neuropsychology of schizophrenia

8/2/2019 Neuropsychology of Schizophrenia

1/9

Neuropsychology of schizophreniaW h a t are the implications of intellectual an dexp eriential abn ormalit ies for the neurobiology ofschizophrenia?Chris FrithWellcome Department of Cognitive Neurology, Institute of Neurology, London, UK

The diagnosis of schizophrenia is largely bas ed on reports of bizarre experiencessuch as having alien thoughts inserted into one's mind. Many patients with thisdiagnosis show a marked intellectual decline and p articular problems with tasksinvolving certain kinds of memory or requiring mental flexibility. Similar patterns ofperform ance can be seen in patients with damage in the prefrontal cortex. However,patients with schizophrenia show a very v aried pattern of impairments relating totheir current mental state. Chronic patients with negative features, such as poverty ofspeech, are most likely to show poor test performa nce, while the presence of severehallucinations an d delusions need not be associated with any imp airment. Acognitive ap pro ach suggests that hallucinations an d delusions result from the patientattributing his own actions to an externa l agency. This error is due to an inab ility todistinguish between e xtern al events and percep tual changes caused by his ow nactions.The basis of this failure could be a functional disconnection between frontalbrain area s concerned with action and posterior areas concerned with perception.

Postal address:Chris frith,

Wellcome Department ofCognitive Neurology,

Institute of Neurology,72 Queen Square, London

V/ON3BCUK

In the continued absence of a biological marker for schizophrenia, thediagnosis is made on the basis of signs and symptoms and remains, to acertain degree, arbitrary. Currently, the most widely used definition is tobe found in the diagnostic manual of the American PsychiatricAssociation (DSM-IV). In cases diagnosed by this scheme, most patientswill have reported bizarre delusions (e.g. thought broadcasting; thoughtsleave the patient's mind and enter the minds of others) and/or prominenthallucinations of a voice (e.g. voices discussing the patient's actions). Inaddition, these experiences will have been prolonged rather thantransient. The symptoms must have been present for at least a weekand there must have been signs of a more general disturbance for at least6 m onths. The majority of patients receiving a diagnosis of schizophrenianever fully recover1. The typical 'chronic' patient will manifest abnormalbehaviour best characterised as lacking volition; poverty of speech,poverty of movement and poverty of ideas. Such patients may not reporthallucinations or delusions, but the vast majority will have experiencedThe Britith Council 1996 British Madical Bulletin 1996;52 (No. 2):618-626


2/9

Neuropsychology of schizophrenia

these at some time in the course of their illness. Confronted with suchcases, the challenge for the neuropsychologist is to provide data fromobservations and experiments which give clues to the nature of the braindysfunctions which underlie the signs and symptoms associated withschizophrenia. There are two fundamental approaches. The descriptiveapproach asks the question, is there a characteristic pattern of intellectualimpairment associated with the diagnosis of schizophrenia? Themechanistic approach asks the question, what psychological processesgive rise to particular signs and symptoms associated with schizophreniaand how do these processes relate to normal brain functioning?

The pat tern of intellectual im pairme ntsGenera/ decline

When Kraepelin used the term Dementia Praecox to label what we nowcall schizophrenia, he was emphasising his observation that such patientsoften show a dramatic and irreversible decline in intellectual functionshortly after the onset of their illness in early adulthood. A number ofrecent studies confirm this observation. Johnstone and her colleagues1attempted to trace all the people in the Harrow Health District who hadbeen discharged with a diagnosis of schizophrenia during a 10 yearperiod . Intellectual functioning was assessed in 283 cases. An estimate ofpremorbid functioning showed no difference from controls, but currentIQ was, on average, 16 points lower2. A similar result was obtained byNelson and her colleagues3 in a group of chronic patients. Theseobservations confirm the reports of many earlier studies4 showing thatintellectual functioning is impaired in schizophrenic patients. While theremay be some patients who were always impaired5, the evidence for adecline in function is striking. It is not clear how rapidly this declineoccurs. It seems to be complete within 5 years of onset of the illness, butmay well start before the first psychotic symptoms appear.

Specific impairmentsIf the brain abnormalities associated with schizophrenia are restricted tocertain areas or systems, then we would expect to find that intellectualimpairment was more marked in some domains than others. There havebeen many studies in which batteries of neuropsychological tests havebeen used to assess schizophrenic patients in the hope of demonstrating acharacteristic pattern of impaired and spared functions. A major problem

BnW.AW.co/Bu/Un 1996^2 (No. 3) 619


3/9

Biological psychiatry

with these (and in indeed all) studies of schizophrenia arises from the factthat the vast majority of patients are on large and chronic amounts ofmedication. Surprisingly, almost nothing is known about the effects ofthese drugs upon test performance. Saykin and his colleagues 6 overcamethis problem by studying chronic patients who were drug free at the timeof testing. Even in this drug free population, the average performance ofthe group was at least one standard deviation below the normal level inall the ten domains that were examined confirming the existence of ageneral intellectual decline. However, in three domains, all concernedwith learning and memory, performance was substantially worse.Although these results are clear cut in terms of test performance, theimplications for underlying brain dysfunction are far less certain. Theauthors conclude that the specific impairments in memory and learningare consistent with damage in medial temporal cortex. However, thereare m any different varieties of learning and mem ory, each associated withdifferent brain systems7. As a measure of verbal memory, Saykin andcolleagues used story recall. The same task was used by Shallice and hiscolleagues8 in an intensive study of a small group of patients. Thesepatients also performed very badly on story recall, but many of the samepatients performed within the normal range on a test of forced choicerecognition of wo rds. Since performance of this latter task is impaired bytemporal lobe lesions, Shallice and colleagues concluded that , if there wasany specific impairment in schizophrenic patients, this was consistentwith frontal rather than temporal damage. A number of other studieshave applied batteries of tests to schizophrenic patients in order toidentify a specific domain of impairment. As Table 1 shows, the resultsare not fully consistent since each study tends to implicate a slightlydifferent domain.

Cognitive impairmentsOne reason for these problems is that no psychological test is 'pure' .Many different underlying processes are necessary for good testperformance. If there is a circumscribed dom ain of impairment associatedwith schizophrenia this will best be described in 'cognitive', rather thanbehavioural, terms. Cognitive processes, such as selective attention,Table 1 Specific neuropsychological impairments in schizophreniaKolb & Wbh aw 10 'Fronted and temporal lobe' tasksSaykin ota/.6 Memo ry and learningMcKenna ( a/. 7 Semantic memoryShallke etal . 1 Executive tasksSho qeir at& May es" Effortful tasks

6 2 0 British Med ical Bulletin 199 6;52 (No . 3)


4/9


planning and episodic memory, exist at the interface between behaviourand brain function. The cognitive impairments have to be inferred fromthe pattern of test performance or, better still, from performance onexperimental tasks designed to test specific hypotheses.

Abnor m a l semantic categorisationPerformance on standardised tests is an obvious starting point formaking hypotheses about underlying cognitive impairments. Followingobservations of specific problems with semantic memory9, Chen and hiscolleagues n used an ingenious technique to examine the boundaries ofsemantic categories. Subjects were asked whether items such as robin,penguin, aeroplane or bell were examples of the category birds. When thetime to make such a categorisation is measured in normal volunteers,items clearly in the category (robin) or clearly outside the category (bell)are associated with rapid decisions, while items on the boundary(penguin) are associated with slow decisions. In terms of these decisionstimes, schizophrenic patients have enlarged semantic categories. Theyshowed the slowest responses for items just outside the category such asaeroplane. Categories such as 'birds' are imposed 'top down' in order toperform the task. We could easily switch to a new category, such asthings that fly, which would give a different ordering of the items listedabove. In terms of such a 'top down' categorisation process, we couldhypothesise that abnormally enlarged categories result from an impair-ment in some kind of 'surround inhibition' function so that items on theedge or just outside the category are not sufficiently inhibited. Resultsconsistent with this idea have been obtained in other studies using verydifferent paradigms13 '14 .

PerseverationElliot and her colleagues15 examined the abnormal behaviour displayedby many schizophrenic patients on tasks such as the Wisconsin cardsorting test. In this task, subjects must learn to sort cards on the basis ofone of various features such as colour and shape. Once the subject haslearned that colour is relevant and shape irrelevant, the rule is changed.The subject must now shift to a new feature and attend to something thatwas previously irrelevant. Elliot and her colleagues argue that there aretwo independent reasons for failing to switch in such a situation: (i) thesubject can not stop responding to the old feature (perseveration): and (ii)the subject can not switch attention to a feature that was previously

BriHth Medical Bulletin 1996;52 (No. 3) 62 1


5/9


irrelevant (learned irrelevance). Using an experimental paradigm whichdistinguished between these two possibilities, it was shown that theproblem lay in continued responding to the old feature and not in switchingattention to a previously irrelevant feature. At present, it is not clear to mewhether this deficit in attention can be related to the problem with semanticboundaries demonstrated by Chen and colleagues12 . This might be becausethe unifying concept has yet to be delineated, but another possibility is thatdifferent cognitive impairments are found in different patients.

Intellectual impa irme nts a nd schizophrenic syndromesAt any on e time, an unselected gro up of schizop hrenic pa tients will differwidely in their current mental state. An individual patient might beexperiencing extreme and continuous hallucinations and delusions, shemight manifest negative features and social withdrawal or she might befunctioning relatively normally. It would seem likely that these stateswould be associated with very different patterns of performance onpsychological tests. A number of studies (see also chapter by Liddle) havefound evidence that three clusters of signs and symptoms are sufficient tocapture the current mental state of most patients 16 . These arePsychomotor Poverty (poverty of speech, action and thought), Dis-organisation (incoherence of speech and incongruity of affect) andReality Distortion (hallucinations and delusions). These clusters areassociated with different patterns of test performance. In the large studyreported by Frith and his colleagues 2, general cognitive impairment wasassociated with poverty and disorganisation, but not with realitydistortion. However, the pattern of impairment differed between thepoverty and the disorganised syndrome. Poverty was associated withomission errors: poor verbal fluency, ignored signals in a vigilance task,while disorganisation w as associated with comm ission errors; odd wo rdsin verbal fluency, inappropriate responses in the vigilance task. Similarresults have been obtained in other studies (see Table 2).T a b l e 2 Cognit ive impairment! and the s igns and symptoms of schizophreniaHallucinations and delusion* Disorganised behaviour Negative features

DedineinrQ 2 D e d m e i n l Q " Odd verbal fluency 2-" Poorverfeal fluency2-17 Commission errors2 Omission errors2 Stroop interference" Stereotyped behaviour1*

622 British Medical Bulletin 1996;52 (No. 3)


6/9

Nouropsychology of schizophrenia

However, the relationship between specific impairments and currentsymptoms is by no means strong. For example, with the semanticanomalies discussed in the previous section, Spitzer and colleagues13found them to be related to formal thought disorder, Beech andcolleagues14 found them related to positive symptoms and Chen andcolleagues12 did not find them to be related to any aspect of current state.In the study of Frith and colleagues2, there were clear cut relationships,but the number of cases was very large (283). A power calculationsuggests that equivalent relationships would not be reliably detected instudies involving less than ~50 cases.If cognitive impairments are closely associated with current mentalstate, then a strong prediction would be that these impairments shouldchange with changes in mental state. This should be a particularlysensitive measure, since noise due to differences between patients wouldbe eliminated. Such studies are difficult to carry out, however, especiallyas they could be confounded by changes in drug trea tment. It is of interestthat Goldberg and his colleagues20 found that general cognitiveimpairments remained even after symptoms had been substantiallyreduced by treatment with clozapine.

Cogn itive mechanisms un derlying specific sy mptom sAll the studies described above are correlational. They address thequestion of whether schizophrenia or certain signs and symptoms areassociated with certain cognitive impairments. As is well known, suchstudies can not address questions of causation and will always beunsatisfactory since an apparent link may be spurious, being mediated bysome hidden factor. There is a sense in which schizophrenia will never besatisfactorily 'explained' in terms of cognitive impairments. Nor can anexplanation solely in terms of a gene or a particular form of brainabnormality be considered satisfactory. In contrast, it is possible to'explain' particular signs or symptoms in cognitive terms 19 and then linkthese to brain abnormalities. Such an explanation involves the descrip-tion of a cognitive mechanism that causes the symptom of interest. Onthe basis of such a description, it should be possible to: (i) designexperiments that reveal the functioning of the relevant cognitivemechanism; (ii) devise procedures which alter the symptom in patients;and (iii) devise procedures which elicit the symptom in normalvolunteers. In this way, we can achieve a truly experimental approachto the study of symptoms rather than relying on associations withcognitive impairments. As yet there is no cognitive account of symptomsin which the aims listed above have been fully achieved. The best

Brihth Mdical Bulletin 1996;32 (No. 3) 6 23


7/9


developed are probably the various attempts to explain auditoryhallucinations and certain delusions in terms of defects in self-monitoring.

Hallucinations and self-monitoringThe basic proposal is that verbal hallucinations occur when the patientexperiences his own speech or thought as coming from an externalsource. A straightforward prediction based on this idea is thathallucinating patients will have difficulty remembering whether a wordthey heard previously was spoken by themselves or by the experimenter.The results from studies using this approach have been equivocal2>21.How ever, as yet, the appropriate paradigm for testing the hypothesis hasnot been used.An elaboration of the self-monitoring theory considers how we knowabout our own actions. One theory is that we know that an action is ourown because of feed-forward signals (corollary discharge) associatedwith the motor command19. These feed-forward signals enable us tomodify behav iour and correct errors very rapidly since we do not have towait for peripheral feedback about the effects of an action. Certainsymptoms, including auditory hallucinations and delusions of control,might arise because the feed-forward signals associated with action arenot properly monitored. In the absence of the feed-forward signal,patients would be dependent on peripheral feed-back for knowing aboutand modifying their own actions. A number of experiments have shownthat, if peripheral visual feedback is eliminated, patients with passivityphenomena, such as delusions of control, have difficulty in making rapiderror corrections and in remembering the actions they have justperformed22"24. These results are consistent with a failure of the feed-forward mechanism. In an experiment specifically aimed at under-standing auditory hallucinations, the patient's own speech was fed backin a distorted form 25. In these circumstances, many patients said th at theycould hear 'another voice saying the same things that I am saying'. Thus,distorting the peripheral feed back caused patients to attribute their ow nvoice to an external source. In this case, we have effectively elicited aclassic symptom by an experimental procedure derived from a cognitivemodel of symptom generation. False attributions were not made bynormal controls or by patients who were currently symptom free or whohad predominantly negative features. The false attributions were mostlikely to be made by patients concurrently experiencing both hallucina-tions anddelusions.

6 2 4 British MtdkalBullttin 1996^2 (No. 3)


8/9


The concept of corollary discharge was developed by physiologists,rather than psychologists26 and so the self-monitoring model can readilybe mapped onto brain function. In the case of vocalisation, there isevidence that signals arising in the anterior cingulate cortex (concernedwith the initiation of vocalisation) are sent to auditory cortex causingmodification of the response of this area to vocalisation 27 . A recent brainimaging study has provided preliminary evidence (see chapter by Liddle)for functional disconnections between these areas in patients withschizophrenia2 8.

ConclusionsWhile there is ample evidence that many patients with a diagnosis ofschizophrenia perform badly on neuropsychological tests, it is not clearthat there is a characteristic pattern of impairment associated with thisdiagnosis. Investigations of particular signs and symptoms are likely to bemore fruitful, particularly if it is possible to develop models that explainthese features in terms of underlying cognitive deficits. Some success hasbeen achieved in explaining auditory hallucinations and certain passivityphe nom ena in terms of a defect in self-monitoring. Such a defect might bethe result of functional disconnections between frontal areas concernedwith initiating actions and posterior areas concerned with the perceptionof the effects of these actions.

AcknowledgementsI have benefited from several discussions about the neuropsychology ofschizophrenia with Rhiannon Corcoran, Uta Frith and Karl Friston.

References1 Johnst one EC. ed. Disabilities and circumstances of schizophrenic pat ie nt s- a follow-up study. Br

] Psychiatry 1 9 9 1 ; 159 (Suppl. 13)2 Frith CD , Leary J, Cahill C, Johnsto ne EC. Disabilities and circumstances of schizophrenic

patients-a follow-up study. IV. Performance on psychological tests: demographic and clinicalcorrelates of the results of these tests. Br ] Psychiatry 1 9 9 1 ; 159 (Suppl. 13) : 26-9

3 Nelson H E, Pantelis C, Carruthe rs K, Speller J, Baxendale S, Barnes TRE. Cognitive functioningand symptomatology in chronic schizophrenia. Psychol Med 1990; 20: 357-65

4 Payne RW. Cognitive abnormalities. In: Eysenck HJ. ed. Handbook of Abnormal Psychology.London: Pitman, 1973; 420-83

Brihsh Medical Bulletin 199622 (No. 3) 62 5


9/9

neuropsychology of schizophrenia

Documents