58ournal ofNeurology, Neurosurgery, and Psychiatry 1995;58:598-606

Comparison of executive and visuospatialmemory function in Huntington's disease anddementia of Alzheimer type matched for degree ofdementia

KW Lange, B J Sahakian, N P Quinn, C D Marsden, TW Robbins

AbstractGroups ofpatients with Huntington's dis-ease and probable dementia ofAlzheimertype (DAT) matched for level of demen-tia on the basis of mini mental stateexamination scores were compared inseveral tests of visual memory and testssensitive to frontal lobe dysfunction.Whereas recall of patients with DATtended to be worse on the Kendrick objectlearning test, the two groups were equiva-lent on tests of sensorimotor ability anddelayed matching to sample perfor-mance. By contrast, the patients withHuntington's disease were significantlyworse on tests of pattern and spatialrecognition, simultaneous matching tosample, visuospatial paired associates,and on three tests sensitive to frontal lobedysfunction-namely, the Tower ofLondon test of planning, spatial workingmemory, and a visual discriminationlearning and reversal paradigm. Theimpairments in these tests, however, didnot always qualitatively resemble thoseseen in patients with frontal lobe damageand may be more characteristic of pri-mary neostriatal deficit. In the visual dis-crimination paradigm the patients withHuntington's disease were significantlyworse than the patients with DAT at thesimple reversal stage, where they dis-played significant perseveration to thepreviously rewarded alternative. Theresults are consistent with the hypothesisthat patients with Huntington's diseaseexhibit deficits in tests sensitive to fron-tostriatal dysfunction and that this formof inteilectual deterioration is qualita-tively distinct from that seen inAlzheimer's disease.

(_ Neurol Neurosurg Psychiatry 1995;58:598-606)

Keywords: Huntington's disease; Alzheimer's disease;cognition

There is currently much interest in whetherdementias arising from different causative fac-tors exhibit qualitatively different patterns ofcognitive deficit. This is an issue of someimportance, not only in practical terms, butalso for understanding the underlying neuralsubstrates of the disorders and the specificity

of cognitive deficits over and above gener-alised intellectual deterioration due to demen-tia. Previously, Albert et all and Cummingsand Benson2 have proposed a distinctionbetween dementia associated with primarysubcortical neuropathology (for example,basal ganglia disorders such as Parkinson'sand Huntington's diseases) and dementia inwhich the primary neuropathology is in theassociation neocortex (for example, dementiaof Alzheimer type (DAT)). This distinctionbetween "subcortical" and "cortical" demen-tia has proved somewhat controversial, notleast because clinical rating scales have some-times failed to support it, but also because thenature of some of the cognitive deficits seen insubcortical dementia may be reminiscent ofthe symptoms of frontal lobe damage. Theseare often referred to as examples of executivedysfunction, because they present as problemsof planning, attentional set shifting, and slow-ness of response, rather than the typical signsof posterior cortical damage which lead tosuch symptoms as aphasia, agnosia, andglobal amnesia. It seems likely that striataldysfunction, as occurs in Parkinson's diseaseand Huntington's disease, would lead to simi-lar executive deficits because of the existenceof several corticostriatal loops that highlightthe functional inter-relationships between dif-ferent parts of the frontal cortex and the basalganglia.3 Of particular significance forHuntington's disease in view of the primarysite of its striatal neuropathology may be theanatomical and functional relations that existbetween the caudate nucleus and the pre-frontal cortex.3 Despite these neurobiologicalconsiderations and evidence that impairmentof daily functioning in patients with earlyHuntington's disease more likely results fromdeficits in executive function than motor dis-ability,4 there has been relatively little analysis ofexecutive dysfunction in Huntington's disease.5

Explicit comparisons among these differentforms of dementia have also been lacking6although there have been several recent com-parisons of Huntington's disease and DAT ondifferent forms of memory. Hodges et a17compared a group of patients with DAT and agroup with Huntington's disease, matched foroverall level of dementia, on a battery of testsof semantic and episodic memory. Whereasthe patients with DAT were significantly moreimpaired on measures of delayed verbal andfigural episodic memory, patients with

Department of ClinicalNeurology, Institute ofNeurology, UniversityofLondon, London,UKKW LangeN P QuinnC D MarsdenDepartment ofExperimentalPsychology, UniversityofCambridge,Cambridge, UKB J SahakianTW RobbinsCorrespondence to:Dr B J Sahakian,Department of Psychiatry,University of Cambridge,Addenbrooke's Hospital,Hills Road, Cambridge CB22QQ, UK.Received 23 June 1994and in revised form27 October 1994Accepted 16 December 1994

598

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from

Comparison of executive and visuospatial memory function in Huntington's disease and dementia ofAlzheimer type matchedfor degree ofdementia 599

Huntington's disease did worse on letter flu-ency, often associated with frontal lobe dam-age. There have been few comparisonsbetween patients with Huntington's diseaseand DAT on tests of executive function,5although there is growing evidence for differ-ences between DAT and Parkinson's diseasein this regard. For example, Litvan et all8showed that demented patients withParkinson's disease perform substantiallyworse than patients with DAT on tests ofexecutive function, such as the ability to per-form the Wisconsin card sorting test(WCST), whereas semantic and episodicmemory were significantly worse in thepatients with DAT.We have found that patients with

Parkinson's disease who are early in thecourse of the disease and yet to take medica-tion are significantly worse than patients earlyin the course of DAT in tests of visual dis-crimination learning and attentional set shift-ing analogous to the WCST.9 '° As we havealso shown that the attentional set shifting testis sensitive to damage to the frontal cortex,"we have hypothesised that patients withParkinson's disease may be especially prone tofrontal type deficits early in the course of thedisease. 12 Whereas patients with DAT can

also be expected to be impaired in tests ofexecutive function as the disease progresses,both neuroimaging and neuropsychologicalevidence supports the hypothesis that anteriorcortical functions are relatively more immuneto disruption." For example, a recent studyshowed that whereas patients with DAT can

be severely impaired on a test of spatial work-ing memory that is sensitive to frontal lobedysfunction, they do not necessarily failbecause of the impaired strategy that has beenshown to account for much of the frontaldeficit.'4

Consequently, the main aim of this studywas to compare patients with Huntington'sdisease and DAT matched for level of demen-tia on tests of executive function that we havepreviously shown to be selectively sensitivenot only to frontal lobe damage" '1 but also tobasal ganglia disorders other thanHuntington's disease, such as Parkinson's dis-ease, progressive supranuclear palsy, and mul-tiple system atrophy.'6 17 The first hypothesisto be tested was thus that such tests of execu-tive function may be more sensitive to cogni-tive deficits in Huntington's disease thanDAT. Whereas patients with basal gangliadisorder do exhibit selective impairments on

these tests, previous evidence has indicatedthat the precise nature of the deficits may differqualitatively compared with those seen afterfrontal lobe lesions, perhaps reflecting the dif-ferent contributions of the striatal and corticalcomponents of the common neuroanatomicalcircuitry, termed corticostriatal loops.'2 17

Thus a secondary hypothesis was that thenature of any executive deficits seen in theHuntington's disease group might be differentfrom those found after frontal lobe damage.

There were several aspects of the studyrequiring careful control. Firstly, to assess the

selectivity of the executive deficits inHuntington's disease, we used several testsfrom the Cambridge neuropsychological testautomated battery (CANTAB) that haveexplicitly been used to assess visuospatiallearning and memory in Alzheimer's disease,'8some of which have been shown to be moresensitive to temporal rather than frontal lobedamage.'9 Assessment of visual memory andlearning in Huntington's disease has also notpreviously received much attention.5Secondly, it was important to match for thedegree of clinical dementia so that any differ-ences in specific cognitive functions could notbe attributed to non-specific intellectual deter-ioration. For the purposes of matching weused a validated measure of global dementia,the mini mental state examination (MMSE).20Another screening instrument for dementia,Kendrick's object learning test (taken fromthe Kendrick cognitive tests for the elderly)2'was also used and it was possible to matchpatients according to a premorbid index ofIQ.22 To gauge the absolute levels of impair-ment in the Huntington's disease and DATgroups we have also provided data for sepa-rate large groups of normal controls matchedas far as possible for age and premorbid esti-mates of verbal IQ to the DAT andHuntington's disease groups. As might beexpected from the recent standardisation21 ofmuch of the CANTAB battery, the perfor-mance of the younger control group wasslightly superior to that of the older group oncertain tests. Whereas this might affect the rel-ative degree of age related impairmentbetween the Huntington's disease and DATgroups, it could not account for any inferiorlevels of executive performance in theHuntington's disease group in the hypothesisunder test. Therefore, for the purposes of thisstudy we focused on a direct comparison ofthe performance of the two patient groups.

MethodsSUBJECTSConsecutively presenting patients withHuntington's disease or mild or moderateprobable DAT were assessed. In all, 17patients with probable DAT and 13 withHuntington's disease were tested. Completedata were only obtained on 13 (83%) patientswith DAT and 10 (77%) patients withHuntington's disease, because of fatigue andlack of comprehension of the instructions onsome tests. As most patients with DAT wereprimarily in the early stages of the disease withmainly mild symptoms, it was feasible tomatch the two groups for level of dementia, asassessed with a clinical rating scale (see later).

Patients with dementia of the Alzheimer's typeAll thirteen patients attended the MaudsleyHospital and all were seen on an outpatientbasis through the memory clinic.24 Patientswere assessed and diagnosed by a consultantpsychiatrist in concordance with a secondpsychiatrist as having DAT following criteriaoutlined for "probable Alzheimer's disease"

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from

Lange, Sahakian, Quinn, Marsden

Table 1 Clinical characteristics ofpatient group

Sex DiseaseNo Age M:F duration NART KOLT MMSE

HD 10 51-5 (3 3) 3:7 7-4 (0 9) 106-6 (4 9) 18-4 (2-4) 20-3 (1-4)DAT 13 69-2 (2 0) 9:4 3-67 (0 33) 111 (2-1) 13-0 (1-5) 19 2 (1-6)t (21) 4-81** 0-98 2-00 0-53

**P<0-01.Data shown are mean (SEM) values; NART = National adult reading test22; KOLT = Kendrickobject learning test2; MMSE = mini mental state examination20; HD = Huntington's disease;DAT = dementia of Alzheimer type.

by McKhann et al.25 All patients were consid-ered to be in the mild and moderate stages ofthe disease, with eight patients in stage 1 andfive patients in stage 2 of the clinical dementiarating (CDR) scale, in which CDR stage 1 ismild, CDR stage 2 is moderate, and CDRstage 3 is severe dementia, as defined byHughes et al.26 To determine the diagnosis,patients were first seen by a psychiatrist onattending the memory clinic, where theyreceived a semistructured clinical interviewand a physical examination.27 After the physicalexamination, which included the testing ofvisual acuity with Snellen cards, blood wastaken for laboratory investigations.Unenhanced CT was then performed with theMaudsley Hospital GE 9800 scanner to deter-mine the extent of brain atrophy, and excludeother forms of dementia, especially multi-infarct dementia. The patients (nine men andfour women) included in the study were thenseen on a subsequent visit by a clinical psy-chologist for detailed neuropsychologicalassessment, including the tests described inmore detail later. Table 1 gives further detailsof the 13 patients.

Patients with Huntington's diseaseTen patients with Huntington's disease wereincluded in the study. The diagnosis wasbased on personality or intellectual change,together with the presence of chorea or thetypical impairment of voluntary movement.These symptoms were not present at birth,were insidious in onset, had gradually becomeworse, and were associated with a definitefamily history of at least one other memberwith these characteristic neurological symp-toms and signs. At the time of investigation,the severity of the hyperkinetic involuntarymovements was rated according to the choreascale of the quantified neurological examina-tion, on a scale ranging from 0 to 25 points28;this patient group scored a mean of 14-8(SEM 1 1). None of the patients had signifi-cantly impaired visual acuity or hearing. Atthe time of the study, the patients did notexhibit acute confusion or acute psychosis andwere not receiving any medication known toaffect the CNS. Informed consent wasobtained from all patients participating in thestudy.

ControlsThe main purpose of the study was explicitlyto compare patients with Huntington's dis-ease and those with DAT. To gauge the rela-

tive degree of age related impairment, controldata for the main dependent variables havebeen provided from a large group of healthynormal control subjects (table 2). These weredrawn from a large population of volunteersaged between 55 and 69 years from London,Cambridge, and Newcastle-upon-Tyne(North-East Age Research Panel) in therecently published standardisation of many ofthe tests of the CANTAB battery. The con-trols had no history of neurological or psychi-atric disorder.

MATERIALS AND PROCEDURESThe main testing procedures were taken fromthe Cambridge neuropsychological test auto-mated battery (CANTAB), a series of com-puterised paradigms run on an Acorn BBCMaster microcomputer with a high resolutionMicrovitec colour monitor and a Microvitec(Touchtech 501) touch sensitive screen. Thegeneral rationale and utility of the battery forclinical investigations has recently beenexplained in some detail.29 Subjects wereseated roughly 0 5 m from the monitor and itwas explained that they would have torespond to stimuli by touching the screen.They were introduced to the apparatus byway of a sensorimotor screening task in whichthey were asked to respond to a series of flash-ing crosses on the screen by placing the indexfinger of their preferred hand on the centrepoint of each cross. The finger had to be heldin place for six seconds at which time anothercross appeared. After a short demonstrationseries in which three consecutive crosses weretouched, subjects were presented with a seriesof 10 crosses to touch at six second intervals.Measures were taken of the latency to point(to the nearest 0-01 s) and the accuracy of theinitial pointing contact with the screen (com-puted in terms of vectorial displacement fromthe centre of the cross in arbitrary units basedon X-Y co-ordinates). No subjects wereexcluded on the basis of their deficient perfor-mance on this task.

Tests of visual memory and learningIn the first part of the test session, all patientsreceived computerised tests of visual memoryand learning identical to those previouslydescribed in detail.'8 These included tests ofpattern and spatial recognition, simultaneousand delayed matching to sample, and a condi-tional visuospatial associative learning task. Inthe pattern recognition task, abstract visualstimuli are displayed one by one in the centreof the screen before recognition memory istested in a two alternative forced choice task,where the original stimulus is presented with anovel one and the subject has to pick the orig-inal. The test consists of two lists of 12 stimulieach, comprising 24 in all. In the spatialrecognition task, open white squares are pre-sented in a number of spatial locations on thescreen, before recognition memory for loca-tion is tested, again with a two alternativeforced choice procedure. There were fourblocks of five stimuli each, comprising 20 inall. In the matching to sample test, the first

600

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from

Comparison ofexecutive and visuospatial memoty function in Huntington's disease and dementia ofAlzheimer type matchedfor degree ofdementia 601

phase (perceptual matching) consists of thesimultaneous presentation of a complexabstract stimulus, varying in both colour andshape, with four matching stimuli. Subjectsare required to choose the matching stimulusin a four alternative forced choice procedure.In the second phase (memory matching), thematching stimuli are presented in five trialblocks after a delay of 0, 4, 8, or 16 seconds,followed finally by a retest at zero seconds.This delayed matching to sample test thusassesses the forgetting of the stimuli to berecognised after a short delay.The conditional visuospatial paired associ-

ates test is identical to one described as thedelayed response task in Sahakian et al. 18Here, however, the emphasis was on thelearning as distinct from mnemonic aspects ofthe task. Abstract visual stimuli were pre-sented one by one in a ring of six (and in thelast stage, eight) boxes arranged around thescreen. The stimuli were then re-presentedone by one in the centre of the screen and thesubject indicated in which box they had beenpreviously presented. Subjects were given up

to 10 trials to learn the location of each stimu-lus of the set before proceeding to the nextstage. Failure to reach criterion at any stageresulted in the premature termination of thetest. The test began with only one stimuluslocation to be remembered and then graduallyincreased to two, three, six, and eight stimuli.

Tests ofplanning and spatial working memorySpatial short term memory task-In this com-

puterised Corsi block tapping task,'0 spatialshort term memory capacity was determinedfrom the ability of subjects to remember a

sequence of squares on the screen, as des-cribed in detail in an earlier publication.'5 Spatialshort term memory span was calculated as thehighest level at which the subject successfullyrecalled at least one sequence of boxes.

Spatial working memory task"-In this taskthe subject was required to search throughspatial arrays of boxes to find tokens. Impor-tantly, once a blue token had been foundwithin a particular box, then that box wouldnever be used to hide another token. On eachtrial, the total number of blue tokens to befound corresponded to the number of boxeson the screen as every box was used just once.

Errors were scored according to the num-

ber of occasions on which a subject returnedto open a box in which a blue counter hadalready been found. After four practice trialswith three boxes, there were four test trialswith each of four, six, and finally eight, boxes.The patient groups were compared in termsof the total number of errors summed acrossthe 12 test trials.

Planning task-This is a modification of theTower of London task" in which the subjecthas to move coloured "balls" on the screenfrom an initial arrangement to one corre-

sponding to the goal arrangement shown inthe top half of the screen, as described indetail previously.'5The starting position of the balls was varied

such that in any particular trial the solution

could only be reached after a minimum oftwo, three, four, or five moves. Subjects wereinstructed to examine the position of the ballsat the beginning of each problem and attemptto solve it in the minimum possible number ofmoves. This was both given to them verballyand displayed on the screen throughout eachtrial. They were encouraged not to make thefirst move until they were confident that theycould execute the entire sequence needed tosolve the problem. The maximum movesallowed corresponded to twice the minimumnumber possible plus one, or plus two in thecase of "five move" problems. The latencymeasures for thinking time previouslyemployed to assess neurosurgical cases'5 andpatients with basal ganglia disorders'2 are notreported in detail here.

Visual discrimination/attentional set shiftingparadigm-The computerised visual discrimi-nation/attentional set shifting paradigm hasbeen described in detail elsewhere.9 Briefly,subjects are trained on a series of visual dis-criminations that vary in two perceptualdimensions, one of which is relevant and oneof which is irrelevant, on the basis of feedbackprovided automatically by the computer. Atcritical points subjects are required first tomaintain attention to different examplarswithin the same dimension (intradimensionalshift) and then to shift attention to the previ-ously irrelevant dimension (extradimensionalshift). The other stages are defined andexplained in previous papers9 and in theresults section. For each stage, continuationto the next one was dependent on a criterionof six successive correct responses beingreached. If criterion was not reached at the50th trial of a stage, then the test was discon-tinued and subjects did not proceed to thenext stage. More detailed explanation andrationales for the exact design of the test canbe found in previously published articles.9'0 32

STATISTICAL ANALYSISFor most of the dependent variables, analysisof variance (ANOVA)33 was used to comparethe Huntington's disease and DAT groups.Data were transformed where appropriate(when there was a positive skew on latencyvariables). For most of the test variables, theANOVA model was a two factor design thatincluded a between subjects factor (group)and a within subject factor (for example, diffi-culty level). Student's t tests were applied totest the significance of differences betweenmeans where factorial ANOVA was notrequired.

For the attentional set shifting task, thedata for the numbers of subjects passing orfailing each stage of the test were cast intocontingency tables and analysed by the likeli-hood ratio method.34 35 This method is particu-larly useful (a) for analysing data with smallcell frequencies, as occurs in some of thedata to be presented, and (b) for partitioninginhomogeneities in the contingency table byadditive, orthogonal contrasts. The resulting"information" statistic (2i) is distributed as x2.

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from

Lange, Sahakian, Quinn, Marsden

Table 2 Means (SEM) for the normal control samples

For Huntington's disease group For DATgroup

Age (y)NARTSpatial spanSpatial working memory

(Between search errors)Tower of London test

(perfect solutions)

Age (y)NARTPattern recognition/24Spatial recognition/20Matching to sample:SimultaneousDelays Os

4s8s16sOs (retest)

Paired associates learning:Total trials to criterion

Working memory and planning battery(n = 55) (n = 222)55-6 (0-6) 67-3 (0-1)

116-5 (1-0) 117-9 (0-4)5-3 (0-1) 5-0 (0-1)

39-6 (3 2) 46-1 (0 9)

8-2 (0 2)/12 7-5 (0-2)/12Visual memory battery

(n = 46) (n = 39)56-7 (0 5) 66-1 (1-0)

117-5 (1 1) 114-2 (1-6)20-9 (0 4) 19-8 (0 4)16-4 (0 3) 16 5 (0 4)

4-85 (0-1)/5 4-87 (0 05)4 30 (0 06) 4-61 (0-05)4-20 (0-08) 4-67 (0-15)4-20 (0 20) 4-26 (0-15)4-40 (0 19) 4 41 (0-11)4 70 (0-17) 4-58 (0-11)

13-3 (0-4) 14-2 (0-4)

Table 3 Experimental neuropsychological tests (mean scores (SEM))

Huntington'sdisease DAT Significance (dl)

Sensorimotor screening:Accuracy (error score) 43-3 (6-6) 49-1 (3-0) t(21) = 0 9, NSLatency (s) 2-24 (0 39) 1-71 (0 23) t(21) = 1-2, NS

Visual memory:Pattern recognition (/24) 11 9 (1-01) 14-5 (0-76) t(21) = 2.12*Spatial recognition (/20) 10 4 (0-56) 12 8 (0-62) t(21) = 2-74**Spatial span 2-7 (0-15) 4-0 (0-4) t(21) = 3-08**

Matching to sample:Simultaneous/5 4 0 (0 3) 4-8 (0-1) t(21) = 2.61*

Delay/5:0 s 3-1 (0-4) 3-3 (0-4)4 s 33 (03) 3.7 (03)8 s 2-5 (0-1) 2-5 (0-4)16s 1-8(02) 25(04)

Group (HD v DAT)F(1,21) = 0-9Group x delay,F(3,63) = 0-6

0 s retest 3 0 (0 3) 3-8 (0 3) t(21) = 1-63Paired associates, visuospatial learming:

Total trials to criterion 46-4 (4-0) 38-5 (4 3) t(21) = 1-31

*P < 0-05; **P < 0-01.

ResultsCOMPARISON WITH MATCHED CONTROLGROUPSFor all the measures to be described, bothpatients with Huntington's disease and thosewith DAT were significantly worse than theirrespective control groups of normal volun-teers with comparable ages and premorbidestimated IQ (using the national adult readingtest (NART); table 2).

SENSORIMOTOR SCREENING TESTThe two groups did not differ significantly oneither of the two indices of performance,latency and accuracy of pointing (table 3).

VISUAL MEMORY AND LEARNING (TABLE 3)The patients with Huntington's disease weresignificantly worse on both tests of patternand spatial recognition memory. The patientswith Huntington's disease were also worse onsimultaneous matching to sample, but the dif-ference disappeared during the delay condi-tions (0, 4, 8, 16 seconds and retest at zeroseconds), although, as expected, performancedeclined significantly over the delays for thetwo groups overall (F(3,63) = 9-84, P <0 001). For the paired associates, conditionallearning test, the patients with Huntington'sdisease took more trials to reach criterion forthe test as a whole over all stages, but this dif-ference was not statistically significant. Interms of patients reaching criterion at eachstage, however, none of the 10 patients withHuntington's disease successfully passed thetest when the locations of six patterns had tobe remembered, but a significantly higherproportion (five out of 13) of the patients withDAT were successful at this stage QX2 = 6-76,d f = 1, P < 0 01).

Figure 1 Mean (SEM)values for between andwithin search errors on thespatial working memorytestfor Huntington'sdisease (HD) and probabledementia of the Alzheimertype (pDAT) groups.

120 e-

100 H

I- HD= pDAT

80 H

60

40

20H

oBetween errors Within errors

TESTS OF SPATIAL WORKING MEMORY ANDPLANNINGSpatial spanThe patients with Huntington's disease had sig-nificantly shorter spatial span scores (table 3).

Spatial working memoryFigure 1 shows that the patients withHuntington's disease committed significantlymore between and within search errors in thistask (F(1,21) = 10-03, P < 0 005). There wasalso a significant patient group x set sizeinteraction (F(3,63) = 5 09, P < 0 005) as thedifference between the groups only emergedafter the difficulty level of three boxes.

Tower ofLondon testThe analyses concentrated on three measuresof accuracy for the first block of problems.Figure 2 shows that the patients withHuntington's disease were significantly worsein terms of "perfect solutions" at each stage ofdifficulty (across two, three, and four moveproblems) (main effect of patient group, t(2 1)= 3-08, P < 0-001) . Two other measures ofaccuracy, total solutions and excess moves,failed to reach significance. Mean (SEM)scores respectively were: HD; 5 3(0 4),8-0(1-3): DAT; 4 7(0 4); 6X8(1X7).

u)o

I-41)

602

T

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from

Companison of executive and visuospatial memory function in Huntington's disease and dementia ofAlzheimer type matchedfor degree ofdementia 603

Figure 2 Tower ofLondon planning test.Proportion ofproblemssolved in the minimumnumber of moves at eachstage of difficulty (upperpanel) and percentagecorrect (lower panel).Abbreviations asforfig 1.

Figure 3 Performance inthe visual discriminationparadigm. The proportionofpatients successfullyreaching criterion (616) ateach stage of the test isshown. SD = simplediscrimination stage; SDR= simple discriminationreversal; C-D = compounddiscrimination withseparate elements; CD =

compound discrimination;CDR = compounddiscrimination reversal;IDS = intradimensionalshift; IDR =

intradimensional reversal;EDS = extradimensionalshift, EDR =

extradimensional reversal.Other abbreviations as forfig 1. See Downes et al 9forfurther explanation. Notethe selective impairment inreversal (SDR) learning inthe Huntington's disease(HD) group.

1.0

0.8

40

C)

o

c

0

0.

0

a-

CL

0.6

0.4

0-2 e-

0

80

70

0)oc)a)

0)

a)

* HDo pDAT

2 3Difficulty

60

50 _-

40 _-

30 _-

20 _

10

o0Total perfect solutions

80K

c

0

a).r

C1)0

M

c)Cu

co

0)

a)0)

C1)01)0-

* HIo PE

60

40

20

0

SD SDR C-D CD CDR IDS IDR EDS EStage

Visual discriminationlattentional set shifting testFigure 3 shows that the patients withHuntington's disease performed significantlyworse at the series of visual discriminations.Although the two groups performed equiva-lently on the test of simple visual discrimina-

41 tion (one out of 13 patients with DAT failedat this stage), as many as six out of 10 of thepatients with Huntington's disease failed atthe next stage, of simple discrimination rever-sal, when the previously correct stimulusbecame incorrect and vice versa. This differ-ence was significantly different (X2(1) = 7d17,P < 0-01). In addition, the nature of the fail-ure of these six patients with Huntington's

L| disease was particularly illuminating because4 the average number of errors made for the 50

trials was 42-5 (SEM 3-3), indicating a highdegree of perseveration of responding towardsthe initially correct stimulus. The rate of attri-tion of subjects across the remaining stages ofthe test was parallel and roughly equivalent.The poor performance of both groups can beseen from our unpublished observations thatmost (about 70%) of 342 elderly normalcontrols successfully negotiate all parts ofthe test, most of the failures normally occur-ring at the extradimensional shift stage. Inthis large sample, there was no significant dif-ference between the younger (50-59) andolder (60-69) control groups (see also Owenet al!2 for representative data on a smallersample).

DiscussionThis neuropsychological comparison of DATand Huntington's disease has shown that,when matched for level of dementia, patientswith Huntington's disease are significantlyinferior to patients with DAT in three tests ofcognitive function (spatial working memory,visual discrimination/attentional set shifting,and the Tower of London planning task) pre-viously shown to be sensitive to frontal lobedamage and basal ganglia dysfunction. The

D results are thus consistent with hypotheses)AT that the neural substrates of many of the cog-

nitive deficits in Huntington's disease arecentred on the caudate nucleus3638 but thatadditional cortical atrophy may also be signifi-cant.37The neural specificity of the three main

cognitive tests may be gauged from the factthat the Tower of London task is insensitiveto temporal lobe excisions and amygdalohip-pocampectomy,'9 as well as more generally to

-o posterior cortical damage.3' Performance onthe visual discrimination/attentional set shift-ing paradigm has similarly been shown not tobe affected by temporal lobe excisions oramygdalohippocampectomy,19 and both ofthese lesions have much less effect on perfor-mance on the spatial working memory taskthan either frontal lesions" or basal ganglia

-0 dysfunction.9 17 By contrast, certain other testsfrom the CANTAB battery such as visual pat-tern recognition or delayed matching to sample

:DR are more affected by lesions of the temporallobe structures than frontostriatal damage.'9

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from

Lange, Sahakian, Quinn, Marsden

Therefore, the more pronounced executivedeficits shown by the patients withHuntington's disease have a degree of neuralspecificity, consistent with the existence ofgreater frontostriatal pathology inHuntington's disease than in DAT.The inferior performance of patients with

Huntington's disease was not limited to testsclearly requiring executive function, as theywere also significantly impaired relative topatients with DAT in tests of perceptualmatching to sample, a visual pattern recogni-tion task, a spatial recognition task, spatialspan, and in learning sets of visuospatialpaired associates. Performance on a difficultvisual delayed matching to sample task wasalso impaired relative to controls, but not to agreater extent than for patients with DAT.Similar results have been reported for patientslate in the course of Parkinson's disease'839possibly due to additional pathology that mayaffect functioning in cortical regions otherthan the prefrontal cortex, such as the tempo-ral lobe. The greater magnitude of thesedeficits in Huntington's disease than in DAT,however, suggests an alternative (though notmutually incompatible) possibility, that thelate cognitive deficits in Parkinson's diseaseand Huntington's disease reflect pathology ofthe striatum, most likely the caudate nucleus.

Before comparing these results with others,obvious explanations of the differences shouldbe considered. Although the patients withHuntington's disease and those with DATdiffered in age, the greater deficits in the caseof the Huntington's disease group cannot beartefacts of relative impairments produced byage related changes in the normal population,as the Huntington's disease group wasyounger and control performance on thesetests usually deteriorates slightly with age.Nor can it be said that the patients withHuntington's disease were globally impairedon all tests. Thus accuracy was apparentlyequivalent in delayed matching to sample forpatients with Huntington's disease and DATwhen visual patterns had to be retained over ashort delay. The two groups were carefullymatched for degree of dementia usingFolstein's MMSE, and the patients with DATperformed almost significantly worse than thepatients with Huntington's disease onKendrick's object learning test (taken fromthe Kendrick cognitive tests for elderly peo-ple). An explanation of the results in terms of agreater motor deficit in Huntington's diseasesomehow interfering with cognitive perfor-mance of the touch screen tasks also seemsmost unlikely. Both accuracy and latency ofresponding in the sensorimotor screening test,which required the subject to point to selectedlocations on the screen, were equivalent in thetwo groups. Furthermore, many of thedeficits, such as the gross impairment of simplereversal in the patients with Huntington's dis-ease, have to be set against their relatively suc-cessful performance in the simplediscrimination phase of the task.

These results are consistent with severalexamples of double dissociations between

Huntington's disease and DAT in otherdomains of cognitive function. Patients withHuntington's disease are generally superior topatients with DAT on tests of verbal recogni-tion,40 episodic memory,7 some aspects ofsemantic memory,74' and implicit (priming)memory,42 retrograde amnesia,4344 visuocon-structional performance45 (though see Hodgeset at7), but inferior on tests of mental arith-metic,5 procedural learning,46 47 retrieval (suchas letter and category fluency7 48), and egocen-tric perception.45 Without attempting toaccount for the whole range of cognitivedeficits in Huntington's disease, this patternof deficits is broadly consistent with thegreater involvement of frontostriatal than pos-terior cortical pathology in this disorder (seealso Brandt and Bylsma5), as the data pre-sented here imply. It was nevertheless inter-esting to find that the patients withHuntington's disease were more impaired insome of the tests of visual memory and learn-ing such as pattern and spatial recognition,and visuospatial paired associates learningthan the patients with DAT, although thereverse was the case for the Kendrick objectlearning test, which involves the verbal recallof pictures of everyday objects. The relativeimpairment in pattern and spatial recognitionmemory in Huntington's disease may seem tocontradict earlier results,40 but the patientswith Huntington's disease and those withDAT were not explicitly matched for degreeof dementia in that earlier study. Moreover,the relative impairment in pattern recognitionmemory in Huntington's disease was notfound in the related test of delayed matchingto sample. The impairment in simultaneousmatching to sample in the Huntington's dis-ease group relative to DAT, supports the find-ing of visual perceptual impairmentsunderlying naming deficits.49 The difference isalso reminiscent of other impairments in thistest that we have reported in patients withbasal ganglia disorders, compared with spar-ing in mild DAT. 18 39The relatively novel feature of the present

study is its focus on tests sensitive to executivecontrol over performance and to frontal lobedysfunction, aspects of cognition inHuntington's disease that have been the sub-ject of only a few investigations.5'5' In theintroduction, it was pointed out that deficitsin tests of executive function may neverthelessdiffer qualitatively between patients withfrontal lobe lesions and basal ganglia disorderssuch as Parkinson's disease. For example,demented patients with Parkinson's diseasehave been found to have excessive problemswith the Wisconsin card sorting test.8 Closerinspection of this result, however, shows thatthis deficit was restricted to the initial phase,when subjects are striving to attain the firstset, rather than to switching sets. In the pre-sent study, deficits for patients withHuntington's disease also occurred early inthe visual discrimination paradigm thatdivides the Wisconsin card sorting test intomore elementary cognitive components,notably at the simple reversal stage.

604

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from

Comparison of executive and visuospatial memory function in Huntington's disease and dementia ofAlzheimer type matchedfor degree ofdementia 605

Impairments at this early stage of the test arenot necessarily hallmarks of frontal lobe (orexecutive) dysfunction. In anatomical terms,whereas reversal learning in animals isimpaired by damage to the medial orbito-frontal cortex5" and the region of the striatumto which it projects,54 such deficits are alsoknown to occur-for example, after damageto the amygdala.55 A more telling piece of evi-dence was that the performance of patientswith Huntington's disease, unlike those withDAT was very perseverative in nature, possi-bly implicating dysfunction of the orbito-frontal cortex and associated outflow via thecaudate nucleus. Jones and Mishkin5" foundthat lesions to the orbitofrontal cortex pro-duced just this form of deficit in monkeys,whereas lesions to the amygdala produced morepervasive deficits in learning new stimulus-reward associations. There is evidence thatthis form of perseveration in reversal learningmay be distinct from failures to shift atten-tional set at the extradimensional shift stage,55which we have found to be a sensitive indicatorof cognitive impairment in patients early inthe course of Huntington's disease (unpub-lished findings), as well as in other forms ofbasal ganglia disease, including Parkinson'sdisease, Steele-Richardson-Olsewski syn-drome, and multiple system atrophy.'6 17Other forms of perseveration, such as occur intests of verbal fluency, are not apparently amajor feature of Huntington's disease.5 TheHuntington's disease deficit in visual discrimi-nation reversal learning is of further theoreti-cal interest because it can be viewed as adifficulty in learning new stimulus-responsehabits, theorised by some authors56 to be anexpression of the procedural learning that hasbeen postulated to form part of the functionof structures within the basal ganglia. Itseems, however, that the deficit is not somuch one of acquiring new visual habits, as ofinhibiting previously acquired ones.The impairments in spatial working mem-

ory were also consistent with a pronouncedpropensity for perseveration, as the number ofbetween search errors (inappropriately return-ing to a previously successful choice) was veryhigh, as great numerically as in any other dis-order we have examined, including frontallobe damage. This perseverative tendency wasalso manifest in terms of repeated, althoughineffective, responses within a searchsequence (within search errors). Studies withgroups of neurosurgical patients with frontallobe lesions1I and other groups with basal gan-glia disease'6 17 as well as experimental animalsin directly analogous tasks55 support the viewthat the pattern of performance on this taskmost likely reflects dysfunction of frontostri-atal circuitry.

It was hardly surprising to find planningperformance in Huntington's disease to beseverely affected, given that componentcapacities, such as spatial span and spatialworking memory were themselves so severelyimpaired. The deficit in planning was mani-fest even on the easy problems (for example,"two movers"), although extensive practice

with one and two move problems ensured thatthe Huntington's disease and DAT groupsunderstood the requirements of the task.Impairment at such an early stage, however,suggests that it is not truly a planning dys-function as such, but represents a failure ofperceptual schemas (standard arrangementsof the set of three balls leading to routinechoices) accurately to elicit the correct combi-nation of moves. This requires further investi-gation, because it is reminiscent of a failure incontention scheduling among sensory cuedresponse options described by Shallice31rather than a loss of supervisory function,equivalent to executive control. In neuralterms these two mechanisms have been linkedrespectively to basal ganglia57 and frontal lobefunction31 and so the Huntington's diseasedeficit is probably consistent with its wellestablished striatal (especially caudate)pathology. In general, such differences in theprecise pattern of deficits in the tests of execu-tive function between patients withHuntington's disease and those with frontallobe damage are consistent with the secondaryhypothesis outlined in the introduction thatthe precise nature of the deficits in thesepatient groups may depend on the distinctways in which normal functioning of the cor-ticostriatal loops is disrupted.

Overall, it is apparent that the cognitivedeficits in the later stages of Huntington's dis-ease can be every bit as wide ranging as inmoderate DAT, including not only executivefunctions, but also fundamental forms of pro-cessing that further undermine the operationof executive processes. Even at this later stage,however, it is clear that there exist qualitativedifferences in cognitive function betweenthese conditions. Presumably, these differ-ences would be accentuated in the earlieststages of both diseases, and the future focus ofwork on Huntington's disease should perhapsconcentrate on its early stages, and onpatients at risk for the disease. Using the samebattery of tests described here in a series ofcross sectional comparisons of patients withParkinson's disease at various stages, we havereported9 12 that the tests employed here thatare sensitive to frontal lobe dysfunction arethe most sensitive to early Parkinson's disease,whereas later stages of the disease are associ-ated with significant declines in visual mem-ory and learning39 that may be related to theprogressive neuropathological changes occur-ring in Parkinson's disease. It is to be hopedthat a similar cross sectional or alternatively,longitudinal, study of Huntington's diseasewould help to relate different aspects of thecognitive impairments associated with the dis-ease to discrete components of the underlyingneuropathology.

This work was supported by a programme grant from theWellcome Trust. We thank Professor R Levy for referrals ofpatients with DAT, Professor P Rabbitt for providing controlsubjects, and Merle James, Andrew Lawrence, and Dr AdrianOwen for their assistance. We are also grateful to the residentsand staff of the Sue Ryder Home, Stagenhoe Park,Hertfordshire, for their kind support of this study. KWL wassupported by the Deutsche Forschungsgemeineschaft.

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from

Lange, Sahakian, Quinn, Marsden

1 Albert M, Feldman RG, Willis AL. The "subcorticaldementia" of progressive supranuclear palsy. J NeurolNeurosurg Psychiatry 1974;37: 121-30.

2 Cummings JL, Benson DF. Subcortical dementia: neu-ropsychology, neuropsychiatry and pathophysiology.Arch Neurol 1984;41:874-9.

3 Alexander GE, DeLong MR, Strick PL. Parallel organiza-tion of functionally segregated circuits linking basalganglia and cortex. Annu Rev Neurosci 1986;9:357-81.

4 Bamford KA, Caine ED, Kido DK, Plassche WM,Shoulson I. Clinical-pathological correlation inHuntington's disease: a neuropsychological and com-puted tomography study. Neurology 1989;39:796-801.

5 Brandt JA, Bylsma FW. The dementia of Huntington'sdisease. In: Parks RW, Zec RF, Wilson RS, eds.Neuropsychology ofAlzheimer's disease and other dementias.New York: Oxford University Press, 1993:265-82.

6 Brown RG, Marsden CD. "Subcortical dementia": theneuropsychological evidence. Neuroscience 1988;25:363-387.

7 Hodges JR, Salmon DP, Butters N. Differential impair-ment of semantic and episodic memory in Alzheimer'sand Huntington's diseases: a controlled prospectivestudy. J Neurol Neurosurg Psychiatry 1990;53: 1089-95.

8 Litvan I, Mohr E, Williams J, Gomez C, Chase TN.Differential memory and executive functions indemented patients with Parkinson's and Alzheimer's dis-ease. J Neurol Neurosurg Psychiatry 199 1;54:25-9.

9 Downes JJ, Roberts AC, Sahakian BJ, Evenden JL, MorrisRG, Robbins TW. Impaired extra-dimensional shift per-formance in medicated and unmedicated Parkinson'sdisease: evidence for a specific attentional dysfunction.Neuropsychologia 1989;27: 1329-43.

10 Sahakian BJ, Downes JJ, Eagger S, et al. Sparing of atten-tional relative to mnemonic function in a subgroup ofpatients with dementia of the Alzheimer type.Neuropsychologia 1990;28: 1197-213.

11 Owen AM, Roberts AC, Polkey CE, Sahakian BJ, RobbinsTW. Extradimensional versus intradimensional set shift-ing performance following frontal lobe excisions, tempo-ral lobe excisions or amygdala-hippocampectomy inman. Neuropsychologia 1991;29:993-1006.

12 Owen AM, James M, Leigh PN, et al. Fronto-striatal cog-nitive deficits at different stages of Parkinson's disease.Brain 1992;115:1727-51.

13 Parks RW, Haxby JF, Grady CL. Positron emissiontomography in Alzheimer's disease. In: Parks RW, ZecRF, Wilson RS, eds. Neuropsychology of Alzheimer's dis-ease and other dementias. New York: Oxford UniversityPress, 1993:459-88.

14 Sahgal A, Lloyd S, Wray CJ, et al. Does visuospatial mem-ory in Alzheimer's disease depend on the severity of thedisorder? International Jrournal of Geriatric Psychiatry1992;7:427-36.

15 Owen AM, Downes JJ, Sahakian BJ, Polkey CE, RobbinsTW. Planning and spatial working memory followingfrontal lobe lesions in man. Neuropsychologia 1990;28:1021-34.

16 Robbins TW, James M, Lange KW, Owen AM, QuinnNP, Marsden CD. Cognitive performance in multiplesystem atrophy. Brain 1992;115:271-91.

17 Robbins TW, James M, Owen AM, et al. Cognitive deficitsin progressive supranuclear palsy, Parkinson's diseaseand multiple system atrophy in tests sensitive to frontallobe dysfunction. J Neurol Neurosurg Psychiatry 1994;57:79-88.

18 Sahakian BJ, Morris RG, Evenden JL, et al. A comparativestudy of visuospatial memory and learning in Alzheimertype dementia and Parkinson's disease. Brain 1988;111:695-718.

19 Owen AM, Sahakian BJ, Semple J, Polkey CE, RobbinsTW. Visuospatial short term recognition memory andlearning after temporal lobe excisions, frontal lobeexcisions or amygdala-hippocampectomy in man.Neuropsychologia 1995;33:1-24.

20 Folstein MF, Folstein SE, McHugh PR. Minimental state.J Psychiatr Res 1975;12: 189-98.

21 Kendrick DC. Kendrick cognitive tests for the elderly.Windsor: NFER Nelson, 1985.

22 Nelson HE. National adult reading test manual. Windsor:NFER Nelson, 1982.

23 Robbins TW, James M, Owen AM, Sahakian BJ, McInnesL, Rabbitt P. Cambridge neuropsychological test auto-mated battery (CANTAB): a factor analytic study of alarge sample of normal elderly volunteers. Dementia1994;5:266-81.

24 Philpot MP, Levy R. A memory clinic for the early diagno-sis of dementia of the Alzheimer type. InternationalJ7ournal of Geriatric Psychiatry 1987;2:195-200.

25 McKhann G, Drachman D, Folstein M, Katzman R, PriceD, Stadlin EM. Clinical diagnosis of Alzheimer's dis-ease: report of the NINCDS-ADRDA work group underthe auspices of the Department of Health and HumanServices Task Force on Alzheimer's disease. Neurology1985;34:939-44.

26 Hughes CP, Berg L, Danziger WL, Coben LA, MartinRL. A new clinical scale for the staging of dementia. BrJPsychiatry 1982;140:566-72.

27 Eagger SA, Levy R, Sahakian BJ. Tacrine in Alzheimer'sdisease. Lancet 1992;337:989-92.

28 David AS, Jeste DV, Folstein MF, Folstein SE. Voluntarymovement dysfunction in Huntington's disease and tar-dive dyskinesia. Acta Neurologica Scandinavica 1987;75:130-9.

29 Sahakian BJ, Owen AM. Computerised assessment inneuropsychiatry using CANTAB. 7 R Soc Med 1992;85:399-402.

30 Milner B. Interhemispheric differences in the localizationof psychological processes in man. Br Med Bull 1971;27:272-7.

31 Shallice T. Specific impairments in planning. In:Broadbent DE, Weiskrantz L, eds. The neuropsychology ofcognitive function. London: The Royal Society, 1982:199-209.

32 Roberts AC, Robbins TW, Everitt BJ. The effects ofintradimensional and extradimensional shifts on visualdiscrimination in humans and non-human primates. Q JExp Psychol 1988;40B:321-41.

33 Winer BJ. Statisticalpinciples in experimental design. 2nd ed.New York: McGraw-Hill, 1971.

34 Kullback S. Information theory and statistics. New York:Dover, 1968.

35 Robbins TW. A critique of the methods available for themeasurement of spontaneous locomotor activity. In:Iversen LL, Iversen SD, eds. Handbook ofpsychopharma-cology. Vol 7. New York: Plenum Press, 1977;37-82.

36 Starkstein SE, Brandt J, Bylsma F, Peyser C, Folstein M,Folstein SE. Neuropsychological correlates of brain atro-phy in Huntington's disease: a magnetic resonanceimaging study. Neuroradiology 1992;34:487-9.

37 Berent S, Giordani B, Lehtinen S, et al. Positron emissiontomographic scan investigations of Huntington's disease:cerebral metabolic correlates of cognitive function. AnnNeurol 1988;23:541-6.

38 Weinberger DR, Beran KE, Iadrola, Driesen N, Zec R.Prefrontal cortical blood flow and cognitive function inHuntington's disease. J Neurol Neurosurg Psychiatry1988;51:94-104.

39 Owen AM, Bekinska M, James M, et al. Visuospatialmemory deficits at different stages of Parkinson's dis-ease. Neuropsychologia 1993;31:627-44.

40 Moss M, Albert MS, Butters N, Payne M. Differential pat-terns of memory loss among patients with Alzheimer'sdisease, Huntington's disease and alcoholic Korsakoffsyndrome. Arch Neurol 1986;43:239-46.

41 Troster AI, Salmon DP, McCullogh D, Butters N. A com-parison of category fluency effects associated withAlzheimer's and Huntington's disease. Brain Lang 1989;37:500-13.

42 Bylsma FW, Rebok GW, Brandt J. Long-term retentionof implicit learning in Huntington's disease.Neuropsychologia 1991;29: 1213-21.

43 Beatty WW. Remote memory for visuospatial informationin patients with Huntington's disease. Psychobiology1980;17:431-4.

44 Beatty WW, Goodkin DE, Monson N, Beatty PA,Hertsgaard D. Retrograde amnesia in patients withAlzheimer's disease or Huntington's disease. NeurobiolAging 1988;9:181-6.

45 Brouwers P, Cox C, Martin A, Chase T, Fedio P.Differential perceptual-spatial impairment inHuntington's and Alzheimer's dementias. Arch Neurol1984;41: 1073-6.

46 Heindel WC, Salmon DP, Shults CW, Walicke PA,Butters N. Neuropsychological evidence for multipleimplicit memory systems: a comparison of Alzheimer's,Huntington's and Parkinson's disease patients.J Neurosci 1989;9:582-7.

47 Knopman D, Nissen MJ. Procedural learning is impairedin Huntington's disease; evidence from the serial reac-tion time task. Neuropsychologia 1991;29:245-54.

48 Butters N, Wolfe J, Granholm E, Martone M. An assess-ment of verbal recall, recognition and procedural mem-ory. Cortex 1986;22:11-32.

49 Hodges JR, Salmon DP, Butters N. The nature of thenaming deficit in Alzheimer's and Huntington's disease.Brain 1991;114:1547-58.

50 Rothlind JC, Brandt J. A brief assessment of frontal andsubcortical functions in dementia. JfNeuropsychiatry ClinNeurosc 1993;5:73-7.

51 Pillon B, Dubois B, Ploska A, Agid Y. Severity and speci-ficity of cognitive impairment in Alzheimer's,Huntington's and Parkinson's disease. Neurology 1986;36:1179-85.

52 Fedio P, Cox CS, Neophytides A, Canal-Frederick G,Chase TN. Neuropsychological profile of Huntington'sdisease patients and those at risk. In: Chase TN, WexlerNS, Barbeau A, eds. Advances in Neurology. Vol 23, NewYork: Raven Press, 1979;239-55.

53 Jones B, Mishkin M. Limbic lesions and the problem ofstimulus-reinforcement associations. Expt Neurol 1972;36:362-77.

54 Divac I, Rosvold HE, Szwarcbart MK. Behavioral effectsof selective ablation of the caudate nucleus. 3tournal ofComparative Physiology and Psychiatry 1967;63:184-90.

55 Roberts AC, De Salvia MA, Wilkinson LS, et al.6-hydroxydopamine lesions of the prefrontal cortex inmonkeys enhance performance on an analogue of theWisconsin Card Sorting test: Possible interactions withsubcortical dopamine. J Neurosc 1994;14:2531-44.

56 Mishkin M, Malamut B, Bachavalier J. Memories andhabits: Two neural systems. In: Lynch G, McGaugh JL,Weinberger NM, eds. Neurobiology of learning and mem-ory. New York: Guildford Press, 1984:65-77.

57 Robbins TW, Sahakian BJ. The neuropsychological signif-icance of the behavioural effects of psychomotor stimu-lant drugs. In: Creese I, ed. Stimulants. New York:Raven Press, 1983:301-38.

606

on Septem

ber 13, 2020 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.58.5.598 on 1 May 1995. D

ownloaded from