Journal ofNeurology, Neurosurgery, and Psychiatry 1994;57: 1389-1394

Initial letter and semantic category fluency inAlzheimer's disease, Huntington's disease, andprogressive supranuclear palsy

Anne Rosser, John R Hodges

AbstractTen patients with dementia ofAlzheimer's type, 10 patients with pro-gressive supranuclear palsy, and 10patients with Huntington's disease were

compared on two types of verbal fluencytask-namely, initial letter fluency andcategory (semantic) fluency. The groupswere carefully matched for overall levelofdementia on the dementia rating scale,and were compared with 25 age matchednormal controls. The controls found let-ter fluency more difficult than categoryfluency, and this relative pattern of per-formance was repeated in the progressivesupranuclear palsy and Huntington'sdisease groups, although both groupswere significantly impaired on bothtasks. By contrast, patients withAlzheimer's disease performed just as

poorly as the progressive supranuclearpalsy and Huntington's disease groups on

the category tasks, but were significantlyless impaired at letter fluency, perform-ing at near normal levels on thistask. From these results, it is suggestedthat the performances of patients withprogressive supranuclear palsy andHuntington's disease relate largely toinitiation and retrieval problems sec-

ondary to disruption of frontostriatal cir-cuits, whereas in Alzheimer's disease, thepoorer performance on category fluencyis due principally to the breakdown ofsemantic knowledge, which probablyreflects temporal neocortical involve-ment.

(J Neurol Neurosurg Psychiatry 1994;57:1389-1394)

University NeurologyUnit, Addenbrooke'sHospital, Cambridge,UKA RosserJ R HodgesCorrespondence to:Dr J R Hodges, Departmentof Neurology,Addenbrooke's Hospital,Hills Road, Cambridge CB22QQ, UK.Received 29 March 1994and in revised form5 July 1994.Accepted 8 July 1994

Alzheimer's disease, which affects predomi-nantly medial temporal and posterior corticalregions, is associated with a different patternof cognitive impairment from dementias asso-

ciated with damage to subcortical structuressuch as progressive supranuclear palsy,Huntington's disease, multiple sclerosisrelated dementia, and AIDS dementia com-plex. 14 Patients with Alzheimer's diseasecharacteristically present with deficits ofmemory, aphasia, or visuospatial impairment,whereas subcortical dementias producepatients who are slow, apathetic, forgetful,and have difficulty manipulating new infor-mation.' Recent studies of patients withAlzheimer's disease using theoretically moti-vated neuropsychological tests have shownevidence of episodic memory impairment that

is secondary to poor encoding and storage ofnew material, as well as an accelerated rate offorgetting.2 With disease progression there isalso a breakdown of structure and organisa-tion of semantic memory.6-8 By contrast,memory is less impaired in subcorticaldementias, such as Huntington's disease, andthe main difficulty seems to be the operationof effective retrieval strategies to search forinformation from memory stores.2389Alzheimer's disease and subcortical types ofdementia (Huntington's disease and progres-sive supranuclear palsy) have been shown todiffer on the memory and initiation subtests ofthe dementia rating scale, when matched foroverall level of dementia on that scale, withpatients with Alzheimer's disease performingpoorly on tests of memory, and patients withHuntington's disease and progressivesupranuclear palsy performing poorly on testsof initiation."l 12

Disorders of language have also long beenrecognised as part of the symptomatology ofmany progressive dementias, and there aretheoretical reasons to believe that differentdementias may differ on tests of language pro-duction, such as letter and category fluency.Letter fluency requires subjects to generatewords beginning with certain letters, whereasin category fluency tests patients are asked togenerate exemplars from a given category (forexample, animals, vegetables, householdobjects). Tests are usually conducted in thesetting of a time constraint, for example, oneminute per letter or category. Performance onboth tasks depends upon frontostriatal circuitsthat control aspects of executive function(including attention, initiation, and retrievalprocesses) and working memory. Also, thetwo types of task depend on the integrity ofthe stores from which the examples areretrieved. In the case of initial letter fluency,the phonologically based word store is clearlycritical, whereas category fluency depends onthe intactness of semantic memory.8 1314 Wewould argue, therefore, that frontostriataldeficits, of the type found in patients withsubcortical dementias, should produceequally severe impairment on letter and cate-gory fluency tasks. By contrast, disorders thatcause breakdown in the organisation ofsemantic memory, such as Alzheimer's dis-ease, should result in more pronouncedimpairment on category fluency.

In the case of Alzheimer's disease, there hasbeen some controversy concerning its relativeimpact on letter and category fluency. Whereassome workers have found disproportionate

1389

on Decem

ber 26, 2019 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.11.1389 on 1 Novem

ber 1994. Dow

nloaded from

Rosser, Hodges

impairment on category fluency tests," 10 13"5others have found equal impairment on bothtypes of fluency test.1617 Recent studies thathave compared groups of patients withAlzheimer's disease and Huntington's disease,matched for overall level of disease severity,have, however, lent support to the hypothesisthat cortical and subcortical dementia causedifferential effects on letter and category flu-ency."8 Before accepting the generality of thishypothesis it is important to compare patientswith a wider range of diseases. Although pro-gressive supranuclear palsy is an obviouscandidate as the pathological changes arerestricted, almost entirely, to subcorticalstructures,'9 20 few studies have comparedpatients with progressive supranuclear palsyand other diseases, and none have includedboth letter and category fluency tests. Milbergand Albert2l reported a double dissociationbetween naming (with the Boston namingtest) and letter fluency in groups of patientswith Alzheimer's disease and progressivesupranuclear palsy, but did not investigatecategory fluency in their patients. Pillon et al22compared the performance of groups ofpatients with Alzheimer's disease, progressivesupranuclear palsy, Huntington's disease, andParkinson's disease on a wide ranging batteryof tests including two measures of verbal flu-ency (animals and words beginning with M);the progressive supranuclear palsy group weremore impaired than the Alzheimer's diseaseand Parkinson's disease groups but becausethe results of the two fluency tests were com-bined, it is not possible to comment on rela-tive levels of performance on the two tasksacross the groups.

In the study reported here, patients withAlzheimer's disease, Huntington's disease,and progressive supranuclear palsy were care-fully matched for overall level of dementia onthe dementia rating scale, and were alsomatched for age and educational level. Allgroups were given letter and category fluencytests and were compared with normal agematched controls. We predicted thatHuntington's disease and progressivesupranuclear palsy would produce equallysevere impairment in letter and category flu-ency, and that Alzheimer's disease would pro-duce disproportionately more impairment incategory fluency.

Materials and methodsSUBJECTSThe study involved 55 subjects in total. Tenpatients had a diagnosis of progressivesupranuclear palsy according to criteriadescribed by Lees,2' 10 patients hadHuntington's disease as defined by chorea,intellectual decline and a positive family his-tory, and 10 patients had a diagnosis ofproba-ble Alzheimer's disease according to theNational Institute of Neurological Disorderand Stroke and the Alzheimer Disease andRelated Disease Association (NINCDS-ADRDA).'4 The patients with Alzheimer'sdisease were selected from a larger cohort of

about 50 patients undergoing longitudinalneuropsychological assessment to match theprogressive supranuclear palsy and Hunting-ton's disease groups in terms of age, educa-tion, and score on the Dementia Rating Scale.

Examination and laboratory tests for thedifferential diagnosis of dementia were carriedout on all patients to exclude other causes ofdementia. All patients had brain MRI or CT.Twenty five community dwelling normal

volunteers from the MRC Applied PsychologyUnit's subject panel formed the normal con-trol group. These subjects were selected toinclude the same age and educational range asthe patient groups.

NEUROPSYCHOLOGICAL TESTSEach subject was assessed on the DementiaRating Scale with the standard methoddescribed by Mattis.'5 Two commonly usedverbal fluency tasks-letter and category flu-ency-were presented to all subjects in thefour groups. Subjects were asked to generateas many words as possible in one minute. Forletter fluency, three trials were performed withthe letters F, A, and S and subjects wereinstructed to exclude proper nouns and thesame word with different suffixes (fix, fixed,fixing etc.). For category fluency tests, sub-jects were given one minute for each of threeliving categories (animals, birds, and watercreatures) and three man made categories(household objects, vehicles, and musicalinstruments).8The principal score used for intergroup

analysis was the total number of correctresponses. For letter fluency, the summedcorrect responses of the three letters was used.For category fluency, the summed correctresponses of the three living and three manmade categories were analysed separately.

Errors were classified as perseverations(repetition of an item from earlier in the list orthe same word stem with a different suffix),intrusions (the inclusion of an item fromanother category or, in the case of letter flu-ency, beginning with the wrong letter) andothers (bizarre and inappropriate responsesthat could not be classified as one of the othererror types).

ANALYSESOne way analysis of variance (ANOVA) wasused to analyse differences between groups.Where main group effects were found, posthoc comparisons were made by t tests withNewman-Keuls correction.

DEMOGRAPHIC DATATable 1 gives the demographic data. Therewas no significant difference between thegroups in terms of age (F(3,51) = 2-8,p > 0 05) or educational level (F (3,5 1) =0-64, p > 0 05). The patient groups were alsowell matched on the Dementia Rating Scale,although as expected, the normal controlsscored higher; one wayANOVA showed a sig-nificant main effect for groups (F(3,5 1) =24-45, p < 0-0001) and post hoc analysisshowed significant (p < 0 05) differences

1390

on Decem

ber 26, 2019 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.11.1389 on 1 Novem

ber 1994. Dow

nloaded from

Initial letter and semantic categoryfluency in Alzheimer's disease, Huntington's disease, and progressive supranuclear palsy

Table 1 Age, education, and dementia rating score total scores for the controls andpatients with dementia ofAlzheimer's type (DA T), progressive supranuclear palsy (PSP),and Huntington's disease (HD)

Controls DAT PSP HD(n = 25) (n =10) (n =10) (n =10)

Age (y) (mean (SD)) 69 (8 4) 67-2 (9 9) 65-8 (8 6) 59 9 (8 0)Range 53-85 51-80 52-82 43-72Education (y) (mean (SD)) 10-7 (2-3) 11-1 (2 5) 10-3 (2 6) 9-7 (1-5)DRS (mean (SD)) 139-9 (3-6) 121-4 (8 3) 122-2 (14-4) 121-5 (6 2)Range 130-144 107-132 90-138 112-133

Table 2 Error data for letterfluency and living categoryfluency tests (figures weresimilarfor non-living categories) expressed as a percentage of total exemplars generatedforthe control subjects and patients with dementia ofAlzheimer's type (DA T), progressivesupranuclear palsy (PSP), and Huntington's disease (HD).

Category fluency (iving)Letterfluencyperseveration Intrusions Perseverations Intrusions Other(/J) (%/) (%) (/%) No)

Control 1-2 07 1-4 07 1.0DAT 3-1 1.1 7-7 2-2 -PSP 7-0 3-2 6-1 2-6 0-2HD 79 39 2-8 1-4 2-1

between the patient groups and controlgroups, but no differences betweenAlzheimer's disease, Huntington's disease,and progressive supranuclear palsy groups.

ResultsThe figure shows results of verbal and cate-gory fluency tests in Alzheimer's disease,Huntington's disease, progressive supranu-clear palsy, and normal control groups.Normal elderly subjects were able to recall amean of 44-5 (SD 9-9) exemplars for the threeletters F, A, S combined, compared with 34-8(14.0) for Alzheimer's disease, 20-7 (12.9) forprogressive supranuclear palsy, and 17-8(11 8) for Huntington's disease groups. A oneway ANOVA showed a highly significant

Performance of the controlsubjects and patients withdementia ofAlzheimer'stype (DAT), progressivesupranuclear palsy (PSP)and Huntington's disease(HD) on letter (F,A,S)and semantic category(three living and threeman made categoriesshown separately) fluencytests showing the meannumber of items correctforeach group (SEM).

607

50

+J0

a)

v

o0

z

crCD

| FAS|ECI LivingIE Man made

40

30

20

10

0

Controls DAT

overall group effect (F= 17.55 (3,51) p <0-0001). Post hoc pairwise analyses showedthat the normal controls produced signifi-cantly higher scores than each of the patientgroups. In addition, patients with Alzheimer'sdisease scored significantly higher thanpatients with Huntington's disease or pro-gressive supranuclear palsy, but those withHuntington's disease or progressive supranu-clear palsy were not significantly differentfrom each other. In terms of the proportionalreduction relative to the controls' mean score,the Alzheimer's disease group showed a 22%reduction whereas the progressive supra-nuclear palsy and Huntington's diseasegroups showed 54% and 60% reductions,respectively.

For category fluency, the scores for totalcorrect living and man made items wereanalysed separately; however, the differencesbetween the groups were identical for bothparts of the test. As shown in previous studies,normal controls scored more highly on thecategory than the letter fluency tests. Again,there was a highly significant group effect forboth living (F = 25-5 (3,51), p < 0-0001) andman made (F= 32.71 (3,51), p < 0*0001)categories. Post hoc analyses showed signifi-cant differences between normal controls(mean scores 57-4 (12-7) and 55-1 (8'5) forliving and man made, respectively) and allthree patient groups (Alzheimer's disease 31 - 1(6.1) and 32-5 (9-6) for living and man made,progressive supranuclear palsy 31d1 (14-2)and 28-4 (12-9) for living and man made, andHuntington's disease 28-8 (8 5) and 27-7(9'6) for living and man made). By contrastwith the letter fluency tests, however, therewas no significant difference betweenAlzheimer's disease, Huntington's disease,and progressive supranuclear palsy groups onperformance in either living or man madecategory fluency tests. In terms of the propor-tional reduction, relative to the mean score ofthe control group, the Alzheimer's disease,progressive supranuclear palsy, andHuntington's disease groups showed reduc-tions of 46%, 46%, and 50% respectively forthe living categories, with virtually identicalfigures for the man made categories.

Table 2 shows the error rates, expressed asthe percentage of the number of exemplarsproduced for each group. For all tasks, thepatient groups produced more errors thancontrols. On letter fluency, the rate of perse-verative errors was almost twice as high forthe Huntington's disease and progressivesupranuclear palsy groups compared with theAlzheimer's disease group. This differencealmost produced a significant group effect(p < 0.065). For category fluency, there wasan overall group effect for perseverative errors(F = 3-88 (3,51), p < 0-014) and post hoccomparisons showed that patients withAlzheimer's disease produced significantlymore perseverative errors than controls orpatients with Huntington's disease. There wasno significant difference between patientswith Huntington's disease and controls.Patients with progressive supranuclear palsy

1391

on Decem

ber 26, 2019 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.11.1389 on 1 Novem

ber 1994. Dow

nloaded from

Rosser, Hodges

produced more errors than controls, but didnot differ significantly from controls orpatients with Alzheimer's disease. The pro-portions of intrusion errors were not signifi-cantly different between groups for any of thefluency tasks.

DiscussionIn this study, two different types of verbal flu-ency task (letter and category) were given topatients with Alzheimer's disease, Hunting-ton's disease, and progressive supranuclearpalsy, all of whom had mild levels of demen-tia, and the scores were compared with thoseof normal elderly controls. Patients were wellmatched for overall level of dementia andother demographic details. Normal controlsperformed less well on letter fluency than oncategory fluency, a finding consistent withother studies.' 15 18 It is interesting to speculateon the reasons for this difference. We assumethat identical executive or supervisoryprocesses are involved in the initiation andmonitoring of both tasks, and that the differ-ence reflects the nature of semantic, asopposed to phonological, representations, orthe specificity of the retrieval cues involved inthe two tasks. The representation of semanticknowledge is clearly a fundamental aspect ofhuman cognition that has been shown to pre-cede linguistic competence in infants.26Furthermore, there is a considerable body ofevidence that knowledge is organised by cate-gory (for example, living v man made items;land animals).27 When performing any cate-gory fluency tasks it is essential to accesssemantic stores: activation of an initial, andusually highly prototypical exemplar (forexample, cat or dog for the category "ani-mals"), leads to automatic activation ofclosely related semantic neighbours. If auto-matic activation fails, subjects can also usemore active search strategies. By contrast, let-ter fluency must be performed at the phono-logical level of word representation withoutreference to meaning, and the spread of acti-vation within the phonological lexicon mayproceed less rapidly than at the semantic level.In addition, the cue "animals" or "musicalinstruments" addresses a very specific subsetof knowledge, whereas the cue "words begin-ning with A" is relatively underspecified as itapplies to a significant proportion of allknown words.

Whatever the reasons for this difference,the relation was maintained in theHuntington's disease and progressivesupranuclear palsy groups, who wereimpaired on both tasks. The number of cor-rect exemplars produced by the patients withHuntington's disease and progressivesupranuclear palsy was remarkably similar,and there was no significant differencebetween these two groups on either of the flu-ency tests. For Alzheimer's disease, however,the pattern was different. Patients withAlzheimer's disease performed significantlybetter on letter fluency than patients witheither progressive supranuclear palsy or

Huntington's disease, although they were sig-nificantly impaired compared with controls.By contrast, patients with Alzheimer's diseasewere relatively more impaired on tests of cate-gory than on letter fluency (46% v 22%reduction compared with controls). This isreflected in the fact that intergroup analysesshowed no difference between Alzheimer'sdisease, Huntington's disease, and progressivesupranuclear palsy on category fluency tests.Thus the patients with Alzheimer's diseasedisplay a reversal of the normal pattern of per-formance. This finding is in agreement withsome recent studies showing that patientswith Alzheimer's disease are more impairedon category than letter fluency tests.' 1314Our interpretation of these findings is that

Alzheimer's disease causes impairment ofsemantic memory early in the course of thedisease so that category fluency, which seemsto be heavily dependent on intact semanticmemory, is affected more severely than letterfluency. There is now overwhelming evidencethat patients with Alzheimer's disease showimpairment on a range of tests dependent onsemantic memory including category fluency,picture naming, word-picture, and picture-picture matching, generation of word defini-tions, and identification offamous faces.68 132829Furthermore, recent work has established thatthis impairment is almost certainly due tobreakdown of semantic knowledge, ratherthan failure to access memory stores.68Semantic memory has not been as extensivelyinvestigated in patients with subcorticaldementias but the evidence to date suggeststhat it does not break down to the same extentin Huntington's disease.3 7 30 For instance, alongitudinal study by Hodges et al3 showedthat semantic memory degraded more quicklyin Alzheimer's disease than in Huntington'sdisease over the course of a year, whereas initi-ation and recall deteriorated more quickly inpatients with Huntington's disease. A morerecent comparison of patients with progres-sive supranuclear palsy and Alzheimer's dis-ease failed to show differences between thetwo groups on tests of semantic memory,although the Alzheimer's disease group wassignificantly more impaired on tests ofepisodic memory." In this study the sugges-tion was made that the impairment in the pro-gressive supranuclear palsy group may reflecta failure of access to semantic representations.

Patients with Huntington's disease andprogressive supranuclear palsy performed tothe same level as patients with Alzheimer'sdisease on category fluency tasks, but weremore impaired on letter fluency. In otherwords, the "normal" pattern of relative per-formance on the two tasks was preserved,although patients were impaired in both. It islikely that the poor performances of patientswith Huntington's disease and progressivesupranuclear palsy reflects impaired initiationand retrieval strategies which, as discussed inthe introduction, play a key part in both flu-ency tasks. This result is in accordance withother studies of patients with Huntington'sdisease.3 17 Randolph et al32 investigated

1 392

on Decem

ber 26, 2019 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.11.1389 on 1 Novem

ber 1994. Dow

nloaded from

Initial letter and semantic category fluency in Alzheimer's disease, Huntington's disease, and progressive supranuclear palsy

performance on a category fluency task in"uncued" and "cued" conditions in patientswith Alzheimer's disease, Huntington's dis-ease, and Parkinson's disease. They foundthat patients with Alzheimer's disease per-formed equally badly in both conditions,whereas patients with Huntington's disease orParkinson's disease performed relatively bet-ter with cueing. On the basis of this finding,they concluded that patients with Alzheimer'sdisease were impaired due to degradation ofsemantic memory, whereas patients withHuntington's disease and Parkinson's diseasefail due to impaired retrieval mechanismsrather than degradation of knowledge whichexplains their improved performance withcueing.The differences in performance on letter v

category fluency are likely to relate to the dif-ferences in pathology in the cortical dementiaof Alzheimer's disease compared with the pre-dominantly subcortical pathology of progres-sive supranuclear palsy and Huntington'sdisease. The neural substrate of semanticmemory remains unsettled but current evi-dence implicates the temporal neocortex asthe most important region.3334 Patients withthe syndrome of semantic dementia, in whichthere is progressive yet selective loss of seman-tic memory with relative preservation of otherlinguistic (phonological and syntactic) andnon-verbal cognitive (for example, complexperceptual and visuospatial) abilities, all havestructural or functional changes in the tempo-ral lobe(s)35; on MRI, the areas most involvedseemed to be the temporal neocortex, with anemphasis on the middle and inferior temporalgyri. Similarly, patients with relatively selec-tive loss of semantic memory after herpes sim-plex virus encephalitis typically showdestruction of the temporal neocortex.'637 Thebreakdown of the semantic memory is likely,therefore, to reflect temporal neocorticaldamage.

Impairment of initiation and retrievalstrategies in progressive supranuclear palsyand Huntington's disease is likely to relate todamage in subcortical frontostriatal circuits.There is pathological, dynamic scanning, andneuropsychological evidence available to sup-port this assumption. Although corticalpathology is known to occur in progressivesupranuclear palsy and Huntington's diseasethe bulk of the neuropathological changesinvolve subcortical structures, at least at theearly stage of the disease.'92038 Positron emis-sion tomography shows a frontal pattern ofhypoperfusion in progressive supranuclearpalsy, which is in keeping with the suggestionthat subcortical pathology results in func-tional frontal deactivation.3940 Moreover,there is now clear evidence that both progres-sive supranuclear palsy and Huntington's dis-ease produce a profile of neuropsychologicaldeficits that parallels that found in patientswith frontal cortical damage.211122122 Forinstance, Robbins, Sahakian, and coworkershave shown that patients with progressivesupranuclear palsy and Huntington's diseaseshow pronounced impairment on their com-

puterised battery of attentional and executivetasks (CANTAB), whereas patients with earlyAlzheimer's disease perform relatively nor-mally on some of the same tasks from theCANTAB battery.4'A3 Other frontal lobe syn-dromes have been shown to impair verbal flu-ency in a similar way: a non-aphasic patientwith trauma induced bilateral frontal lobelesions, described by Randolph et al,29 wasimpaired on verbal fluency tests but improvedto normal after cueing, suggesting that theprefrontal cortex may be critical in maximis-ing the search strategies. Similar results wereobtained in a detailed single case study of apatient with frontal Pick's disease who exhib-ited a severe reduction in letter and categoryfluency in the absence of any evidence ofsemantic memory impairment on other mea-sures.44 As Pick's disease is a cortical disorder,but one which produces a "subcortical" pro-file of behavioural and neuropsychologicaldeficits, it would perhaps be better to adoptthe term "frontostriatal dementias" to encom-pass patients with both subcortical and pro-gressive frontal neurodegenerative diseases.

In summary, we suggest that the differen-tial performance of patients with Alzheimer'sdisease compared with those with Hunting-ton's disease and progressive supranuclearpalsy, on tests of verbal fluency is due tobreakdown of semantic knowledge structurein Alzheimer's disease and to the disruption ofsubcortical-frontal circuits in progressivesupranuclear palsy and Huntington's disease,and that these differences relate to the differ-ences in underlying pathology.

This research is supported in part by an MRC project grant toJRH. We thank Naida Graham for her help with the statisticalanalysis.

1 Cummings JL. Subcortical dementia. Oxford: OxfordUniversity Press, 1990.

2 Heindel WC, Salmon DP, Butters N. Cognitiveapproaches to the memory disorders of dementedpatients. In: Sutker PB, Adams HE, eds. Comprehensivetextbook of psychopathology. 2nd ed. New York: PlenumPress, 1993.

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

4 Moss MB, Albert MS, Butters N, Payne M. Differentialpatterns of memory loss among patients withAlzheimer's disease, Huntington's disease and alcoholicKorsakoff s syndrome. Arch Neurol 1986;43:239-46.

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

6 Chertkow H, Bub D. Semantic memory loss in dementiaof Alzheimer type. What do various measures measure?Brain 1990;113:397-417.

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

8 Hodges JR, Salmon DP, Butters N. Semantic memoryimpairment in Alzheimer's disease: failure of access ordegraded knowledge? Neuropsychologia 1992;30:301-14.

9 Butters N, Wolfe J, Martone M, Granholm E, Cermak LS.Memory disorders associated with Huntington's disease:verbal recall, verbal recognition and procedural memory.Neuropsychologia 1985;23:729-43.

10 Butters N, Wolfe J, Granholm E, Cermak LS. An assess-ment of verbal recall, recognition and fluency abilitiesin patients with Huntington's disease. Cortex 1986;22:11-32.

11 Rosser AE, Hodges JR. The dementia rating scale inAlzheimer's disease, Huntington's disease and progres-sive supranuclear palsy. J Neurol 1994 (in press).

12 Salmon DP, Kwo-on-Yuen PF, Heindel WC, Butters N,Thal U. Differentiation of Alzheimer's disease andHuntington's disease with the Dementia Rating Scale.Arch Neurol 1989;46:1204-6.

13 Martin A, Fedio P. Word production and comprehension inAlzheimer's disease: the breakdown of semantic knowl-edge. Brain Lang 1983;19:124-41.

1 393

on Decem

ber 26, 2019 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.11.1389 on 1 Novem

ber 1994. Dow

nloaded from

Rosser, Hodges

14 Butters N, Granhohm E, Salmon DP, Grant I, Wolfe J.Episodic and semantic memory: a comparison ofamnesic and demented patients. Jf Clin Exp Neuropsychol1987;9:479-97.

15 Monsch AU, Bondi MW, Butters N, Salmon DP,Katzman R, Thal U. Comparisons of verbal fluencytasks in the detection of dementia of the Alzheimer type.Arch Neurol 1992;49:1253-8.

16 Rosen W. Verbal fluency in aging and dementia. J ClinNeuropsychol 1980;2: 135-46.

17 Ober BA, Koss E, Friedland RP, Delis DC. Processes ofverbal memory failure in Alzheimer'-type dementia.Brain Cogn 1985;4:90-103.

18 Monsch AU, Bondi MW, Butters N, Paulsen JS, SalmonDP, Brugger P, Swenson MR. A comparison of categoryand letter fluency in Alzheimer's disease andHuntington's disease. Neuropsychol 1994;8:25-30.

19 Jellinger KA, Bancher C. Neuropathology. In: Litvan I,Agid Y, eds. Progressive supranuclear palsy: clinical andresearch approaches. New York: Oxford University Press,1992:44-89.

20 Steele JC, Richardson JC, Olszewski J. ProgressiveSupranuclear Palsy: A heterogeneous degenerationinvolving the brain stem, basal ganglia, and cerebellum,with vertical gaze and pseudobulbar palsy, nuclear dys-tonia, and dementia. Arch Neurol 1964;10:333-59.

21 Milberg W, Albert M. Cognitive differences betweenpatients with progressive supranuclear palsy andAlzheimer's disease. J7 Clin Exp Neuropsychol 1989;11:605-14.

22 Pillon B, Dubois B, Ploska A, Agid Y. Severity and speci-ficity of cognitive impairment in Alzheimer's,Huntington's, and Parkinson's diseases, and ProgressiveSupranuclear Palsy. Neurology 1991;41:634-43.

23 Lees AJ. The Steele-Richardson-Olszewski syndrome. InMarsden CD, Fahn S, eds. Movement disorders 2.London: Butterworths, 1987:272-87.

24 McKhann G, Drachman D, Folstein M, et al. Clinicaldiagnosis of Alzheimer's disease: report of theNINCDS-ADRDA Work Group under the auspices ofDepartment of Health and Human Services Task Forceon Alzheimers Disease. Neurology 1984;34:939-44.

25 Mattis S. Dementia rating scale. Windsor: NFER-Nelson,1992.

26 Mar, 1ler JM, Bauer PJ, McDonough J. Separating theshet;p from the goats: Differentiating global categories.Cogn Psychol 1991;23:263-98.

27 Shalice T. From neuropsychology to mental structure.Cambridge: Cambridge University Press, 1988.

28 Bayles KA, Tomoeda CK. Confrontational namingimpairment in dementia. Brain Lang 1983;19:98-114.

29 Hodges JR, Patterson KP. Is semantic memory consis-tently impaired early in the course of Alzheimer's dis-ease?: neuroanatomical and diagnostic implications.Brain 1994 (in press).

30 Smith S, Butters N, White R, Lyon L, Granholm E.Priming semantic relations in patients with Huntington'sdisease. Brain Lang 1988;33:27-40.

31 van der Hurk P, Hodges JR. Episodic and semantic memoryin Alzheimer's disease and progressive supranuclearpalsy: a comparative study. J Clin Exp Neuropsychol 1994(in press).

32 Randolph C, Braun AR, Goldberg TE, Chase TN.Semantic fluency in Alzheimer's, Parkinson's, andHuntington's disease: Dissociation of storage andretrieval failures. Neuropsychology 1993;7:82-88.

33 Damasio AR, Damasio H, Tranel D, Brandt JP. Neuralregionalization of knowledge access: preliminary evi-dence. Symposia on quantitative biology. Cold SpringHabour Laboratory Press 1990;55: 1039-47.

34 Patterson KP, Hodges JR. Disorders of semantic memory.In: Baddeley A, Wilson B, Watts F, eds. Handbook ofmemory disorders. Chichester: Wiley 1994 (in press).

35 Hodges JR, Patterson K, Oxbury S, Funnell E. Semanticdementia: Progressive fluent aphasia with temporal lobeatrophy. Brain 1992;115:1783-806.

36 Pietrini V, Nertempi P, Vaglia A, Revello MG, Pinna V,Ferro-Milone F. Recovery from herpes simplexencephalitis: selective impairment of specific semanticcategories with neuroradiological correlation. J NeurolNeurosurg Psychiatry 1988;51: 1284-93.

37 Warrington EK, Shailice T. Category specific semanticimpairments. Brain 1984;107:829-54.

38 Roos RAC. Neuropathology of Huntington's disease. In:Vinken GW, Bruyn GW, Klawans HL, eds. Handbook ofclinical neurology. Vol 49. Amsterdam: Elsevier, 1986.

39 Parks RW, Haxby JV, 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.

40 Goffinet AM, De Volder AG, Gillian C, et al. Positronemission tomography demonstrates frontal like hypome-tabolism in progressive supranuclear palsy. Ann Neurol1989;25:131-9.

41 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's type.Neuropsychologia 1990;28:1 197-213.

42 Sahakian BJ, Owen AM. Computerized assessment inneuropsychiatry using CANTAB. J R Soc Med 1992;85:399-402.

43 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.

44 Hodges JR, Gurd J. Remote memory and lexical retrievalin a case of frontal Pick's disease. Arch Neurol 1994;51:821-7.

1394

on Decem

ber 26, 2019 by guest. Protected by copyright.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.57.11.1389 on 1 Novem

ber 1994. Dow

nloaded from