Social cognition in frontotemporal dementia and Huntington’s disease

Neuropsychologia 41 (2003) 688–701

Social cognition in frontotemporal dementia and Huntington’s disease J.S. Snowden a,∗ , Z.C. Gibbons a , A. Blackshaw a , E. Doubleday a , J. Thompson a,b , D. Craufurd b , J. Foster c , F. Happé d , D. Neary a a

d

Cerebral Function Unit, Greater Manchester Neuroscience Centre, Hope Hospital, Salford M6 8HD, UK b University Department of Medical Genetics, St. Mary’s Hospital, Manchester M13, UK c Department of Psychology, University of Western Australia, Crawley, Perth, WA 6009, Australia Social Genetic and Developmental Psychiatry Research Centre, Institute of Psychiatry, King’s College, London, UK Received 19 February 2002; received in revised form 2 October 2002; accepted 2 October 2002

Abstract Frontotemporal dementia (FTD) and Huntington’s disease (HD) are degenerative disorders, with predominant involvement, respectively of frontal neocortex and striatum. Both conditions give rise to altered social conduct and breakdown in interpersonal relationships, although the factors underlying these changes remain poorly defined. The study used tests of theory of mind (interpretation of cartoons and stories and judgement of preference based on eye gaze) to explore the ability of patients with FTD and HD to interpret social situations and ascribe mental states to others. Performance in the FTD group was severely impaired on all tasks, regardless of whether the test condition required attribution of a mental state. The HD group showed a milder impairment in cartoon and story interpretation, and normal preference judgements. Qualitative differences in performance were demonstrated between groups. FTD patients made more concrete, literal interpretations, whereas HD patients were more likely to misconstrue situations. The findings are interpreted as demonstrating impaired theory of mind in FTD, as one component of widespread executive deficits. In HD the evidence does not suggest a fundamental loss of theory of mind, but rather a tendency to draw faulty inferences from social situations. It is concluded that social breakdown in FTD and HD may have a different underlying basis and that the frontal neocortex and striatum have distinct contributions to social behaviour. © 2002 Elsevier Science Ltd. All rights reserved. Keywords: Frontotemporal dementia; Huntington’s disease; Social cognition, theory of mind; Behaviour

1. Introduction Frontotemporal dementia (FTD) and Huntington’s disease (HD) are degenerative brain disorders that affect frontostriatal systems. FTD is a predominantly neocortical disorder, characterised by radical alterations in personality, emotions, and social, interpersonal conduct [12,25,35,44,46,47,56]. Behavioural changes include disinhibition, tactlessness, and loss of social proprieties [12,37,44,45,47]. Cognitive assessment typically shows deficits predominantly in frontal executive functions [47,57], indicating deficits in abstraction, problem solving, attention, mental set shifting, sequencing, and mental generation of information. Patients are not clinically amnesic, although formal memory test performance is often inefficient, attributed to executive impairments. Neuroimaging [48,62] and pathological studies [41] of FTD demonstrate severe frontal and anterior temporal ∗ Corresponding author. Tel.: +44-161-787-2561; fax: +44-161-787-2993. E-mail address: [email protected] (J.S. Snowden).

neocortical atrophy, which may be largely confined to orbital regions, or (particularly with progression of disease) more widespread extending into anterior cingulate and dorsolateral frontal cortex. Modest pathological changes in the striatum reflect the emergence of striatal neurological signs usually relatively late in the disease course. Huntington’s disease (HD) is a predominantly subcortical disorder, distinguished clinically by its characteristic involuntary movements [30]. Patients’ social conduct is altered, albeit less profoundly than in FTD, and there is frequently severe breakdown in interpersonal relationships. Patients are often described as self-centred, lacking in sympathy and empathy, and mentally inflexible, sometimes with fixed ideas, which may not be consistent with the prevailing view or available evidence. As in FTD deficits have been reported in the processing of emotions [23,32,59]. Cognitive changes are predominantly in the realm of frontal executive function [14], although generally less marked in degree than in FTD, and memory impairment is ascribed to inefficient encoding and retrieval strategies rather than a primary failure of retention.

0028-3932/02/$ – see front matter © 2002 Elsevier Science Ltd. All rights reserved. PII: S 0 0 2 8 - 3 9 3 2 ( 0 2 ) 0 0 2 2 1 - X

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

Pathological [40,63] and structural neuroimaging [6,38] studies of HD have demonstrated marked atrophy of caudate and putamen, which form the dorsal part of the striatum or neostriatum. This is present even in the early stages of disease [4], and has been reported in some studies in pre-symptomatic individuals who carry the HD mutation [5]. Some frontal neocortical atrophy may also occur later in the disease course [7], assumed to be at least partly (although not necessarily exclusively) secondary to striatal differentiation [40]. Thus, FTD and HD represent complementary disorders in which there is a virtual, although not exclusive, double dissociation with respect to the distribution of degenerative change within the frontal neocortex and striatum. FTD and HD thus provide ideal models for the study of frontal-striatal function. The striatum has traditionally been recognised for its importance in the domain of motor functioning, in the execution of learned motor plans [42]. Conditions such as HD attest to its crucial role also in cognition. The identification of parallel and segregated frontal-subcortical circuits, distinguished by their areas of origin in the frontal cortex [3,43], has led to the notion that the striatum is intimately linked functionally to the cerebral cortex. The assumption is that analogous cognitive deficits may arise from disruption at different levels (i.e. frontal cortical or striatal) of the circuit. Commonalities between FTD and HD with respect to the prominence of behavioural changes and pattern of cognitive deficits are thus unsurprising. Nevertheless, it cannot be inferred that deficits underlying FTD and HD are identical. Executive tasks make multiple demands, so that test scores may mask fundamental differences in the reason for failure. Similarly, disordered social behaviour might have different underlying substrates. Comparative studies of FTD and HD ought to clarify the nature of change in each condition. Moreover, in view of the predominance of frontal neocortical changes in FTD and of striatal changes in HD, such studies provide the potential for improving knowledge of the relative contributions of the frontal lobes and striatum in behaviour and cognition. Traditional executive tasks do not capture the full range of abnormalities in FTD and HD and may be a relatively poor predictor of the patient’s functioning in daily life. Indeed, some patients with FTD, in whom the pathology is confined to the orbital regions of the frontal lobes, perform relatively well on conventional executive tasks, despite impaired judgement and gross breakdown in their social conduct in daily life [39,57]. Such a finding is consistent with reports that lesions of the orbital frontal lobes may give rise to severe breakdown in social behaviour in the context of normal executive functioning [15,16,54]. In HD, disorganised behaviour and breakdown in interpersonal relationships in daily life are often prominent clinical features, outweighing changes in neuropsychological test performance. There are at least two factors that are likely to contribute to the relative insensitivity of traditional tests to some of the changes in

689

FTD and HD. Traditional neuropsychological tests are structured and typically require a constrained set of responses. By contrast, everyday life situations are open-ended, and require self-generated structure and organisation. Secondly, traditional tests are impersonal, whereas everyday life involves social interaction. Neuropsychological tasks that are both open-ended and involve interpretation of social scenarios are likely to be particularly informative in FTD and HD because they mirror the daily life situations in which FTD and HD patients so dramatically fail. They may also have the potential to reveal fundamental differences between FTD and HD. Recent years have seen an accumulation of literature on social cognition [1,2]. A core component of social functioning is the capacity to attribute independent mental states to others and to predict other people’s behaviour on the basis of their mental states, a capacity known as “theory of mind” [9,36,51]. There is a growing body of evidence from both neuroimaging [10,18,20,21,27] and brain lesion studies [24,29,53,60,61] that the frontal lobes have a pivotal role in theory of mind. However, to date there have been no direct comparisons in performance on tests of social cognition between patients with predominantly frontal neocortical and predominantly striatal pathology. Clinical observation of patients with FTD and HD leads to the prediction that performance on tests that require interpretation of event scenarios is likely to differ. FTD patients typically lack insight into the change in their own behaviour and appear oblivious of the effects that their behaviour has on others, leading to the prediction that such patients show a genuine loss of theory of mind. By contrast, at clinical interview HD patients may make pertinent and insightful remarks about the effects of their illness on a close relative (e.g. “It is hard on my husband having to do everything for me. He must get very fed up”). Such apparent cognisance of others’ mental states leads to the prediction that social breakdown in HD arises for reasons other than a primary inability to ascribe mental states to others. In FTD, a purported problem in theory of mind is unlikely to be exclusive. FTD patients commonly show concreteness of thought. A concrete interpretation of events would be expected to be manifest in a general difficulty in the interpretation of social scenarios, even when they do not depend on attribution of mental states. The present study investigates the ability of FTD and HD patients to interpret social situations and explores by means of error analysis possible differences between the two groups. The study involved four tasks drawn from the literature on social cognition that have been used to address theory of mind. The tasks differ with respect to their level of difficulty. The cartoon and story tasks (tasks 1–3) make relatively great mental demands on the patient raising the possibility that they may exceed the capabilities of some patients for reasons that have little to do with social cognition per se. The judgement of preference task (task 4) examines the capacity for mental state attribution while minimising

690

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

executive demands. It was predicted that FTD patients would show a general impairment, compared to normal, in their ability to interpret situations, but that this should be particularly pronounced when attribution of mental states is required. It was anticipated that HD patients would show impaired performance relative to normal, but to a lesser extent than FTD. It was predicted, however, that HD patients would not show a disproportionate impairment in tasks dependent upon mental state attribution compared to tasks involving interpretation of non-social scenarios. The pattern of performance ought to clarify the nature of patients’ deficits more precisely and may help to identify factors that contribute to the breakdown in social functioning in FTD and HD.

2. Methods 2.1. Participants Two patient groups and a normal control group took part in the study. Informed written consent was obtained from participants and/or their next of kin. The study was granted approval by the local Ethics Committee. 2.1.1. Frontotemporal dementia (FTD) The FTD group comprised 13 consecutive patients, 9 men and 4 women, referred to a Neurology Department Specialist Dementia Clinic who met clinical criteria for FTD [46]. Diagnosis was based on historical information, neurological examination and neuropsychological assessment, and supported by structural (magnetic resonance) and functional (single photon emission computerised tomography) brain imaging. The presenting feature in all cases was personality change and all patients had demonstrable frontal executive impairments on cognitive evaluation. Patients were in the mild to moderate stages of the disease, and were physically well. Neurological examination was entirely normal in 10 patients. The remaining three patients showed a mild degree of limb rigidity. Neuroimaging showed changes predominantly in orbital frontal cortex in nine patients, and widespread frontal lobe changes in the remaining four patients. Demographic information and clinical features are shown in Table 1. The patients had a mean Mini Mental State Examination (MMSE) [19] score of 22. The table also shows scores on category and letter fluency tests [58] (total

number of animals and words beginning with F generated in 1 min) and the number of categories achieved in the modified version of the Wisconsin Card Sorting test [49]. The low scores highlight the presence of frontal executive deficits in the FTD group. 2.1.2. Huntington’s disease (HD) The HD group consisted of 13 consecutive patients, 5 males and 8 females, attending a regional HD clinic. In all, the presence of HD had been confirmed by genetic testing and all showed the characteristic choreiform movement disorder, and cognitive changes associated with the disorder. Patients had a mean Total Functional Capacity score [55] of 9.5 indicating that they were in the mild to moderate stages of disease. They had a mean motor deficit score of 26/124, range 5–55, as measured by the Unified Huntington’s Disease Rating Scale (UHDRS) [31], consistent with mild to moderate disease. Six patients were taking prescribed medications for the treatment of mood changes, particularly irritability. The remaining seven patients were on no medication. No imaging data were available. A definitive diagnosis of HD can be made on the basis of the clinical features and genetic test, so that neuroimaging was not clinically justified. Patient demographics and clinical data are shown in Table 1. They were younger than the FTD patients (t = 3.5, P < 0.002), commensurate with the younger onset age of HD, and they had been clinically symptomatic for longer (t = 2.2, P = 0.04), consistent with the more protracted course of HD. The HD group did not differ from the FTD group with respect to MMSE or category and letter fluency scores. However, performance was less impaired on the Wisconsin Card Sorting test, as measured by the greater number of categories achieved (t = 10.7, P < 0.0001). 2.1.3. Controls The control group consisted of 18 people who were spouses of participants in the patient groups. All were healthy individuals who had no history of neurological disease, head injury or alcohol abuse. The control group covered a wider distribution of ages than the two patient groups (Table 1), reflecting the fact that they were drawn from the spouses of both groups. The mean difference in age between controls and the two patient groups did not reach statistical significance.

Table 1 Demographic, clinical and neuropsychological characteristics Group

Number

Male:female

Age (years)

Duration illness

MMSE

Animals per minute

F words per minute

WCST categories

FTD HD Control

13 13 18

9:4 5:8 8:10

60 (7) 50 (7) 49 (23)

3 (2) 6 (3) n/a (n/a)

22 (6) 25 (3) n/a (n/a)

12 (4) 14 (4) n/a (n/a)

6 (5) 8 (4) n/a (n/a)

1.4 (1.9) 5.5 (0.9) n/a (n/a)

The data represent mean (S.D.).

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

2.2. Task 1: single cartoon abstraction 2.2.1. Materials The materials, taken from Happé et al. [28], consisted of 12 humorous cartoons. In six, designated ‘mental’ cartoons, the humour related to a cartoon character’s mistaken belief or deception, so that humour appreciation required inference of a person’s mental state. In one cartoon, for example, a man is shown cuddling a young woman who is sitting on his lap, while, with his free hand, he is tapping a ping-pong ball with a bat. The humour lies in the fact that an older woman sitting in the adjacent room, within earshot but out of view of the couple, is deceived into believing that the man is playing table tennis, whereas in reality he is otherwise occupied. In six cartoons, designated ‘physical’ and matched for difficulty with the ‘mental’ cartoons, the humour related to physical properties or anomalies in the cartoon and did not require inference of a person’s mental state. For example, in one cartoon, a line of musicians is shown entering the stage door, each carrying a musical instrument case. One man has no head, but is carrying a head-shaped instrument case. Illustrative examples of cartoons are given by Happé [28]. 2.2.2. Procedure The cartoons were presented in randomised order, in accordance with Happé et al. [28], and subjects were asked to describe what was funny about each. Responses were transcribed verbatim and the time taken to respond was recorded. Subjects were prompted with “anything else?” to encourage as full a response as possible. Cartoons remained in full view until their response was complete. 2.2.3. Scoring Performance accuracy was measured using the scoring system devised by Happé et al. [28]. Three points were awarded for a full and explicit explanation, two points for a partial or implicit explanation, one point for reference to relevant parts of the cartoon, but without further explanation and zero for patently incorrect responses including omissions. Examples of the marking criteria are given in Happé et al. [28]. Scores for each test item were summated, yielding a total maximum score of 18 (6 × 3) for each of the two cartoon types. Errors, defined as responses yielding a less than perfect score, were further classified as follows: (i) omissions (“don’t know” responses), (ii) concrete responses (itemisation of elements without integration), (iii) descriptions (responses limited to a description of the cartoon, involving integration of elements but no inferences that go beyond the cartoon’s content), (iv) misconstructions (responses that go beyond a description of the cartoon’s contents but involve drawing faulty inferences) and (v) partial responses (responses that involve correct inferences but are incomplete or implicit rather than explicitly stated). Accuracy measures and error classification were rated independently by four raters, who were blinded to clinical diagnosis. The ratings

691

used in the analyses were based on a consensus from the four raters. In addition to accuracy measures and error classification, the total number of words contained in each response, and the number of action verbs (e.g. walking, pushing) and mental state verbs (e.g. thinking, expecting) were calculated. Word and verb counts were calculated by a single author (ZG), and verified by two others. Calculations were made blind to diagnosis. 2.3. Task 2: cartoon pairs 2.3.1. Materials The materials, taken from Happé et al. [28], consisted of 10 cartoon pairs, one of which was humorous and the other was not, having had the key humorous element replaced. Five of the humorous cartoons were of a ‘mental’ type, in that appreciation of the humour depended on understanding of a cartoon character’s ignorance, false belief or act of deception. Five humorous cartoons were of a ‘physical’ type, in that the humour was based on physical properties of the cartoon and did not require inferences about a character’s mental state. Illustrative examples of cartoon pairs are given by Happé [28]. 2.3.2. Procedure Cartoon pairs were presented side by side in accordance with Happé et al [28], with the left–right order being counter-balanced across items. Subjects were asked to select which cartoon of the pair they considered to be the funny one. Accuracy of selection and time to respond were recorded. Subjects were then asked to describe why the cartoon was funny. As in the previous task, subjects were cued with “anything else?” to elicit as full a response as possible. Cartoons remained in full view until the response was complete. Responses were scored as for task 1, in terms of accuracy measure, error types and word and verb count. 2.4. Task 3: story comprehension 2.4.1. Materials The materials were drawn from Happé et al. [28] and had originally been adapted from a study of theory of mind in autism [26]. They consisted of 16 short passages, 8 of which were of a ‘mental-type’ and 8 ‘physical’. The mental stories involved false belief, acts of deception, bluff and double bluff and were followed by questions that required an inference about a character’s thoughts, feelings or intentions. The physical stories involved logical or practical situations, and although the stories also contained people, questions required inferences about physical causation or logical consequence and not about a character’s mental state. 2.4.2. Procedure Participants were asked to read each passage silently, as described by Happé [28], and to inform the examiner when

692

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

they had done so, following which a question related to the passage would be asked. Subjects were advised to absorb as much of the information contained in the passage as possible, as they would be unable to refer back to the passage later. When participants indicated that they had assimilated the story the page was turned to reveal the question, which was read aloud to them. The question, but not the story, remained in front of the subject during their response. Mental and physical stories were blocked and presented in counter-balanced order, in accordance with Happé et al. [28]. The examiner recorded the time taken to study the passage and the answers given. The standardised scoring scheme devised by Happé et al. [28] was adopted. Answers were credited with two points for correct answers that gave a full and explicit account, one point for partial or implicit answers and no points for incorrect responses. Examples of stories and of the scoring criteria are given in Happé et al. [28]. In addition, a classification scheme was devised to characterise the nature of errors, similar to that used in the cartoon tasks. Errors were recorded as omissions (don’t know responses), concrete responses (reiterations of parts of the passage), descriptions (responses limited to a description of the story, without drawing inferences), and misconstructions (bizarre or incorrect inferences). As for the cartoon tasks, the total number of words, the number of action verbs, and the number of mental state verbs contained in each response were calculated. 2.5. Task 4: judgement of preference 2.5.1. Materials The task was based on one described previously by Baron-Cohen et al. [8] and involved the ability to judge

preference based on eye gaze. Unlike the previous tasks it involved a structured, forced-choice, rather than an open-ended response. The materials consisted of 48 A4-size cards presented in landscape format, each showing the cartoon outline of a face, positioned centrally and four coloured pictures of items belonging to a single category (e.g. apple, strawberry, banana, pineapple) one in each of the four corners of the card. The eye gaze of the face (upper-left, lower-left, upper-right or lower-right) was directed towards one of the four pictures. Across the 48 items, six object categories were used: cartoon characters, fruits, bakery items, houses, jumpers, and cars, each category having eight exemplars. For the first 24 test items (arrow condition), a heavy black arrow was also present, which pointed to one of the pictures other than that towards which the face’s eye gaze was directed. The remaining 24 cards (neutral condition) were a duplicate of the first 24, with the exception that no arrow was present. The direction of eye gaze and the arrow position was pseudo-random, occurring in each of the four positions an equal number of times across the stimulus set. An example of the test stimuli is shown in Fig. 1. 2.5.2. Procedure Each card was presented individually, using a blocked presentation, the arrow condition being presented first. Participants were instructed to point to the one of the four pictures on the card that the central face “likes best”. Responses were recorded on a score sheet by the examiner. Participants were not given feedback about their choices. On completion of the task, participants whose responses did not accord with the direction of eye gaze of the cartoon face were re-presented with the stimuli and asked to point to the picture that the face “is looking at”. They were also shown the four pictures on a card devoid of face and arrow and asked to indicate

Fig. 1. Example of stimuli used in task 4 involving judgement of preference.

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

which one of the four was their personal favourite. Choices were recorded. The control “looking at” task was not administered to all subjects. The preference task was judged to be so easy that it was expected to present no difficulty to a normal adult. Given that a correct judgement of preference implied knowledge of the item being looked at it seemed superfluous, and perhaps slightly insulting, to ask also for a “looking at” judgement. Nevertheless, the “looking at” task was deemed critical for those people who had difficulty on the judgement task, to distinguish specific problems in mental state attribution from general problems in attention or other executive function. The arrow condition and neutral condition were scored separately, each item being credited one point if the patient’s picture selection accorded with the direction of eye gaze of the central face. Errors in the arrow condition were coded as ‘arrow’ if the participant had incorrectly selected the picture corresponding to the direction of the arrow, ‘favourite’ if the participant had incorrectly chosen their personal favourite, ‘perseveration’ if the participant pointed to the same item position as their immediately preceding response, and ‘random’ if an incorrect choice did not fit into any of the above error types. The errors made in the neutral condition were coded as above, but without the ‘arrow’ error type.

3. Results 3.1. Task 1: single cartoon abstraction 3.1.1. Time taken to respond The patient groups did not differ significantly in terms of the time taken to respond for either mental or physical cartoons.

693

3.1.2. Performance accuracy Fig. 2 illustrates the mean scores for the cartoon explanations across groups and cartoon types. A repeated measures ANOVA showed a main effect of group (F (2, 41) = 17.3, P < 0.0001), but no effect of cartoon type, nor group by cartoon-type interaction. Post hoc Tukey analyses showed that controls performed significantly better than both FTD (P < 0.0001) and HD patients (P = 0.004). There was a trend towards better performance in HD compared to FTD (P = 0.09). 3.1.3. Error type A preliminary examination of error patterns showed that error type was not influenced by cartoon type, so analyses of errors were based on a summation from both mental and physical cartoons. FTD patients made significantly more omissions (t = 2.7, d.f. 29, P = 0.01) and concrete responses (t = 4.1, d.f. 29, P < 0.0001) than the control group, and more concrete responses (t = 2.7, d.f. 24, P = 0.01) than the HD group. In contrast, the HD patients made more misconstruction errors than both the control (t = 3.9, d.f. 29, P = 0.001) and the FTD (t = 3.8, d.f. 4, P = 0.001) groups. When they made errors, control subjects were more likely than the HD (t = 3.7, d.f. 29, P = 0.001) and the FTD (t = 4.0, d.f. 29, P < 0.0001) groups to produce partially correct responses. The following are examples of responses to the ‘ping-pong’ cartoon, described in Section 2.2.1, in which a man deceives an older woman in the adjacent room into believing that he is playing table tennis by tapping a ping-pong ball while he cuddles his younger female companion. An FTD patient’s concrete response consisted of: “He’s bouncing the ball on the table”. An HD patient’s misconstruction response consisted of: “They’re having a bit of nooky while the wife’s

Fig. 2. Mean accuracy scores for single cartoon interpretation (task 1) as a function of cartoon type.

694

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

sat in there. She’s thinking ‘At least he’s leaving me alone. Peace!”’. 3.1.4. Word and verb count A repeated measures ANOVA comparing the number of words per response elicited by the three groups for the two types of cartoon revealed a main effect of cartoon type (F (1, 39) = 12.6, P = 0.001), but no effect of group nor group by cartoon type interaction. Mental cartoons elicited lengthier responses than physical cartoons. A repeated measures ANOVA comparing the number of action verbs produced by the three groups for the two cartoon types showed no main effect of group, but a small effect of cartoon type (F (1, 41) = 7.0, P = 0.01). More action verbs were produced for physical than for mental cartoons. There was no interaction effect of group by cartoon type. A repeated measures ANOVA comparing the number of mental state verbs produced showed a main effect of group (F (2, 41) = 7.9, P = 0.001), a main effect of cartoon type (F (1, 41) = 63.0, P < 0.0001), and an interaction effect of group by cartoon type (F (2, 41) = 9.5, P < 0.0001). Post hoc analyses showed that FTD patients produced significantly fewer mental state verbs than controls (P = 0.001). Other group comparisons were non-significant. As expected, mental cartoons elicited more mental state verbs than physical cartoons. The interaction effect resulted from a smaller disparity in number of mental state verbs for the two cartoon types in the FTD group (t = 2.1, P = 0.06) compared to the HD group (t = 4.9, P < 0.0001) and controls (t = 7.1, P < 0.0001). 3.1.5. Frequency analysis The frequency with which overall performance accuracy scores were superior or inferior in the ‘mental’ compared to the ‘physical’ cartoon condition was calculated for each group and frequency values were submitted to a Sign test. FTD patients were significantly more likely to perform worse in the ‘mental’ than the ‘physical’ condition (two-tailed test P < 0.01), whereas other groups showed no significant bias. 3.2. Task 2: cartoon pairs 3.2.1. Choice of cartoon pair The number of correct choices made by the three groups for mental and physical cartoons is shown in Table 2. A repeated measures ANOVA comparing the number of correct Table 2 Correct selection of forced-choice cartoons Group

Mental cartoons

Physical cartoons

FTD HD Control

3.2 (1.8) 4.1 (1.2) 4.6 (0.5)

1.8 (1.3) 3.4 (1.2) 4.6 (0.7)

The data represent mean (S.D.).

choices for the two cartoon types showed a main effect of group (F (2, 41) = 19.6, P < 0.0001), and cartoon type (F (1, 41) = 11.9, P = 0.001), and an interaction effect of group by cartoon type (F (2, 41) = 3.9, P = 0.03). Post hoc analyses revealed that controls were significantly more accurate than the FTD patients (P < 0.0001), and there was a trend towards greater accuracy of controls compared to HD patients (P = 0.07). HD patients made more correct choices than FTD patients (P = 0.002). More correct choices were made for mental than physical cartoons, particularly in the patient groups compared to controls. 3.2.2. Time taken to respond A repeated measures ANOVA showed no difference in response times in the three groups and there was no effect of cartoon type on response time. 3.2.3. Accuracy of interpretation Fig. 3 illustrates scores for the three groups across cartoon types. A repeated measures ANOVA showed a main effect of group (F (2, 41) = 21.4, P < 0.0001), but no effect of cartoon type or group by cartoon-type interaction. Post hoc analyses showed that controls performed significantly better than both FTD (P < 0.0001) and HD patients (P = 0.007). HD patients achieved higher scores than FTD patients (P = 0.01). 3.2.4. Error type A preliminary examination of error patterns showed that error type was not influenced by cartoon type, so analyses of errors were based on a summation from both mental and physical cartoons. FTD patients made more omission and single element concrete responses than HD patients (t = 2.2, d.f. 23, P = 0.04; t = 3.4, d.f. 23, P = 0.003, respectively) and controls (t = 3.3, d.f. 28, P < 0.003; t = 3.8, d.f. 28, P = 0.001, respectively). HD patients, by contrast, made more misconstruction errors than FTD patients (t = 3.7, d.f. 23, P = 0.001) and controls (t = 4.2, d.f. 29, P < 0.0001). Control subjects made significantly more partially correct responses than both FTD (t = 4.3, d.f. 29, P < 0.0001), and HD patients (t = 2.3, d.f. 29, P = 0.03). 3.2.5. Word and verb count A repeated measures ANOVA showed that the groups did not differ in terms of the number of words contained in their responses. There was also no effect of cartoon type on length of response and no interaction effect. A repeated measures ANOVA comparing the number of action verbs produced by the three groups showed a small effect of group (F (2, 41) = 4.1, P = 0.02). FTD patients produced fewer overall action words than controls (P = 0.03). There was no main effect of cartoon type. There was however an interaction effect of group by cartoon type (F (2, 41) = 6.9, P = 0.003), reflecting the fact that whereas control subjects tended to produce more action verbs for physical than mental cartoons, the HD group showed the reverse effect.

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

695

Fig. 3. Mean accuracy scores for forced-choice cartoon interpretation (task 2) as a function of cartoon type.

A repeated measures ANOVA comparing the number of mental state verbs produced by the three groups showed a main effect of group (F (2, 41) = 4.6, P = 0.02), and of cartoon type (F (1, 41) = 25.0, P < 0.0001), and an interaction effect of group by cartoon type (F (2, 41) = 3.8, P = 0.03). Post hoc analyses revealed that FTD patients produced fewer mental state verbs than controls (P = 0.01). As expected more mental state verbs were produced for mental than for physical cartoons, but the difference was proportionally smaller for the FTD group. 3.2.6. Frequency analysis The frequency with which overall performance accuracy scores were superior or inferior in the ‘mental’ compared to the ‘physical’ cartoon condition was calculated for each group and frequency values were submitted to a Sign test. Control subjects were significantly more likely to perform worse in the ‘mental’ than the ‘physical’ condition (two-tailed test P < 0.02), whereas the patient groups showed no significant bias. 3.3. Task 3: story comprehension 3.3.1. Time taken to respond A repeated measures ANOVA comparing the time to respond by the three groups for the two story types showed a main effect of group (F (2, 27) = 4.1, P = 0.03), but no effect of story type or interaction effect. HD patients were slower to respond than FTD patients (P = 0.04), and tended to be slower than control subjects (P = 0.06). No other group comparisons were significant.

3.3.2. Accuracy Fig. 4 shows accuracy scores for the story comprehension test in the three groups for the two story types. A repeated measures ANOVA showed a main effect of group (F (2, 30) = 19.3, P < 0.0001), but no effect of story type or interaction effect. Control subjects performed better than both FTD (P < 0.0001) and HD patients (P = 0.01), and HD patients performed better than FTD patients (P = 0.01). 3.3.3. Error type FTD patients were more likely than controls to make omission errors (t = 2.4, d.f. 22, P = 0.03), concrete responses (t = 4.1, d.f. 22, P < 0.0001) and description responses (t = 5.2, d.f. 22, P < 0.0001). FTD patients also made more concrete responses (t = 2.1, d.f. = 13, P = 0.05) and description responses (t = 3.0, d.f.13, P = 0.01) than HD patients. HD patients showed a trend towards more misconstruction errors than the control group (t = 1.9, d.f 25, P = 0.07). Control subjects were more likely than FTD patients to make partially correct responses (t = 3.6, d.f. 22, P = 0.002). 3.3.4. Word and verb count There was no difference in the overall number of words per response produced by each of the three groups and length of response was not influenced by story-type. A repeated measures ANOVA comparing the number of action verbs produced by the three groups showed no effect of group, story type, or interaction effect of group by story type. A repeated measures ANOVA comparing the number of mental state verbs produced by the three groups showed

696

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

Fig. 4. Mean accuracy scores for story interpretation (task 3) as a function of story type.

no main effect of group, but a main effect of story type (F (1, 30) = 56.2, P < 0.0001), and a trend towards an interaction effect of group by story type (F (2, 30) = 3.1, P = 0.06). More mental state verbs were elicited for mental than for physical stories, but this increase was smaller for the FTD group than the other groups. 3.3.5. Frequency analysis The frequency with which overall performance accuracy scores were superior or inferior in the ‘mental’ compared to the ‘physical’ story condition was calculated for each group and frequency values were submitted to a Sign test. No group showed a statistical bias towards better performance for one or other story type. 3.4. Task 4: determining preference from eye gaze 3.4.1. Accuracy Fig. 5 shows mean accuracy scores for each group, when a distracting arrow was present and absent. A repeated measures ANOVA showed a main effect of group (F (2, 39) = 5.5, P = 0.008), but no effect of condition (arrow versus no arrow), and no interaction effect of group by condition. Post hoc analyses showed that FTD patients made more errors than both HD patients (P = 0.03) and controls (P = 0.01). HD patients did not differ from controls and performance in both approached ceiling levels. 3.4.2. Error types In view of the rarity of errors made by the HD and control groups, their responses were not subjected to analysis of error type. In the FTD group incorrect responses were dominated by “favourite” errors, which accounted for 71% of all

incorrect responses (patients selected their personal favourite picture disregarding eye gaze). Although a relatively high number of incorrect responses (39%) accorded with the direction of the arrow, in the majority of these instances the response also corresponded to the patient’s favourite picture, so that the basis for the correct choice was ambiguous. The lack of a statistical effect of condition (arrow present versus arrow absent), implying that the presence of the arrow had no overall effect on performance accuracy, suggests that incorrect choices were largely being made on the basis of personal favourite and not arrow direction. One single FTD patient appears to represent an exception to this general rule. In the arrow condition 20 of his 22 incorrect responses corresponded to the arrow direction. Perseverations of a single response position and random errors were rare in all patients. 3.4.3. Judgement of eye direction Three of the FTD patients had exhibited chance level performance in the judgement of preference task. These individuals were subsequently asked to indicate which of the four items the cartoon face was looking at. Two FTD patients had no difficulty carrying out the “looking at” task and scored, respectively 100 and 92% correct, compared with performance, respectively of 21 and 17% correct in the “preference” task. These differences in performance for the “like” and “look at” tasks were highly significant (McNemar test χ 2 = 17.1, P < 0.001; χ 2 = 16.1, P < 0.001 for the two patients, respectively). By contrast the third patient persisted in selecting her own personal favourite item, disregarding the direction of eye gaze of the cartoon face. Her accuracy score of 25% correct was unchanged from her chance level score in the preference task.

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

697

Fig. 5. Mean score for judgement of preference (task 4) as a function of the presence or absence of a distractor arrow. Table 3 Correlation between social cognition and standard executive test scores Social task

Executive task Category fluency (n = 12)

Letter fluency (n = 13)

WCST categories (n = 11)

(a) FTD Task 1 Task 2 Task 3 Task 4

(single cartoons) (forced-choice cartoons) (stories) (preference judgement)

0.58∗ 0.69∗ 0.13 0.28

0.62∗ 0.53 0.01 0.36

0.57 0.75∗∗ 0.58 0.20

(b) HD Task 1 Task 2 Task 3 Task 4

(single cartoons) (forced-choice cartoons) (stories) (preference judgement)

0.38 0.37 0.73∗∗ 0.37

0.04 0.02 0.38 0.28

0.09 0.08 0.11 0.71∗

∗

P < 0.05. P < 0.001.

∗∗

3.5. Relationship of performance on social cognition and standard executive tests

3.6. Relationship of FTD performance to distribution of frontal atrophy

For each of the four social cognition tasks a total performance accuracy score was calculated. In the case of the cartoon and story tasks, this involved summating accuracy scores for the mental and physical conditions. In the case of the judgement of preference task scores for the arrow and no arrow conditions were summated. The relationship between total accuracy scores and performance on the category fluency (animals generated in one minute), letter fluency (F words generated in one minute) and Wisconsin Card Sorting test (WCST) was examined. Table 3 shows the correlation between performance on the experimental tasks and standard executive tests. Modest relationships were found but these were not consistent across tasks or for the two groups.

Nine FTD patients showed predominant orbitofrontal abnormalities on neuroimaging, whereas four patients showed widespread frontal lobe changes, extending into dorsolateral frontal cortex. Patients with widespread changes performed worse that those with more circumscribed changes on the cartoon and story tasks (P < 0.05) but not the preference task.

4. Discussion Both FTD and HD groups were impaired relative to controls in their interpretation of humorous cartoons and story vignettes. The FTD group was more severely affected than

698

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

the HD group, consistent with the grosser social breakdown and executive failure demonstrable in FTD. Previous investigations using the same cartoon and story materials have shown a disproportionate impairment on the mental compared to the physical conditions in patients following right hemisphere stroke [28] and in a patient following frontal lobe surgery [29]. The dissociation was interpreted as demonstrating impairments in theory of mind associated respectively with right hemisphere and frontal lobe lesions. It was predicted that FTD patients would show a similar disproportionate impairment for mental compared to physical items, whereas HD patients would show no such dissociation. The findings were not entirely in accordance with prediction. Both patient groups were essentially comparably impaired for the two types of test material, poorer scores for mental compared to physical material being demonstrated in FTD only in task 1. Analysis of error responses provides clues to the basis for the lack of dissociation, whilst also drawing attention to important qualitative differences between FTD and HD. In the FTD group, the largest proportion of errors for the cartoon tasks were omissions and concrete responses. Patients reported not knowing what was funny about cartoons and either failed to produce a response, or itemised elements without further explanation (e.g. “There’s a car, there’s a child”). Thus, not only did patients fail to go beyond the contents of the cartoon and draw inferences they also failed to integrate elements of the cartoon into a thematic narrative. Failures occurring at this relatively low-order level of cognition inevitably applied equally to mental and physical test material. These concrete-type responses contrasted strikingly with the errors of control subjects, which largely constituted partial responses: responses that were insufficiently detailed or inferences were implicit rather than explicitly stated. Controls typically did draw inferences, going beyond the literal elements of the cartoon. In the HD group, a large proportion of errors in the cartoon tasks were of the ‘description’ type: a full and integrated commentary on the contents of each cartoon was provided but without drawings inferences beyond the physical contents. Such errors occurred to some extent in all groups and were of no differentiating value. Of more theoretical interest is the presence of misconstruction errors. For a substantial proportion of items HD patients did draw inferences that went beyond the physical contents of the cartoon. They abstracted and formulated hypotheses, including hypotheses about a character’s feelings or belief. However, those inferences deviated from the conventional interpretation. For example, the usual interpretation of the ‘ping-pong’ cartoon described in Section 2.2.1 is that the man is deceiving the older woman into thinking he is playing table tennis. The response “They’re having a bit of nooky while the wife’s sat in there. She’s thinking ‘At least he’s leaving me alone. Peace!”’ suggests a diametrically opposite response: the older woman is not deceived. Such eccentric interpretations constituted a trademark of HD, in that they occurred

at least once in all but one of the HD patients, yet were virtually absent in other groups. Group differences cannot be explained in terms of notional differences in severity of impairment between FTD and HD. Misconstruction errors did not occur even in relatively high-functioning FTD patients, whose overall level of accuracy on the cartoon tests was comparable to that of HD subjects. They cannot, moreover, be attributed to scoring bias, because responses were evaluated blind to diagnosis. The tendency to misconstrue cut across stimulus type, being present both for mental and physical cartoon types. A similar pattern of errors occurred for the story task. FTD patients were more likely than other groups to give concrete responses, reiterating parts of the story without drawing inferences. By contrast, there was a trend for HD patients to make misconstruction errors, drawing faulty inferences from stories. Controls were more likely to give partially correct responses. Thus, the findings suggest a consistent pattern of performance regardless of the nature of the stimulus material. The cartoon and story tasks are both relatively demanding. Concrete responses in FTD patients might potentially have arisen due to task complexity: the requirement to integrate information and draw inferences might simply have imposed too great a mental executive demand on the patient. General executive deficits may have masked more specific deficits in mental state attribution. The face test is important in that it is undemanding, requiring no active mental manipulation or integration of information and it can be achieved by children as young as 3 years. Participants merely point to one of four pictures that the cartoon face prefers, preference being determined by direction of eye gaze. Nevertheless, FTD patients as a group showed an impaired ability to carry out the task, frequently disregarding eye gaze direction and basing their selection of preference on their own personal favourite. The fact that at least some patients had no difficulty determining which item the face was looking at suggests that failures on preference judgement could not be ascribed to executive deficits such as inability to attend to the test stimuli. Moreover, all patients complied with the task when asked for their personal preference suggesting that failures are unlikely to be secondary to comprehension impairment. Expression of personal preference might be construed as constituting a pre-potent response that the patient is no longer able to override or inhibit and is consistent with the impairments in response inhibition typical of frontal lobe dysfunction. However, the concrete mode of response, based on their own personal preference, is compatible also with the notion that FTD patients have an egocentric world-view in which they fail to recognise or attribute to others a mental state that differs from their own. Such an interpretation is consistent with relatives’ reports that FTD patients are oblivious to the feelings and needs of others. FTD patients, unlike children with autism [8], were not typically greatly influenced by the presence of an arrow, suggesting that external environmental stimuli were a less potent factor in guiding

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

responses than patients’ own internal mental state. However, there was an exception to this general rule. Responses in one FTD patient consistently corresponded to the arrow direction, highlighting heterogeneity within the FTD population. HD patients had no difficulty on the preference judgement task. This is consistent with earlier findings that they could draw inferences about another person’s emotions, thoughts or beliefs and with clinical observations that HD patients make pertinent and insightful remarks about the effects of their illness on a close relative. Nevertheless, insight demonstrated at a cognitive level, is frequently not matched by commensurately considerate, sympathetic or empathic behaviour in the patient’s daily life. There is a mismatch between what the patient says and does. HD patients show alterations in the processing of emotion [23,59] and one possibility is that HD patients’ lack of sympathy and empathy arises more at an emotional than a cognitive level. In any event, it cannot be attributed to an inability per se to attribute mental states. By contrast, FTD patients exist in an egocentric world, in which they do not ascribe independent mental states to others, a factor likely to contribute to their loss of capacity for sympathy and empathy. If HD patients can infer mental states in others but FTD patients cannot then this should be reflected in the number of mental state terms used in interpreting cartoons and stories. Consistent with prediction HD patients did not differ from controls with respect to the number of mental terms used. Conversely, FTD patients produced significantly fewer mental state terms, despite a comparable overall length of responses. This may partly reflect patients’ tendency to itemise elements, without integration into a coherent narrative. Indeed, on the cartoon pairs task FTD patients showed a reduction in the number of action verbs as well as mental state verbs. Nevertheless, a significant reduction in action verbs was demonstrated on a single task only, whereas a reduction in mental state verbs was a more pervasive finding. This disparity suggests that at least one contribution to FTD patients’ poor test performance is a failure to engage in mentalising and in the attribution of mental states to others. It is of relevance that a SPECT imaging study [10] demonstrated activation of orbitofrontal cortex in subjects required to judge whether words represented mental state terms. All the FTD patients in the present study had demonstrable involvement of orbitofrontal cortex on MR and SPECT imaging. A number of studies have demonstrated dissociations in performance on theory of mind and traditional frontal executive tasks [11,17,24,29,34,39,52], interpreted as evidence for the independence of theory of mind and executive skills. Nevertheless, it is reasonable to suppose that executive impairments will have a secondary impact on performance on theory of mind tasks, and some studies have found a relationship between performance on the two types of tasks [13,53]. In one study [13] patients with left anterior lesions, like FTD patients in the present study, failed to make non-literal interpretations, a finding ascribed to their tendency to attend only to the most salient aspect of the relevant information. The

699

authors argued that impaired executive function provided a sufficient explanation of the impaired story comprehension performance, without the need to invoke an additional theory of mind impairment. Nevertheless, they acknowledged the possibility that there may be two routes to impairment in theory of mind tasks, arising from disruption either to broader executive processes or to specific theory of mind ability. We would adopt such a view. The FTD group was not disproportionately impaired for social than for physical cartoons and stories compared to controls, and FTD patients showed worse performance for social compared to physical stimuli only on task 1, suggesting that general executive impairments contribute substantially to test performance. Indeed, poorer overall performance was generally seen in those patients with widespread frontal lobe atrophy. Nevertheless, the relationship between performance on social cognition and standard executive tests was relatively modest and not systematic across tasks. Moreover, on the preference judgement task some FTD patients failed to ascribe preference, yet had no difficulty reporting direction of eye gaze. The two tasks (“Which one does he like?” versus “Which one is he looking at?”) make comparable demands on attention and differ only with respect to the need for mental state attribution. Performance differences provide evidence for a specific impairment in theory of mind. We would argue that FTD patients have a genuine impairment in theory of mind, but that in many patients, in whom frontal lobe atrophy is severe and widespread this constitutes only one of a variety of deficits. General executive impairments will have an inevitable impact on performance on theory of mind tasks and may mask more specific deficits in theory of mind. Deficits in theory of mind independent of executive function might be expected early in the course of disease when pathological change is relatively confined to orbitofrontal regions [57]. Later in the disease course, with extension of pathology into dorsolateral regions, the picture will be increasingly coloured by additional executive deficits. Complementary findings and anatomical interpretation come from a study of the relationship between empathy, which requires the capacity to appreciate another’s thoughts and feelings, and cognitive flexibility, as measured by conventional executive tasks [22]. Differences between findings in the present study and another study of FTD patients [24], in which impairments were demonstrated on theory of mind but not control tasks, are likely to be attributed to differences in severity. The patients in the latter study had a substantially higher mean MMSE score than the present FTD patients (27 versus 22) and performed better on the WCST (4.4 versus 1.4 categories), suggesting that they were at an earlier stages in their illness. HD patients in the present study showed little convincing evidence of deficits in theory of mind. Nevertheless patients performed abnormally on tasks requiring interpretation of situations. Their tendency to make misconstruction errors has a resonance with their functioning in daily life. Relatives’ reports suggest that people with HD sometimes interpret

700

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701

events and social interactions in ways that do not accord with the norm. It is possible that their tendency to misconstrue may underlie or at least contribute to the fixed ideation and intransigence that are characteristic of some people with HD. The relatively weak correlation between performance on social cognition and standard executive tasks exemplifies the fact that the problems encountered by HD patients in open-ended tasks (and indeed normal social situations) may not be adequately reflected in standard executive tests. We have interpreted differences between FTD and HD as reflecting the fact that FTD is largely a frontal neocortical disorder and HD a disorder of the striatum. However, in disorders that affect frontostriatal circuitry a common assumption is that analogous deficits will arise regardless of the level of the circuit at which disruption occurs. Why then should FTD and HD be qualitatively different? If FTD and HD involve different striatofrontal circuits[3,43] then might this explain qualitative differences? FTD is thought to progress in an orbital-to-dorsolateral direction [57], whereas the dorsal to ventral striatal progression in HD [63] suggests the reverse. Nevertheless, differential involvement of circuits is an unlikely explanation. The classification of frontostriatal circuits [50] and the extent to which they are parallel [33] is itself not without controversy. Moreover, the FTD group included patients both with a relatively circumscribed orbitofrontal atrophy and with widespread frontal atrophy, presumably involving each of the dorsolateral, orbitofrontal and anterior cingulate loops, yet none showed an HD-like pattern of error response. Furthermore, poor performance on executive tasks has characteristically been associated with dorsolateral frontal lobe pathology [43], yet it was the FTD patients who performed the more poorly on these tasks suggesting dorsolateral frontal dysfunction at least as great as in HD. In FTD frontal cortical grey and white matter are comparably affected [41], whereas in HD there is imaging [7] and pathological [40] evidence of a disproportionate involvement of white matter. In FTD maximal loss of neurones occurs in the superficial layers II and III, resulting primarily in loss of cortico-cortical connections, whereas in HD pyramidal neurones in deeper layers V and VI, which subserve cortico-subcortical projection fibres, are most involved. We would suggest that the key distinction underlying performance differences is that FTD involves primarily frontal neocortex and its cortico-cortical afferents whereas HD is a disorder of the striatum and its cortico-subcortical connections. The study highlights the value of open-ended tasks involving interpretation of social situations, in exploring deficits arising from disorders of the frontostriatal system. The findings suggest that, despite superficial similarities in the pattern of cognitive disorder and altered social conduct in FTD and HD, qualitative differences exist in the nature of underlying deficits. In FTD there is a loss of theory of mind but that additional executive deficits colour patients’ performance on theory of mind tasks. In HD there is no convincing evidence of a loss of theory of mind. Future studies

need to address the qualitative characteristics of performance in laboratory-based and real-life social situations to clarify more precisely the basis for social breakdown in FTD and HD. References [1] Adolphs R. Social cognition and the human brain. Trends in Cognitive Sciences 1999;3:469–79. [2] Adolphs R. The neurobiology of social cognition. Current Opinion in Neurobiology 2001;11:231–9. [3] Alexander GE, De Long MR, Strick PL. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annual Review of Neuroscience 1986;9:357–81. [4] Antonini A, Leenders KL, Spiegel R, Meier D, Vontobel P, Weigell-Weber M, et al. Striatal glucose metabolism and dopamine D2 receptor binding in asymptomatic gene carriers and patients with Huntington’s disease. Brain 1996;119:2085–95. [5] Aylward EH, Brandt J, Codori AM, Mangus RS, Barta PE, Harris GJ. Reduced basal ganglia volume associated with the gene for Huntington’s disease in asymptomatic at-risk persons. Neurology 1994;44:823–8. [6] Aylward EH, Li Q, Stine OC, Ranen N, Sherr M, Barta PE, et al. Longitudinal change in basal ganglia volume in patients with Huntington’s disease. Neurology 1997;48:394–9. [7] Aylward EH, Anderson NB, Bylsma FW, Wagster MV, Barta PE, Sherr M, et al. Frontal lobe volume in patients with Huntington’s disease. Neurology 1998;50:252–8. [8] Baron-Cohen S, Campbell R, Karmiloff-Smith A, Grant J, Walker J. Are children with autism blind to the mentalistic significance of the eyes? British Journal of Developmental Psychology 1995;13:379–98. [9] Baron-Cohen S, Leslie AM, Frith U, Baron-Cohen S. ‘Does the autistic child have a theory of mind’? Cognition 1985;21:37–46. [10] Baron-Cohen S, Ring H, Moriarty J, Schmitz B, Costa D, Ell P. Recognition of mental state terms: clinical findings in children with autism and a functional neuroimaging study of normal adults. British Journal of Psychiatry 1994;165:640–9. [11] Baron-Cohen S, Robertson MM. Children with either Autism, Gilles de la Tourette Syndrome or both: mapping cognition to specific syndromes. Neurocase 1995;1:101–4. [12] Bathgate D, Snowden JS, Varma A, Blackshaw A, Neary D. Behaviour in frontotemporal dementia, Alzheimer’s disease, and vascular dementia. Acta Neurologica Scandinavica 2001;103:367–78. [13] Channon S, Crawford S. The effects of anterior lesions on performance on a story comprehension test: left anterior impairment on a theory of mind-type task. Neuropsychologia 2000;38:1006–17. [14] Craufurd D, Snowden JS. Neuropsychological and neuropsychiatric aspects of Huntington’s disease. In: Harper P, Bates G, Jones L, editors. Huntington’s disease. 3rd ed. 2002. p. 62–94, in press. [15] Damasio AR, Tranel D, Damasio H. Individuals with sociopathic behavior caused by frontal damage fail to respond autonomically to social stimuli. Behavioural Brain Research 1990;41:81–94. [16] Eslinger PJ, Damasio AR. Severe disturbance of higher cognition after bilateral frontal lobe ablation: Patient EVR. Neurology 1985;35:1731–41. [17] Fine C, Lumsden J, Blair JR. Dissociation between ‘theory of mind’ and executive functions in a patient with early left amygdala damage. Brain 2001;124:287–98. [18] Fletcher PC, Happé F, Frith U, Baker SC, Dolan RJ, Frakowiak RS, et al. Other minds in the brain: a functional imaging study of “theory of mind” in story comprehension. Cognition 1995;57:109–28. [19] Folstein MF, Folstein SE, McHugh PR. Mini-mental state. Journal of Psychiatric Research 1975;12:189–98. [20] Gallagher HL, Happé F, Brunswick N, Fletcher PC, Frith PC, Frith CD. Reading the mind in cartoons and stories: an fMRI study of ‘theory of mind’ in verbal and nonverbal tasks. Neuropsychologia 2000;38:11–21.

J.S. Snowden et al. / Neuropsychologia 41 (2003) 688–701 [21] Goel V, Grafman J, Sadato N, Hallett M. Modeling other minds. Neuroreport 1995;6:1741–6. [22] Grattan LM, Bloomer RH, Archambault FX, Eslinger PJ. Cognitive flexibility and empathy after frontal lobe lesion. Neuropsychiatry Neuropsychology and Behavioral Neurology 1994;7:251–9. [23] Gray JM, Young AW, Barker WA, Curtis A, Gibson D. Impaired recognition of disgust in Huntington’s disease gene carriers. Brain 1997;120:2029–38. [24] Gregory C, Lough S, Stone V, Erzinclioglu S, Martin L, Baron-Cohen S, et al. Theory of mind in patients with frontal variant frontotemporal dementia and Alzheimer’s disease: theoretical and practical implications. Brain 2002;125:752–64. [25] Gustafson L. Frontal lobe degeneration of non Alzheimer type II. Clinical picture and differential diagnosis. Archives of Gerontology and Geriatrics 1987;6:209–23. [26] Happé FGE. An advanced test of theory of mind: understanding of story characters’ thoughts and feelings by able autistic, mentally handicapped, and normal children and adults. Journal of autism and developmental disorders 1994;24:129–54. [27] Happé FGE. ‘Theory of mind’ in the brain. Evidence from aPET scan study of Asperger syndrome. Neuroreport 1996;8:197–201. [28] Happé FGE, Brownell H, Winner E. Acquired ‘theory of mind’ impairments following stroke. Cognition 1999;70:211–40. [29] Happé FGE, Malhi GS, Checkley S. Acquired mind-blindness following frontal lobe surgery? A single case study of impaired ‘theory of mind’ in a patient treated with stereotactic anterior capsulotomy. Neuropsychologia 2001;39:83–90. [30] Harper, P.S., Huntington’s disease. London: Saunders; 1996. [31] Huntington’s Disease Study Group. Unified Huntington’s Disease Rating Scale: reliability and consistency, movement disorders 1996;11:136–42. [32] Jacobs DH, Shuren J, Heilman KM. Impaired perception of facial identity and facial affect in Huntington’s disease. Neurology 1995;45:1217–8. [33] Joel D, Weiner I. The connections of the primate subthalamic nucleus: indirect pathways and the open-interconnected scheme of basal ganglia-thalamocortical circuitry. Brain Research Reviews 1997;23:62–78. [34] Karmiloff-Smith A, Klima E, Bellugi U, Grant J, Baron-Cohen S. Is there a social module? Language, face processing and theory of mind in individuals with Williams syndrome. Journal of Cognitive Neuroscience 1995;7:196–208. [35] Lavenu I, Pasquier F, Lebert F, Petit H, Van der Linden M. Perception of emotion in frontotemporal dementia and Alzheimer’s disease. Alzheimer’s Disease and Associated Disorders 1999;13:96–101. [36] Leslie AM. Pretense and representation: the origins of ‘theory of mind’. Psychological Review 1987;94:412–26. [37] Levy ML, Miller BL, Cummings JL, Fairbanks LA, Craig A. Alzheimer’s disease and frontotemporal dementias. Archives of Neurology 1996;53:687–90. [38] Loh EA, Roberts JKA, Mohr E. Structural correlates of neurological signs in Huntington’s disease. Behavioural Neurology 1994;7: 127–34. [39] Lough S, Gregory C, Hodges JR. Dissociation of social cognition and executive function in frontal variant frontotemporal dementia. Neurocase 2001;7:123–30. [40] Mann DMA, Oliver R, Snowden JS. The topographic distribution of brain atrophy in Huntington’s disease and progressive supranuclear palsy. Acta Neuropathologica 1993;85:553–9. [41] Mann DMA, South PW. The topographic distribution of brain atrophy in frontal lobe dementia. Acta Neuropathologica 1993;85: 334–40. [42] Marsden CD. The mysterious motor function of the basis ganglia: the Robert Wartenberg lecture. Neurology 1982;32:514–39.

701

[43] Mega MS, Cummings JL. Frontal-subcortical circuits and neuropsychiatric disorders. Journal of Neuropsychiatry and Clinical Neurosciences 1994;6:358–70. [44] Miller BL, Cummings JL, Villanueva-Meyer J, Boone K, Mehringer CM, Lesser IM, et al. Frontal lobe degeneration: clinical, neuropsychological, and SPECT characteristics. Neurology 1991;41:1374–82. [45] Miller BL, Darby A, Benson DF, Cummings JL, Miller MH. Aggressive, socially disruptive and antisocial behaviour associated with frontotemporal dementia. The British Journal of Psychiatry 1997;170:150–5. [46] Neary D, Snowden JS, Gustafson L, Passant U, Stuss D, Black S, Freedman M, Kertesz A, et al. Frontotemporal lobar degeneration. A consensus on clinical diagnostic criteria. Neurology 1998;51: 1546–54. [47] Neary D, Snowden JS, Northen B, Goulding P. Dementia of frontal lobe type. Journal of Neurology, Neurosurgery, and Psychiatry 1988;51:353–61. [48] Neary D, Snowden JS, Shields RA, Burjan AWI, Northen B, Macdermott N, et al. Single photon emission tomography using 99mTc-HMPAO in the investigation of dementia. Journal of Neurology, Neurosurgery and Psychiatry 1987;50:1101–9. [49] Nelson HE. A modified card sorting test sensitive to frontal lobe defects. Cortex 1976;12:313–24. [50] Parent A, Hazrati LN. Functional anatomy of the basal ganglia 1. The cortico-basal ganglia-thalamo-cortical loop. Brain Research Reviews 1995;20:91–127. [51] Premack D, Woodruff G. Does the chimpanzee have a theory of mind? Behavioural and Brain Sciences 1978;1:515–26. [52] Rowe AD, Bullock PR, Polkey CE, Morris RG. ‘Theory of mind’ impairments and their relationship to executive functioning following frontal lobe excisions. Brain 2001;124:600–16. [53] Saltzman J, Strauss E, Hunter M, Archibald S. Theory of mind and executive functions in normal human aging and Parkinson’s disease. Journal of the International Neuropsychological Society 2000;6: 781–8. [54] Saver JL, Damasio AR. Preserved access and processing of social knowledge in a patient with acquired sociopathy due to ventromedial frontal damage. Neuropsychologia 1991;29:1241–9. [55] Shoulson I. Huntington’s disease: functional capacities in patients treated with neuroleptic and antidepressant drugs. Neurology 1981;31:1333–5. [56] Snowden JS, Bathgate D, Varma A, Blackshaw A, Gibbons Z, Neary D. Distinct behavioural profiles in frontotemporal dementia and semantic dementia. Journal of Neurology, Neurosurgery, and Psychiatry 2001;70:323–32. [57] Snowden JS, Neary D, Mann DMA. Frontotemporal lobar degeneration: frontotemporal dementia, progressive aphasia, semantic dementia. London: Churchill-Livingstone; 1996. [58] Spreen O, Strauss E. A compendium of neuropsychological tests. New York: Oxford University press; 1991. [59] Sprengelmeyer R, Young AW, Calder AJ, Karnat A, Lange H, Hömberg V, et al. Loss of disgust. Brain 1996;119:1647–65. [60] Stone VE, Baron-Cohen S, Knight RT. Frontal lobe contributions to theory of mind. Journal of Cognitive Neuroscience 1998;10:640–56. [61] Stuss DT, Gallup GG, Alexander MP. The frontal lobes are necessary for ‘theory of mind’. Brain 2001;124:279–86. [62] Talbot PR, Lloyd JJ, Snowden JS, Neary D, Testa HJ. A clinical role for 99mTc-HMPAO SPECT in the investigation of dementia? Journal of Neurology Neurosurgery and Psychiatry 1998;64:306–13. [63] Vonsattel JP, Myers RH, Stevens TJ, Ferrante RJ, Bird ED, Richardson EP. Neuropathological classification of Huntington’s disease. Journal of Neuropathology and Experimental Neurology 1985;44:559–77.