.Cognitive Brain Research 5 1996 4954

Neurological basis of skill learningAtsushi Yamadori a,), Takashi Yoshida b, Etsuro Mori c, Hikari Yamashita c

a Section of Neuropsychology, Diision of Disability Science, Tohoku Uniersity Graduate School of Medicine, 2-1, Seiryo-machi, Aoba-ku, Sendai 980-77,Japan

b Neurology Serice, Hyogo Brain and Heart Center at Himeji, Himeji, Japanc Department of Clinical Research, Hyogo Institute for Aging Brain and Cognitie Disorders, Himeji, Japan

Abstract

Skill-learning, i.e. anterograde memory of procedures has been separated from conventional declarative memory, or event- ordata-related memory. This type of memory requires a concerted activity of various neural structures which are not assigned to theacquisition of declarative memory. We employed mirror reading task as a paradigm of skill acquisition memory and tried to elucidatepossible neurological mechanisms involved in the procedural memory process. Ten normal control subjects, 10 early, non-treatedParkinsons, and 9 relatively early spinocerebellar degeneration patients participated in our study. The results showed a clear dissociationbetween declarative memory and mirror reading skill acquisition capacity. Thus, the Parkinsons patients as well as the spinocerebellarpatients showed retardation in acquiring mirror reading skills, while both groups showed normal performance in auditory verbal learningtests and word recognition tests. The facts suggest a possible role of the nigro-striatal system and fronto-ponto-cerebellar system informing these skill-related memory.

Keywords: Procedural memory; Declarative memory; Mirror reading; Parkinsons disease; Spinocerebellar degeneration; Nigro-striatal system; Fronto-ponto-cerebellar system

1. Introduction

In recent years it has been made clear that neurologicalsubstrata of skill acquisition is different from those en-

w xgaged in data-related information acquisition 4 . The for-mer is named procedural memory and the latter declarative

w xmemory 20 . Skills can be defined as such behavioralchanges that the execution of a particular type of percep-tual or motor tasks becomes fast and smooth throughrepetitive experience. Some types of cognitive operationslike a puzzle solution was also said to belong to this

w xdomain 5 .Reading of mirror inverted words mobilizes simple and

w xcomplex cognitive operations at the same time 4 . It issimple in a sense that the task is just to read aloud a wordpresented. For non-aphasic literate subjects it is automaticand requires no effort. It is complex in a sense that eachletter of the stimulus word must be inverted in the mindand these inverted letter images must be combined again torecognize them as an ordinary word. As a perceptual taskit is ideal since no manual motor component is involved.

) . .Corresponding author. Fax: 81 22 717-7360.

The mirror reading task was first reported to be selec-w xtively impaired in Huntingtons disease patients 17 . Thus

Marton and others showed that their patients of Hunting-tons disease, although mildly demented, showed a con-trasting performance in mirror reading paradigm. Theywere retarded in their ability to acquire mirror reading skillbut showed normal ability of verbal recognition memoryof the words used in the test. This is a persuasive evidencethat mirror reading skill is acquired through the basalganglia activities.

Stimulated by these works we studied a possible rela-tionship between the mirror reading skill acquisition andthe subcortical neural structures. As candidates whichshows selective subcortical dysfunction we chose Parkin-sons disease and spinocerebellar degeneration.

2. Parkinsons disease

2.1. Subjects

Newly diagnosed, early, and non-treated Parkinsonsw xdisease patients were recruited for our project 21,23 . The

mean years of age of the patients were 64.1 with standard

0926-6410r96r$17.00 Copyright q 1996 Elsevier Science B.V. All rights reserved. .PII S0926-6410 96 00040-7

( )A. Yamadori et al.rCognitie Brain Research 5 1996 495450

deviation of 6.88. The mean years of education were 8.5with standard deviation of 0.92. Four were males and 6were females. The severity of the stage was evaluated bythe criteria proposed by Yahr which divides the severity of

w xthe disease into 5 grades 13 . Two belonged to stage I,seven to stage II, and one to stage III.

Ten normal subjects served as controls. The mean yearsof age of the control subjects were 67.7 with standarddeviation of 4.45. The mean years of education were 10.3years. Two were males and eight were females.

2.2. Methods

We developed a personal computer system for theMirror Reading Test. Words used as stimuli were 420nouns and proper names. All the words consisted of 6 to 8

Hiragana characters Japanese phonograms representing a. .syllable Fig. 1 . From these word pools a pair of words

was recalled, then reversed into mirror image and dis-played in white on a 14-inch black CRT window in two

.rows Fig. 2 . The paired words were chosen, so that therewas no relation in meaning and pronunciation betweenthem.

A subjects task was to read aloud a displayed pair ofwords as quickly as possible. Reading of the pair consti-tuted one trial. When reading was finished, the examinerpressed a key manually, and the duration from the start ofthe display to the end of reading was recorded in thecomputer automatically in the order of 1 ms. When readingwas incorrect, the subject was so informed, and tried tocorrect his reading. The upper limit of reading time al-lowed for one trial was set to 120 s. If the subject did notsucceed reading within this time limit, his reading timewas recorded as 120 s.

Twenty successive trials, i.e. reading of 40 words,constituted one block of test. Five blocks of test were runin one day session. The test was performed for 3 succes-sive days to evaluate the extent of the reading skill acquisi-tion. A week later, the final 5-block test was performed tocheck the retention of the skill.

Fig. 1. A sample of two common Japanese words in Hiragana syllabo-.gram .

Fig. 2. Examples of a pair of mirror reflected words used for the task. Thetwo words in the Fig. 1 were mirror-reflected. As in the figure, the wordsappeared in white against black background.

In order to evaluate the status of declarative memory,10 of the 20 words in a block were repeatedly displayed inevery block reading, and the other 10 were displayed onlyonce. The former will be called repeated word and thelatter unique word.

After the third day session, a recognition memory testwas performed. Thus a list of 60 words, consisting of 15repeated words, 15 unique words and 30 unrelated words,was presented to the subjects and they were asked to checkwhether they encountered the words in the foregoing mir-ror reading test or not.

2.3. Neuropsychological status

The general neuropsychological background was as- .sessed by Mini-Mental State Examination MMSE , Raven

.Colored Progressive Matrices Test RCPM , Wechslers .Adult Intelligence Scale WAIS and Auditory Verbal

.Learning Test AVLT of Rey.

2.4. Results

The results of performance for MMSE, RCPM andWAIS evaluation for the two groups of the subjects aresummarized in Table 1. As is clear from the table, no

Table 1Background neuropsychological evaluation of Parkinsons patients

. .PD ns10 C ns10 . .mean S.D. mean S.D. . .MMSE 28.9 1.13 29.8 0.40 . .RCPM 29.4 3.32 30.7 4.02 . .WAIS full IQ 104.3 10.06 107.8 12.06 . .verbal IQ 107.0 11.36 110.5 13.39 . .perf IQ 101.4 13.51 103.7 11.69

PD, Parkinsons patients; C, control subjects; MMSE, Mini-Mental StateExamination; RCPM, Raven Colored Progressive Matrices; WAIS,Wechsler Adult Intelligence Scale; perf IQ, performance IntelligenceQuotient.

( )A. Yamadori et al.rCognitie Brain Research 5 1996 4954 51

.Fig. 3. Results of Auditory Verbal Learning Test AVLT . The average . .scores of Parkinsons patients PD and control subjects C were com-

pared. Ordinate denotes the number of words recalled and abscissadenotes number of trial sessions. Trial 6 was performed after interference.

significant difference was present between the patient groupand the control group.

Fig. 3 shows the results of AVLT performance. Againno significant difference was detected between the twogroups. Thus our Parkinsonian group showed normal intel-ligence and normal verbal learning ability.

Fig. 4 summarizes the mirror reading performance. The .mean seconds spent for one trial ordinate was plotted

.against the days and the blocks in a day abscissa . As forthe repeated words, reading time shortened as the numberof the sessions increased both for the Parkinsons andcontrol group. As for the unique words, the control andParkinson group showed a contrasting curve. The controlgroup showed steady improvement of reading speed overblocks and days although it was slower compared to theone for repeated words. On the contrary, the Parkinson

Fig. 4. Results of mirror reading task. Average reading times of a pair ofwords were compared for repeated words and unique words betweenParkinsons and control group. Ordinate denotes mean reading time andabscissa denote blocks and days. Reading times of standard words wereshown on the lower left.

Fig. 5. Results of recognition test for repeated words. Ordinate denotesscores as percent ratio of correct recognition.

group failed in improving their reading speed. They evenshowed a tendency to deteriorate toward the end of theeach day session.

However, once acquired, the reading skill was wellretained at least for a week as the curve on the day 4performance clearly demonstrates.

We also measured the speed of reading for non-in-verted, normal words. There was no difference betweenthe two groups.

The result of recognition memory for repeated words isshown in Fig. 5. There was no significant differencebetween the two groups for the repeated words.

2.5. Comment

It can safely be concluded that the failure of improve-ment of reading speed for mirror inverted unique words inthe Parkinsons subjects after repeated training cannot beattributed to general cognitive deterioration since the groupshowed no such change as normal MMSE, WAIS andRCPM score clearly demonstrated.

Also it is difficult to attribute the failure to memoryimpairment in general, since their ability to memorizewords was just as good as the controls as demonstrated inAVLT and recognition memory of repeated words.

Bradykinesia, often observed in Parkinsons disease,cannot account for this retarded skill acquisition either,since the speed of reading of standard words did not differfrom the control group.

We agree with the previous authors that mirror readingskill acquisition represents a category of memory which isorganized differently from traditional data or event related

w xdeclarative memory 4,17 . Unlike declarative memorywhich is related mainly with the thalamus, hippocampus,and neocortical structures, this perceptuo-verbal skill ac-quisition is very likely related with the subcortical struc-tures. Martons work suggests its possible relation with thecaudate nucleus, or more generally with the striatum.Unlike their patients who suffered from Huntingtons dis-

( )A. Yamadori et al.rCognitie Brain Research 5 1996 495452

ease, our patients suffered from Parkinsons disease. OurParkinsons patients received no medication. The stage oftheir disease was very early and no neurological signsother than those of early Parkinsons disease were demon-strated. These facts strongly suggest that their failure ofmirror reading skill acquisition is somehow related withthe disturbed activities of the nigro-striate dysfunction.

3. Spinocerebellar degeneration

3.1. Subjects

Nine patients whom we diagnosed as suffering fromspinocerebellar degeneration on clinical neurological basis

w xparticipated in the mirror reading test 22 . Five of themwere classified into late cortical cerebellar atrophy and therest into olivo-ponto-cerebellar degeneration. Unlike theParkinson group it was difficult to select the patients bydisease stage or medication. So all the patients who agreedto participate in the project were included. The mean yearsof age of the patient group were 59.4 with standarddeviation of 7.7. The mean years of education was 10.0with standard deviation of 2.1. The average years after theappearance of the first symptoms were 3.1 years withstandard deviation of 1.6 years. The average grades of thedisease in terms of the criteria set by Hirayama which

w xdivides the disease into 5 grades 11 were 1.9 withstandard deviation of 0.8.

The data of the same 10 normal subjects described inthe Parkinson section were used as control.

3.2. Methods

The same Mirror Reading Test that was used for theParkinsonian patients was administered.

General neuropsychological evaluation: MMSE, RCPM,AVLT and WAIS were evaluated.

3.3. Results

3.3.1. Neuropsychological backgroundThe results are summarized in Table 2. Compared to the

control group, WAIS IQ of the SCD patients was signifi-

Table 2Background neuropsychological evaluation of spinocerebellar patients

. .SCD ns9 C ns10 . .mean S.D. mean S.D. . .MMSE 29.4 0.7 29.8 0.40 . .RCPM 28.8 3.4 30.7 4.02) . .WAIS full IQ 94.1 8.7 107.8 12.06) . .verbal IQ 96.9 6.3 110.5 13.39) . .perf IQ 90.1 11.4 103.7 11.69

SCD, Spinocerebellar patients. Other abbreviations as in Table 1. ) Sig-nificant difference compared to the control subjects at the level ofP -0.05.

.Fig. 6. Results of Auditory Verbal Learning Test AVLT . The average .scores of spinocerebellar degeneration patients SCD and control sub-

jects were compared. Ordinate denotes the number of words recalled andabscissa denotes number of trial sessions. Delayed means trial afterinterference.

cantly lower at the level of 0.05%. The scores of MMSEand RCPM showed no significant difference. The learningpattern of AVLT also showed no significant difference

.between the two groups Fig. 6 .

3.3.2. Mirror readingThe results of mirror reading are summarized in Fig. 7.

Reading of repeated words showed gradual shortening ofreading time like the control group. As for reading of theunique words, the control group showed slow but definiteimprovement. On the other hand, the SCD group failed toshow this gradual shortening of reading time. Like theParkinson group, they showed deterioration within the dayafter repeated sessions. There was no tendency of improve-ment between the day 2 and day 3.

The reading skill was well retained at least for a weekas the curve on the day 4 performance clearly demon-strates.

The reading time of non-inverted, standard wordsshowed no difference between the SCD and control group.

Fig. 7. Results of mirror reading task. Average reading times of a pair ofwords were compared for repeated words and unique words betweenspinocerebellar patients and control group. Ordinate denotes mean read-ing time and abscissa denote blocks and days. Reading times of standardwords were shown on the extreme left. Circle represents the results ofcontrol and square of spinocerebellar subjects.

( )A. Yamadori et al.rCognitie Brain Research 5 1996 4954 53

3.3.3. Recognition memory of the stimulus wordsThe SCD patients recognized 95.2% standard deviation

.4.7% of the tested repeated words, while the control .group recognized 84.7% standard deviation 11.2% . Thus

they even showed the superior recognition memory ofrepeated words against the control.

3.4. Comment

Unlike the Parkinsonian group, the SCD patients showeda significant lowering in WAIS IQ. Both the performanceIQ and verbal IQ were low.

However, AVLT as well as recognition memory of therepeated words used in the mirror reading test showednormal performance. The reading speed of standard wordswas also not significantly different from the control. Non-verbal intelligence reflected in RCPM score was alsowithin normal range.

These results led us to a conclusion that skill acquisitionof mirror reading in our SCD patients was impaired inde-pendent of declarative memory and articulatory agility.Since no selection of patients was attempted and thepatient group included degeneration of olivo-ponto-cere-bellar type as well as late cortical cerebellar atrophy type,the affected nervous system is not identical in all thesubjects. It is, however, reasonable to speculate that theretardation of acquisition of mirror reading skill was mostlikely related with the nervous system that was responsiblefor manifesting the cerebellar signs at the clinical level.The fronto-ponto-cerebellar systems might be one of thepossible substrata for this skill acquisition.

4. General discussion

We believe that we were able to demonstrate that theacquisition of mirror reading skill was selectively retardedboth on Parkinsons patients and spinocerebellar degenera-tion patients against normal recognition memory of thestimulus words employed in the mirror reading test. Theother important feature was, although retarded in acquisi-tion, once acquired, the skill was well retained for at leasta week in both groups.

At the earliest phase of the idiopathic Parkinsons dis-ease, affected brain area is believed to be confined to thesubstantia nigra, particularly the zona compacta. The ni-grostriatal fibers originate mainly from cells of this regionw x3 . Although there are many other connections to andfrom the substantia nigra, functional failure of thisdopaminergic nigrostriatal system is believed to be thecause of Parkinsons disease. It is difficult to attribute thisdeficit of skill acquisition to general cognitive disorders orspecifically to memory disorders, since our patients showednone of such signs. The skill building for mirror reading ofunique words must be based on more or less separablefunction somehow related with the nigrostriatal system. It

must also be remembered that the nigrostriatal system isfurther related with the frontal cortex in a complex wayw x8 .

However, the reported data are contradictory. For in-stance, Harrington and others reported that mirror reading

w xwas not impaired on their Parkinson subjects 10 . Similarw xreports followed 2,6 . The problem with these studies are

that test procedures were not identical making direct com-parison difficult. For instance, the skill acquisition was

w xevaluated only in a single day session 2,6 . In addition, thestage of the disease was not mentioned or controlled in allstudies. The nature of the stimuli employed may be an-other factor contributing to the difference. The above citedthree studies were performed for patients who use thealphabetic code. In this code, the letters that should bemirror reversed are 26, but in our Japanese study, letters,i.e. Hiragana phonographs in this study, that should bemirror reversed amounts to 46. This complexity mighthave contributed to clearer separation of mirror readingtask in our study.

In spinocerebellar degeneration, the subject populationwas less homogeneous. Thus it is quite probable that ourpatients may have had lesions outside the cerebellum, atleast in the brainstem especially in the olivary nuclei evenin the LCCA cases in their earliest stage. In more ad-vanced cases, degeneration in the pons, medulla, basal

w xganglia and thalami is expected 14 . Reflecting theserather diffuse involvement of the subcortical nervous sys-tem, mild lowering of general intellectual level has alwaysbeen reported in patients with spinocerebellar degenerationw x7,12 . Our present study was not an exception. Despite ofthis mild intellectual decline, it was clear that the patientsshowed good memory for words, demonstrating that thedeficit of acquiring mirror reading skill was out of propor-tion to other cognitive capacities.

Cerebellar participation in cognitive activities has beenw xspeculated 19 . Leiner and others thought that the cerebel-

lum contributes to some sensory, cognitive, linguistic andw xemotional aspects of behavior 16 . Their citation from a

clinical experience that implied Brocas area-cerebellarcircuitry is particularly interesting. Ito hypothesizes theexistence of a cerebellar microcomplex which acts as anadaptive controller. He postulates these microcomplexesare inserted in cortical systems providing adaptive-learning

w xcapacities 15 . Along with these lines of basic studies,clinical data also point to possible cognitive roles of thecerebellum. Thus, cerebellar cortical degeneration was re-

w xlated with deficits in cognitive planning 9 or cognitivew xfunctions involving sequences 18 . But there are also

authors who are cautious in inviting the cerebellum to thew xcognitive arena 1 . They concluded the deficit in executive

functions, i.e. initiation andror perseveration was a majorfactor of apparent cognitive deficits.

Last but not least, since the affected area is clearlydifferent between Parkinsons disease and spinocerebellardegeneration, it is difficult to attribute their mirror reading

( )A. Yamadori et al.rCognitie Brain Research 5 1996 495454

skill deficit to dysfunction of a single nervous system. Thecommon neural structure affected in both diseases may bethe frontal loop, but the evidence is still too subtle to jumpto such an assumption. Our result implies extreme com-plexity of the nervous organization of mirror reading skillacquisition.

Acknowledgements

This work was supported by a grant from the Ministryof Health and Welfare, Japan.

References

w x1 Appolonio, L.M., Grafman, J., Schwartz, V., Massaquoi, S. andHallet, M., Memory in patients with cerebellar degeneration, Neu-

.rology, 43 1993 15361544.w x2 Bondi, M.W. and Kaszniak, A.W., Implicit and explicit memory in

Alzheimers disease and Parkinsons disease, J. Clin. Exp. Neu- .ropsychol., 13 1991 339358.

w x3 Brodal, A., Neurological Anatomy in Relation to Clinical Medicine,Third Edition, Oxford University Press, New York, 1981, pp. 217218.

w x4 Cohen, N.J. and Squire, L.R., Preserved learning and retention ofpattern-analyzing skill in amnesia: Dissociation of knowing how and

.knowing that, Science, 210 1980 207210.w x5 Cohen, N.J., Eichenbaum, H., Deacedo, B.S. and Corkin, S., Differ-

ent memory systems underlying acquisition of procedural and .declarative knowledge, Ann. NY Acad. Sci., 444 1985 5471.

w x6 Daum, I., Schugens, M., Spieker, S., Poser, U., Schonle, P.W. andBirbaumer, N., Memory and skill acquisition in Parkinsons disease

.and frontal lobe dysfunction, Cortex, 31 1995 413432.w x7 Freedman, M., Cognitive deficits in olivopontocerebellar atrophy:

Implicationns for the cholinergic hypothesis of Alzheimers demen- .tia, Ann. Neurol., 24 1988 200206.

w x8 Gotham, A.M., Brown, R.G. and Marsden, C.D., Frontal cognitivefunction in patients with Parkinsons disease on and off Lev-

.odopa, Brain, 111 1988 299321.w x9 Grafman, J., Litvan, I., Massaquoi, S., Stewart, M., Sirigu, A. and

Hallett, M., Cognitive planning deficit in patients with cerebellar .atrophy, Neurology, 42 1992 14931496.

w x10 Harrington, D.L., Haaland, K.Y., Yeo R.A. and Marder, E., Proce-dural memory in Parkinsons disease: impaired motor but not vi-

.suoperceptual learning, J. Clin. Exp. Neuropsychol., 12 1990323339.

w x11 Hirayama, K., Diagnostic criteria and severity classification of( ) .spinocerebellar degeneration, Naika Internal Medicine , 55 1985

.13341336. in Japanesew x12 Hirono, N., Yamadori, A., Kameyama, M., Mezaki, T. and Abe, K.,

.Spinocerebellar degeneration SCD : cognitive disturbances, Acta .Neurol. Scand., 84 1991 226230.

w x13 Hoehn, M.M. and Yahr, M.D., Parkinsonism: Onset, progression, .and mortality, Neurology, 17 1967 427442.

w x .14 Ito, M., Sobue, I., Komatsuzaki, A. and Hirose, G. Eds. , Neurol-ogy of the Cerebellum, Igakushoin, Tokyo, 1986, pp. 224245.

w x15 Ito, M., A new physiological concept on cerebellum, Re. Neurol., .146 1990 564569.

w x16 Leiner, H.C., Leiner, A.L. and Dow, R.S., The human cerebro-cerebellar system: its computing, cognitive, and language skills,

.Beha. Brain Res., 44 1991 113128.w x17 Martone, M., Butters, N., Payne, M., Becker, J.T. and Sax, D.S.,

Dissociation between skill learning annd verbal recognitiuon in .amnesia and dementia, Arch. Neurol., 41 1984 965970.

w x18 Pascual-Leone, A., Grafman, J., Clark, K., Stewart, M., Massaquoi,S., Lou, J.-S. and Hallet, M., Procedural learning in Parkinsons

.disease and cerebellar degeneration, Ann. Neurol., 34 1993 594602.

w x19 Schmahmann, J.D., An emerging concept. The cerebellar contribu- .tion to higher function, Arch. Neurol., 48 1991 11781187.

w x20 Squire, L.R., Memory and Brain, Oxford University Press, NewYork, 1987.

w x21 Yamadori, A., Yoshida, T., Yamashita, H, Mori, E., Tabuchi, M.,Osumi, Y., Okawa, S. and Uehara, T., Procedural memory ofParkinsons patients, Annual Report of Research Group for Neu-ropsychological and Psychopathological Study on Dementia, Min-

.istry of Welfare, Japan, 1991, pp. 2125. in Japanesew x22 Yamadori, A., Yoshida, T., Yamashita, H, Mori, E., Tabuchi, M.,

Osumi, Y., Okawa, S., Uehara, T. and Shimada, K., Proceduralmemory of spinocerebellar degeneration patients, Annual Report ofResearch Group of Dementia, Ministry of Welfare, Japan, 1993, pp.

.232237. in Japanesew x23 Yamashita, H., Yoshida, T. and Yamadori, A., Mirror reading task

.in Parkinsons disease patients, Jap. J. Neuropsychol., 7 1991133138.