spectrum, reflectingalterations in the structure of the cognitiveand perceptual abilities of those who becomepatients, these being yoked with aberrantneuroanatomy and neurophysiology.

Documents

Biological Psychiatry (Trimble/Biological Psychiatry) || The Schizophrenias