Delayed response tasks and prefrontal lesions in man—Evidence for self generated patterns of behaviour with poor environmental modulation

Neuropsychologia, Printed in Great

Vol. 31, No. 12, pp. 1379-1396. Britain.

00X-3932/93 56.00+0.00 Q 1993 Pergamon Press Ltd

1993.

DELAYED RESPONSE TASKS AND PREFRONTAL LESIONS IN MAN-EVIDENCE FOR SELF GENERATED PATTERNS OF BEHAVIOUR WITH POOR ENVIRONMENTAL MODULATION M. VERIN,A. PARTIOT, B. PILLON, C. MALAPANI, Y. AGID and B. DUBOIS* INSERM U 289 & Service de Neurologie

et Neuropsychologie (Pr. Chain), 47, Bd de I’Hbpital, 75651 Paris Cedex 13, France

Hbpital

de la Sal$triire,

(Received 24 June 1992; accepted 4 May 1993) Abstract-The functions of the frontal lobes in humans are still under debate, mainly because none of the neuropsychological tests used for their assessment is sufficiently specific for frontal dysfunction. In animals, the delayed reaction paradigm is considered to be a specific marker of the function of dorsolateral region of the prefrontal cortex. It seemed of interest, therefore, to attempt to apply this paradigm to patients with recent and limited cortical lesion of vascular origin. The performance of patients with dorsolateral prefrontal lesion (n = 10) was compared to that ofpatients with post-central lesion (n = 10) and control subjects (n = 24), in four experiments: a Delayed Response task in which the correct answer was previously indicated by an explicit cue (externally guided task); Delayed Alternation and Non-Alternation tasks coupled with a Delayed Reversal task in which the patient had to discover the rule by himself in the absence of explicit cues (internally driven tasks). Patients with prefrontal lesion showed a specific deficit in the Delayed Response task, the emergence of a stereotyped behaviour in the Delayed Alternation task and an inability to deduce and to transfer rules (non-alternation and reversal), mainly because of difficulty in abandoning previous behaviours. Our study demonstrates that the prefrontal cortex plays a role in behavioural adaptation to challenging new situations by inhibiting not only ongoing elaborated programmes but also the emergence of previously established automatic programmes. The respective role of the prefrontal cortex and the basal ganglia in these two levels of behavioural organization is discussed.

INTRODUCTION EXPERIMENTALSTUDIES have shown that Delayed Alternation and Delayed Response tasks bring into play the dorsolateral prefrontal cortex. This has been demonstrated by a variety of experimental techniques (uni- or bilateral surgical lesions, uni- or bilateral cooling, stimulation, monocellular recording or metabolic imagery) in several animal species [22,27, 521. While problematic, the use of Delayed Reaction paradigms in humans may permit access to some processes considered to be specific to frontal lobe activity: (1) the faculty to maintain the internal representation of a visuo-spatial stimulus during a delay (Delayed Response task); (2) the capacity to elaborate a concept or a rule (Alternation task); (3) the capacity to shift set (Reversal task). Furthermore, in the absence of neuropsychological tests which are unquestionably specific for frontal lobe function, Delayed Reaction tasks may help to detect frontal lobe disorders in patients. Results obtained with patients with frontal lesions have been contradictory, however. In some studies, performance of Delayed Reaction tasks was reported to be altered: PRIBRAM et al. [Sl] found a deficit in Delayed Alternation in five

*To whom all correspondence

should

be addressed. 1379

M. VERIN et al.

1380

schizophrenic patients with bilateral frontal lobotomy compared with schizophrenics who had not undergone the operation; FREEDMAN and OSCAR-BERMAN [17] reported a deficit in both Delayed Response and Delayed Alternation tasks in six patients with bilateral frontal lesions. On the other hand, CHQROVER and COLE [lo] found no difference between the performance on Delayed Alternation task of 15 patients with frontal lesions and 18 patients with post-central lesions. A similar result was reported by CANAVAN et al. [9] who, in addition, found that the age of the patients accounted for more of the variability in performance on the Delayed Alternation task than did the site of the lesions. The difference in these results may be explained by the diversity of patients studied, and the delay between the occurrence of the lesions and the test session. Patients studied by PRIBRAM et al. [Sl] suffered from schizophrenia, known to alter performance on tasks considered to be sensitive to frontal dysfunction [64]. The effects of lobotomy in these patients were probably not comparable to those of a frontal lesion in an initially healthy brain. Subsequent studies concerned patients presenting heterogeneous (tumoral, traumatic or infectious) and poorly defined lesions (as to their size and locations) which may have modified the function of distant brain structures directly or indirectly. For these reasons, the specificity of Delayed Reaction paradigms for frontal functions in man remains to be demonstrated. We have therefore studied this paradigm in carefully selected patients with isolated unilateral ischemic lesion of the cerebral cortex.

SUBJECTS

AND METHODS

Subjects Twenty-four control subjects with no history ofneurologic or psychiatric disorders, 10 patients with focal lesion of the dorsolateral prefrontal cortex, and 10 patients with focal lesion of the post-central cortex were studied. The three groups were matched for age and level of education (Table 1). In all patients, the lesion was cortical, isolated,

Table

1. Clinical

and neuropsychological

characteristics subjects Control (n=24)

Age (years) Education level (years) Sex Ratio F/M Laterality R/L Side of the lesion R/L Age of the lesion (weeks) MMSE WCST Graphic series” Frontal score b

59.7+ 12.6 10.4* 2.4 1519

24/O 29.3 k 0.7 620 9.5k3.4 57.9+2

of patients

with focal lesion and control

Post-central (n= 10) 62.6+ 14.3 9.4fl.9 317 1o/o 515 2.5 + 1 28.5 + 2.4 (n.s.) 5.5+ 1.2 8.5k3.7 53+6.5

Results are expressed as mean k S.D. MMSE: Mini Mental Status Examination. WCST: Wisconsin Card Sorting Test (number of categories). n.s.: not significant. * P
lesion patients. lesion patients.

Prefrontal (n=lO) 6Ok 12.2 9.9k2.5 317 1010 317 6.7kJ.l 26.6 + 3.8 (ns.) 1.9+ 1.2* 5*4.1* 33.3 + 8.9*

DELAYED RESPONSE TASKSAND

PREFRONTAL LESlONSIN MAN

1381

unilateral, ischemic, and occurred less than 3 months before the experiment. It resulted from an arterial embolism in 8 cases of the prefrontal group and 10 cases of the post-central group, and from a cardiac embolism in 2 cases of the prefrontal group. The lesions. documented by CT scan, were reconstructed according to the atlases of MATSUI and HIRANO [42] and TALAIRACH and SZIKLA [61] and the method proposed by EBELING et al.[IS].The site of each focal lesion is shown in Figs l(A) and l(B) (prefrontal and post-central, respectively). General procedure The subjects were seated in an armchair 50cm away from a microcomputer screen (25 x 19 cm), in front ofa box with two symmetrically placed buttons (one on the right side, the other on the left side). Two identical blue squares (6 x 6 cm) arranged horizontally, 10 cm apart, appeared at the top of the screen, and a small schematized cannon with a 2.5-cm long barrel and a 2-cm wide base in the lower center. The examiner was seated in such a fashion that he could observe the subject without being seen. The examiner noted, without interfering, the subject’s behaviour and any comments he might have made. In each of the tasks, the two squares disappeared for 15 set (the delay). After this delay, the two stimuli reappeared on the screen for the subject to respond. A beep alerted the subject that he must fire at the square ofhis choice (the response) by pushing the corresponding button on the box (right button fires right, left button fires left). The subjects were asked to respond with their dominant hand and to leave this hand in a neutral position in front of the box during the delay. When the subject pushed the button, a circular projectile measuring 3 mm in diameter was shot from the cannon to the square on the chosen side. If the response was correct, the square disappeared and the word “WIN” appeared in its place for a period of 2 set (the reinforcement). Then, the square reappeared. If the choice was incorrect,the square remained visible and there was no reinforcement. The study was divided in two different parts. In the first part of the experiment, which was internally driven, there was no external guidance: the subject had to discover the rule by deductive reasoning, guided by the distribution of reinforcements. There were two different rules to be discovered: an alternation rule (Delayed Alternation task) and a non-alternation rule (Delayed Non.4lternation task), coupled with a reversal (Delayed Reversal task): (1) In the Delayed Alternation task, the subject had to learn to alternate his response (i.e. win-shift strategy), that is to choose the square opposite to the one he chose previously. The first trial was systematically reinforced, no matter which square was chosen by the subject. A correction procedure was used: ifthe good square was not chosen, it remained the correct choice until it was chosen. (2) In the Delayed Non-Alternation task, the subject had to change his strategy and learn not to alternate his response (i.e. win-stay strategy), that is to always choose the same square, which corresponded to the last correct choice of the Alternation task. (3) In the Delayed Reversal task, the subject had to choose the side opposite to the correct response on the Delayed Non-Alternation task. that is to reverse the rule of non-alternation to the opposite side, and to continue with this choice. The oral instructions given at the beginning of the session were: “Infront ofyou are two squares and a small cannon. The two squares are two targets. At each beep, you mustfire at one of the two squares with the help ofone of these two buttons. ryoujire at the correct square, it will disappear and the word “ WIN” willappear on the screen. Ifyoujre at the wrong square, it will not disappear and you will have lost. There is a 15 seconds delay between each trial. It is possible to win each time but it is up to you tofigure out the rule ofthe yame. It is therefore by trying and perhups by making mistakes that you willjnd the rule. Please note, you must win eoery time”. These instructions were repeated until it appeared that the subject fully understood them. The three tasks (Delayed Alternation, Delayed Non-Alternation coupled with Delayed Reversal) were executed one after the others, rule changing automatically, without the subject being informed, as soon as the criterion of 10 consecutive successful trials was met or, failing that, after 80 trials. The score for each subject was recorded automatically by the computer throughout the session. The second part of theexperiment was externally guided. In this situation, the choice between the two squares was indicated by an explicit cue (an arrow), which appeared on the screen before the delay: this was the Delayed Response task (Fig. 2). The arrow, measuring 2 x 1.5 cm, appeared for 2 set under one of the squares (2A) and was selected pseudo-randomly by the computer so that it appeared under one square about as often as under the other. The subject had to remember the position of the arrow during the delay (2B) and, at the signal, fire at the correct square (2C). The oral instructions given at the beginning of the experiment were the same as in the previous experiment, except for the following: “an arrow will appearfor 2 seconds under one ofthe squares. It indicates at which square you should fire when you hear the beep, that is to say after a 15 seconds delay”. Additional neuropsychological tests (Tab/e 1) Patients underwent an additional battery of tests which included Folstein’s Mini Mental Status ExaminationMMSE [16] and tests known to be sensitive to frontal lobe dysfunction: Nelson’s revised version of the Wisconsin Card Sorting Test-WCST [47], a verbal fluency test [6], a graphic series test [40]. Behavioural abnormalities (prehension, imitation, utilization behaviours; inertia, indifference) characteristic of patients with frontal lobe lesions [33] were also assessed and scored each from 4 for normal to 0 for very pathological. A “frontal” score [14] was defined as the sum of the scores on each cognitive and behavioural test believed to lx sensitive to frontal dysfunction (WCST, verbal fluency test, graphic series test and behavioural abnormalities).

1382

M. VERIN et al

(A)

uv

case I

case6

case7

case 3

case8

case4

case9

case5

case 10 Fig. 1A

Data analysis Performance.

(I) For individual subjects, two scores were defined for each of the tasks: the number errors produced before reaching the criterion of 10 consecutive successful trials; the percentage of errors with respect to the number oftrials. The first error made in Delayed Non-Alternation and Reversal tasks was not taken into account, as this error was the expected result of the acquisition of the previous rule. (2) In the Delayed Alternation and Non-Alternation tasks, the ability to generate a rule (rule acquisition ability) was quantified by calculating at each successive trial the percentage ofsubjects who started a successful series of 10 consecutive trials, i.e. subjects who understood the rule. (3) In the Delayed Reversal task, only the score of subjects who had reached the criterion on Delayed NonAlternation task was taken into account, as the aim of this task was to reverse the already acquired non-alternation rule. Statistical analysis. The scores for the different tasks were compared using first the Kruskal-Wallis nonparametric one-way analysis for the three independent groups. If there was a significant group effect, the paired comparisons were made using the non-parametric Mann-Whitney U test for two independent groups because of large variance within the groups. As the number of subjects was small, the percentage of subjects meeting the criterion in the Delayed Alternation and Non-Alternation tasks were compared using the bilateral version of the Fisher x2 test. Correlations were determined by the Spearman rank correlation coefficient.

DELAYED

RESPONSE TASKS AND

PREFRONTAL

LESIONS IN MAN

1383

Fig. 1B. Fig. 1. Representative diagrams of focal lesions based on CT scans, including an external view of the brain and an upper view of a transverse section following the xy axis. Black areas indicate damaged regions. (A) Prefrontal lesions. (B) Post-central lesions.

RESULTS Delayed Alternation task (Table 2, Fig. 3) The Kruskal-Wallis analysis of variance for the three independent groups indicates a significant group effect (H= 14; P
1384

M. VERIN et al.

Fig. 2. Sequence of one trial of the Delayed Response task. (A) External clue (arrow) indicating the reinforced square; (B) indicates the 15 set delay; (C) correct response with disappearance of the chosen square and reinforcement. If uncorrect response (here right), the chosen square does not disappear, there is no reinforcement.

Table 2. Number

of errors

in Delayed

Reaction Control (n = 24)

Delayed

Alternation

Delayed

Non-Alternation

Delayed

Reversal

Delayed

Response

3.lk4.1 (2) 4.3k3.5 (3) l&2.3 (0) Ok0 (0)

tasks: mean

+S.D.

and median

Post-central (n=lO) 10.5* 18.4 (2.5) 3.9k2.5” (3) 2.4 f 2.9” (1.5) 0.4* (0)

1*

(in brackets) Prefrontal (n=lO)

0.1*0.3***tt (0) 24.5 -+ 17 .6**t

(22) 8.5+ 15.5*b 2.5) 2.8+3.2***7 (1.5)

* PcO.05 when compared with control subjects. ** PiO.01 when compared with control subjects. *** P
was 10.5+ 18.4 (20.3%), which was not significantly different from the control group. The ability to find the rule of alternation was also similar to that of the control subjects: 2 of the 10 patients met the criterion directly; half met it by the 10th trial, but 2 of the 10 did not meet the criterion within the 80 trials limit (Fig. 4). In the prefrontal group, only one patient made an error (case 7). The mean number of errors was 0.1 f0.3 (0.8%), which was significantly lower than that of both the control (U=30; P~O.001) and the post-central (U=13; P~0.01) groups. The criterion of

DELAYED

RESPONSE

TASKS AND PREFRONTAL

DELAYED

LESIONS IN MAN

1385

ALTERNATION

60

60

40

30

20

PC

PF

Fig. 3. Number of errors on the Delayed Alternation task. The histograms give the mean values and the bars the values for each individual. (C)control subjects; (PF) patients with prefrontal lesion; (PC) patients with post-central lesion.

alternation was reached significantly faster by prefrontal patients than by both control subjects and patients with post-central lesion (Fig. 4): 9 of the 10 prefrontal patients met the criterion directly on the 1st trial (P~O.001 and P-c 0.01 compared with the control and postcentral groups, respectively), and by the 3rd trial, all of the prefrontal patients had acquired the rule (P-cO.002 and P-co.02 with the control and post-central groups, respectively). Delayed Non-Alternation-Reversal

tasks (Table 2)

Delayed Non-Alternation task (Fig. 5). The Kruskal-Wallis analysis of variance for the three independent groups indicates a significant group effect (If= 12; P
M. VERIN et al.

1386

DELAYED

60

-

ALTERNATION

-

-

CONTROL

I.--“....-

POST-CENTRAL

-___-

PREFRONTAL

I

0

10

30

20

40

50

trials

Fig. 4. Proportion, trial by trial, of control subjects and patients with focal lesion who have completed 10 consecutive successful trials in the Delayed Alternation task. Continuous line represents the control group, dotted line the post-central group and dashes the prefrontal group.

lesion: none of the prefrontal patients reached the criterion directly; half of the patients reached it by the 38th trial and four ofthe 10 (cases 2,3,5 and 9) failed to meet it within the 80 trials limit (Fig. 6). Delayed

Reversal

task (Fig. 7)

The Kruskal-Wallis analysis of variance for the three independent groups indicates a significant group effect (H= 6; P-c 0.05). Seven of the 24 control subjects made errors. The mean number of errors was 1 f2.3 (4.1%). Seventy-one per cent of the subjects (17 of the 24) reached the criterion directly and all subjects had reached it by the 22nd trial (Fig. 8). In the post-central group, 5 of the 8 patients who had acquired the rule of NonAlternation, made errors. The mean number of errors, 2.4k2.9 (8.5%), was not different from that of the control group. The ability of this group to perform the reversal was not different from that of the control subjects: half of the 8 reached the criterion directly, and all had reached it by the 32nd trial (Fig. 8). In the prefrontal group, 5 of the 6 patients who had acquired the rule of Non-Alternation made errors. The mean number of errors was 8.5+ 15.5 (18.7%). The difference was significant when compared to the control group (U= 30; P~0.02). By the 1st trial, only one of the 6 patients (case 7) reached the criterion directly, and by the 80th trial, 1 of the 6 (case 6) had not yet reached the criterion (Fig. 8).

DELAYEDRESPON~ETASKSAND

DELAYED

~REFRONTALLESIONSINMAN

1387

NON-ALTERNATION

60

50

40

30

20

10

0

C

PC

Fig. 5. Number oferrors on the Delayed Non-Alternation task. The histograms give the mean values and the bars the values for each individual. (C)control subjects; (PF) patients with prefrontal lesion; (PC) patients with post-central lesion.

Delayed

Response

task (Table 2, Fig. 9)

The Kruskal-Wallis analysis of variance for the three independent groups indicates a significant group effect (H=21; P
The performance of the control subjects and of the patients was not correlated with either age or education in any of the tasks. There were no significant differences related to sex or the side of the lesion. Correlations between Delayed Reaction tasks and neuropsychological tests performance. In the Delayed Response task, the number of errors of the prefrontal group was correlated with the score on the graphic series (r=0.89; Pt0.05) and with the frontal score (r =0.89; P
1388

M. VERIN et al. DELAYED

NON-ALTERNATION

-

CONTROL

----I

POST-CENTRAL

_____

PREFRONTAL

r----------, I : :

0

! ” .

0

I

10

20

I

I

I

30

40

50

1

60

trials Fig. 6. Proportion, trial by trial, of control subjects and patients with focal lesion who have completed IO consecutive successful trials in the Delayed Non-Alternation task. Continuous line represents the control group, dotted line the post-central group, and dashed the prefrontal group.

DISCUSSION In the absence of unequivocal tests ofhuman frontal functions, we have attempted to apply the methodology of the Delayed Reaction task, which has been demonstrated to evidence prefrontal dysfunction in animals, and has produced some controversial [lo] results in human studies. In order to reevaluate this paradigm, we have adapted the method used with monkeys by JACOBSEN [27], and selected patients with unique, vascular, ischemic lesion, allowing us to compare more homogeneous patient groups than in previous studies. Delayed Alternation task The control subjects easily deduced the rule from the distribution of reinforcement obtained by trial and error. Only 21% of the subjects found the rule by chance, reaching the criterion directly, and no parameter (age, level ofeducation or performance on the additional neuropsychological tests) distinguished them from the others. The patients with post-central lesion performed equally well. In contrast, the behaviour of the patients with prefrontal lesion was surprising since they made significantly less errors than the two other groups. All but one alternated spontaneously. Thus their performance was apparently better than that of the control subjects. This result was unexpected for at least three reasons. Firstly, previous studies of Delayed Reaction tasks have always shown impaired rule acquisition, and especially alternation, in humans [lo, 17,511 and experimental animals [20,46] with frontal lobe lesions. Secondly, the success of these patients in the Delayed Alternation task implies

DELAYED

RESPONSE

TASKS

AND PREFRONTAL

DELAYED

LESIONS

IN MAN

I389

REVERSAL

40

30

: 0

20

I. L

0

10

C

PC

PF

Fig. 7. Number of errors on the Delayed Reversal task. The histograms bars the values for each individual. (C) control subjects; (PF) patients patients with post-central lesion.

give the mean values and the with prefrontal lesion; (PC)

that they have no difficulties in shifting their responses at each trial, whereas an impaired ability to shift mental set is considered to be characteristic of frontal dysfunction [ f1,51,59, 631, and has been proposed to explain the poor performance of animal with prefrontal lesion on the Delayed Alternation task [45]. Finally, the apparent success of patients with prefrontal lesion on the Delayed Alternation task implies that they can maintain their response during the delay, in contradiction with their poor Delayed Response task performance. It can be argued that this paradoxical observation may be due to a non-specific effect at the early period of an ischemic episode. In this hypothesis, however, this behavioural feature would be observed, as well, in the post-central group since the mean age of the lesion in this group is as earlier as that of the prefrontal one. This is not the case, suggesting that the prefrontal behaviour demonstrated in this task is not simply due to a non specific effect of wide-spread disruption. Analysis of the performance of the prefrontal patients may give some clue to their behaviour. All but one expressed the rule spontaneously at the first trial and the last one found it by the 3rd trial while some control subjects did not get it until the 46th trial. This suggests that they did not learn alternation by a deductive reasoning, but simply expressed it spontaneously. This spontaneous tendency to alternate was even irrepressible for some patients: two patients (cases 2 and 3) continued to alternate during the Delayed NonAlternation and Reversal tasks, although the distribution of reinforcements had changed; and one patient (case 6), resumed to alternation even after the acquisition of non-alternation rule, and persisted during the Delayed Reversal task. The persistence of alternation despite changes in environmental clues underlines the

M. VERIN et al.

1390

DELAYED

REVERSAL

60

01 0

5

-

CONTROL

-.-----.

POST-CENTRAL

--___

PREFRONTAL

10

15

20

25

trials Fig. 8. Proportion, trial by trial, of control subjects and patients with focal lesion who have completed 10 consecutive successful trials in the Delayed Reversal task. Continuous line represents the control group, dotted line the post-central group, and dashes the prefrontal group.

strength of this spontaneous behaviour in patients with prefrontal lesion. The tendency to alternate seems unlikely to result from random responses, as nine ofthe 10 patients succeeded directly, or from an abnormal tendency to shift [28,51], given the patients’poor performance on the Delayed Reversal task. It seems rather that spontaneous tendency to alternate was imposed on the patients by a process independently of the reinforcements which, in contrast, guided the performance of the control subjects. The pattern of alternation, induced by environmental triggers, seems to correspond to an implicit and automatic programme which corresponded, by chance, to the rule the subjects had to find, thereby accounting for their apparent good performance. If one admits that the learning process needed to acquire a new rule consists of many phases--data analysis, establishment of a programme, execution and control of this programme, final comparison-which is repeated until the rule is acquired [32,36-381, the performance of the patients with prefrontal lesion indicates that they do not make use of this process since they alternate directly. For this reason, and despite their apparent success on the test, the emergence of a stereotyped programme cannot be considered as a behavioural gain, since it overshadows the heuristic processes leading to environmental adaptation exhibited by both the control subjects and the patients with postcentral lesion. To our knowledge, this is the first time that the emergence ofan uncontrolled stereotyped behaviour in patients with frontal lesions is documented in an experimental study. Luria had clinically observed such automatisms that he described as “the inability to suppress

DELAYED

RESPONSE

TASKS AND PREFRONTAL

LESIONS IN MAN

1391

DELAYED RESPONSE

I PF

Fig. 9. Number of errors on the Delayed Response task. The histograms bars the values for each individual. (C) control subjects; (PF) patients patients with post-central lesion.

give the mean values and the with prefrontal lesion; (PC)

involuntarily emerging inert stereotypes” [39]. For FUSTER [19], these automatisms correspond to routine activities for which the behavioural programme is well established and does not require an adaptative effort: “the frontal patient, like the frontal animal, tends (. . .) to repeat old patterns of behavior even in circumstances that demand change. Repetitious routine seems to preempt what under those circumstances would be more adaptive behavior”. These observations suggest that one of the functions of the frontal lobe is to elaborate goal-directed programmes and to permit their expression by the inhibition of more automatic activities. This loss of ability to inhibit stereotypes may account for many of the behavioural disturbances described in patients with frontal lobe lesions, such as environmental adherence [34], tendency to simplify, and difficulty in maintaining complex instructions [38], because of the emergence of uncontrolled activities which are ordinarily repressed in normal subjects. Delayed Non-Alternation-Reversal

tasks

If the hypothesis proposed to explain the performance of patients with prefrontal lesion on the Delayed Alternation task is correct, they should have problems finding rules other than alternation. This was precisely the case with the Delayed Non-Alternation task, in which the performance of these patients was significantly worse than that of the control subjects and the patients with post-central lesion (Table 2), confirming that patients with prefrontal lesion have specific difficulties with rule acquisition. The deficit may result from an inability either to inhibit the previous rule (shifting deficit) or to elaborate the new rule that is required (learning deficit), or both.

1392

M. VERIN et

al.

Analysis of performance on the Delayed Reversal task may help us to answer this question since, in this task, there is no new rule to be elaborated: the patients need only to shift from one side to the other the previously deduced rule of non-alternation. The performance of control subjects clearly demonstrates that no learning process is needed for the Reversal task: 7 1% reversed the non-alternation rule immediately after their expected error on the first trial. The performance of the patients with post-central lesion was similar to that of the control subjects (Table 2), whereas that of the patients with prefrontal lesion was impaired at least for the first trials (see Fig. 4), indicating a specific set shifting defect in this situation. This suggests that the set acquisition deficit in patients with prefrontal lesion revealed by the Delayed Non-Alternation task, results from difficulty in extracting themselves from a previous behaviour rather than from difficulty in generating a new one. The process of elaboration of a new goal-directed set is prevented by the predominance of the last acquired set. This observation is consistent with results from experimental literature. As shown by HARLOW [24, 251 and MISHKIN [45], monkeys with prefrontal lesion were impaired on reversal tasks although they performed normally on non-reversal discrimination learning tasks. The results obtained suggest that the inability of patients with prefrontal lesion to modify their behaviour according to a changing environment results from a difficulty in inhibiting an already established mental set. Remarks of the patients during the Reversal task may help to understand this difficulty: “Each time I choose this side I win but every time Zchoose the other side Zlose”. The patients were not able, however, to use this observation to make the reversal. The feedback between the environment and the patient seems to persist but remains inoperable. This knowing/doing dissociation in humans with frontal lesion was interpreted by Luria as a defect in regulating behaviour through language [35]. Luria’s patients could repeat the instructions for the task but were unable to use these instructions to better their performance, and seemed little conscious of their errors. MILNER [43], in her study on the Wisconsin Card Sorting Test, interpreted the thought/action dissociation observed in patients with frontal lesion as the result of a more general inability to use external stimuli to guide their responses. As in our study, the patients described by Milner, as well as the patient described by KONOW and PRIBRAM [29], were conscious of their errors but were unable to avoid them. Delayed Response task In this task, there was no rule to learn: the subject had only to remember the position of the arrow during a 15 set delay. In this situation, the control subjects made no errors. They had no difficulty in maintaining an internal representation of the stimulus during the delay and in responding in accordance with the trace. In contrast, seven of the 10 patients in the prefrontal group made errors and their performance was significantly worse than that of both the control and post-central groups. These results suggest the existence of a specific deficit for prefrontal lesion, an hypothesis supported by the significant correlations between the performance of patients on the Delayed Response task and both the frontal score and the score for the graphic series, whereas no correlation with frontal lobe test scores was found for patients with post-central lesion. This specific deficit in the prefrontal group seems unlikely to result from difficulty in understanding the instructions because all of the patients were finally able to meet the criterion of 10 consecutive successful trials. This observation is supported by the fact that patients with prefrontal lesion obtained scores within the normal range on the MMSE (Table

DELAYED

RESPONSE

TASKS AND PREFRONTAL

LESIONS IN MAN

1393

1). A tendency to perseverate is not sufficient to explain the deficit because only 29% of the errors corresponded to the reiteration of the previous response. In contrast, inter- and intraindividual variability in the types of the error comitted by patients with prefrontal lesion favours the hypothesis of an attentional deficit for temporal processing of visuo-spatial stimuli. The Delayed Response paradigm requires that the response is made on the basis of an internal representation maintained during the delay [22, 231 rather than directly in reaction to the external stimulus. During the successive trials of the Delayed Response task, patients with prefrontal lesion might have difficulties with the inhibition of non-pertinent internal cue [56,X$60], such as the trace of a previous stimulus. This type of disorder was described by MILNER, in humans, as a deficit in judging recency [44, 503 and by MISHKIN [45], in monkeys, as a perseveration of central set. In addition to their temporal sequence, internal representations of visuo-spatial cues may be difficult to recall by patients with prefrontal lesion. Such a deficit in short-term spatial representational memory has been reported in monkeys with lesions of the sulcus principalis [22, 231, the region that is homologous to dorsolateral part of the prefrontal cortex in human. To summarize, it may be proposed that impaired execution of the Delayed Response task results from an attentional disorder affecting the handling of visuo-spatial representations and whose anatomical substrate could be the dorsolateral prefrontal cortex. If so, performance on the Delayed Response task may be considered as a behavioural marker of attentional processes requiring an intact prefrontal cortex.

CONCLUSION The spontaneous tendency to alternate on Delayed Alternation task, shown by patients with prefrontal lesion, seems to be automatic, requiring no voluntary mental operations on the part of the patients. Furthermore, their performance on subsequent tasks was characterized by the predominance of previously acquired patterns (alternation prevailed over non-alternation, which in turn prevailed over reversal), despite the presence of external indications to change behaviour, resulting in a not adapted behaviour to goal. The neuronal network involved in Delayed Reaction tasks has been the object of considerable investigation. It appears from these studies that lesions in any structure of the prefronto-striato-pallido-thalamo-prefrontal loops [2, 121 can impair the performance in these tasks [I, 5, 13, 18, 21, 26, 30, 49, 53, 573. There is also evidence that other neuropsychological processes considered to be specific to the frontal lobes can be disrupted by damage to the basal ganglia. BOWEN [7] and OBERG and DIVAC [48] have shown that interruption anywhere along the striato-frontal system may result in the predominance of previously established behaviours. Difficulty in shifting was also observed after both frontal lesions and striatal dysfunction of either degenerative origin, as in idiopathic Parkinson’s disease [62], progressive supranuclear palsy [4] and Huntington’s disease [8], or vascular origin, as in bipallidal ischemic accidents [31]. In the prefrontal group of patients studied here, the absence of striatal lesions might explain the emergence of a stereotyped behaviour such as spontaneous alternation. The striatum, then, could be viewed as the anatomic substrate of pre-elaborated routine behavioural programmes, in the same way that it is considered as the substrate for pre-elaborated motor programmes, according to MARSDEN [41]. These “automatic” programmes would be under constant inhibition by the prefrontal cortex. When environmental triggers activate one of these programmes, the inhibition must

1394

M. VERIN et al.

be released, either in an adapted manner, as in normal subjects, when it corresponds specifically to an assigned goal, or in a not adapted manner, when damaged prefrontostriatal connections can no longer suppress its activation. This hypothesis is in agreement with the model developed by SHALLICE [54,55]in which the Contention Scheduling System intervenes automatically when environmental conditions activate a schema corresponding to a routine behaviour, whereas the Supervisory Attentional System is activated when no adequate routine schema exists. In conclusion, our study suggests the existence of two types of behavioural organization. The first requires elaboration of new behavioural schemas by a learning process, permitting adaptation of the subject to new environmental situations. The second type of organization is independent of the environment and concerns routine and stereotyped behaviours, generated by subcortical structures which are normally repressed by the prefrontal cortex. In this hypothesis, lesion of the prefrontal cortex may release control of these stereotyped behaviours. The subcortical structures involved and the precise nature of their contribution in this behavioural organization remains, however, to be determined. Acknowledgements-The the manuscript.

authors

would like to thank

Dr Merle Ruberg

for her precious

help in the preparation

of

REFERENCES 1. ALEXANDER,G. E., W~rr, E. D. and GOLDMAN-RAKIC, P. S. Neuronal activity in the prefrontal cortex, caudate nucleus and mediodorsal thalamic nucleus during delayed response performance of immature and adult rhesus monkeys. Sot. Neurosci. Abstr. 6, 86, 1980. 2. ALEXANDER, G. E., DELONG, M. R. and STRIK, P. L. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Ann. Rev. Neurosci. 9, 357-381, 1986. 3. ALBEKT, M. L., FELDMAN, R. G. and WILLIS, A. L. The ‘subcortical dementia’ofprogressive supranuclear palsy. J. Neural. Neurosurg. Psychiat. 31, 121-130, 1974. 4. ALBERT, M. L. Subcortical dementia. In Alzheimer’s Disease: Senile Dementia and Related Disorders-Aging, R. KATZMAN, R. D. TERRY and K. L. BICK (Editors), Vol. 7, pp. 173-180. Raven Press, New York, 1978. 5. BATTIG, K., ROSVOLD, H. E. and MISHKIN, M. Comparison of the effects of frontal and caudate lesions on delayed response and alternation in monkeys. J. camp. Physiol. Psycho/. 53, 40&404, 1960. 6. BENTON, A. L. Differential behavioral effects in frontal lobe disease. Neuropsychologia 6, 53-60, 1968. 7. BOWEN, F. P. Behavioral alterations in patients with basal ganglia lesions. Research Publications-Association for Research in Nervous and Mental Disease 55, 169-189, 1976. 8. CAINE, E. D. and FISHER, J. M. Dementia in Huntington’s Disease. In Handbook ofclinical Neurology, F. JAM (Editor), Vol. 46, pp. 305-310. Elsevier, Amsterdam, 1985. 9. CANAVAN, A. G. M., PASSINGHAM, R. E., MARSDEN, C. D., QUINN, N., WYKE, M. and POLKEY, C. E. Prism adaptation and other tasks involving spatial abilities in patients with frontal lobes lesion and patients with unilateral temporal lobectomies. Neuropsychologia 9, 969-984, 1990. 10. CHOROVER, S. L. and COLE, M. Delayed alternation performance in patients with cerebral lesions. Neuropsycholoyia 4, 1-7, 1966. 11. DAMASIO, A. R. The frontal lobes. In Clinical Neuropsychology, K. M. HEILMAN and E. VALENSTEIN(Editors), pp. 339-375. Oxford University Press, London, 1985. 12. DELONG, M. R. and GEOHGOPOULOS, A. P. Motor functions of the basal ganglia. In Handbook ofPhysiologyThe Nervous System--Motor Control, V. B. BRWKS (Editor), Vol. 2, pp. 1017-1061. The American Physiological Society, MD 1981. 13. DIVAC, I., ROSVOLD, H. E. and SZWARCBART, M. K. Behavioral effects of selective ablation of the caudate nucleus. J. camp. Physiol. Psycho/. 63, 184-190, 1967. 14. DUBOIS, B., PILLON, B., LEGAULT, F., ACID, Y. and LHERMITTE,F. Slowing ofcognitive processing in progressive supranuclear palsy. A comparison with Parkinson’s disease. Arch. Neural. 45, 1194-l 199, 1988. 15. EBELING, U., HUBER, P. and REULEN, H. J. Localization of the precentral gyrus in the computed tomogram and its clinical application. J. Neural. 233, 73-76, 1986. 16. FOLSTEIN, M. F., FOLSTEIN, S. E. and MACHUGH, P. R. ‘Mini Mental State’. A practical method for grading the cognitive state of patients for the clinician. J. psychiarr. Res. 12, 189-198, 1975.

DELAYEDRESPONSETASKSAND PREFRONTALLESIONSIN MAN

1395

17. FREEDMAN, M. and OSCAR-BERMAN, M. Bilateral frontal lobe disease and selective delayed-response deficits in humans. Behau. Neurosci. 100, 337-342, 1986. 18. FUSTER, J. M. and ALEXANDER, G. E. Firing changes in cells of the nucleus medialis dorsalis associated with delayed response behavior. Brain Res. 61, 79-91, 1973. 19. FUSTER, J. M. The Prefrontal Cortex, 2nd edition. Raven Press, New York, 1989. 20. GOLDMAN, P. S., ROSVOLD, H. E., VEST, B. and GALKIN, T. W. Analysis of the delayed-alternation deficit produced by dorsolateral prefrontal lesions in the rhesus monkey. J. camp. Physiol. Psychol. 77,212-220,1971. 21. GOLDMAN, P. S. and ROSVOLD, H. E. The effects of selective caudate lesions in infant and juvenile rhesus monkey. Brain Res. 43, 53-66, 1972. 22. GOLDMAN-RAKIC, P. S. Circuitry of primate prefrontal cortex and regulation of behavior by representational memory. In Handbook of Physiology, F. PLUM and U. MOUNTCASTLE (Editors), Vol. 5, pp. 373417. The American Physiological Society, Washington, 1987. 23. GOLDMAN-RAKIC, P. S. and FRIEDMAN, H. R. The circuitry of working memory revealed by anatomy and metabolic imaging. In Frontal Lobe Function and Dysfunction, H. S. LEVIN, H. M. EISENBERGand A. L. BENTON (Editors), pp. 72-91. Oxford University Press, New York, 1991. 24. HARLOW, H. F. and DAGNON, J. Problem solution by monkeys following bilateral removal ofprefrontal areas. I. Discrimination and discrimination reversal problems. J. exp. Psycho/. 32, 351-356, 1943. 25. HARLOW, H. F. Analysis of discrimination learning by monkeys. J. exp. Psycho/. 40, 26-39, 1950. 26. ISSEROFF, A., ROSVOLD, H. E., GALKIN, T. W. and GOLDMAN-RAKIC, P. S. Spatial memory impairments following damage to the mediodorsal nucleus of the thalamus in the rhesus monkeys. Brain Res. 232,97-l 13, 1982. 27. JACOBSEN, C. F. Studies of cerebral function in primates. Camp. Psycho/. Monogr. 13, l-68, 1936. 28. KONORSKI, J. and LAWICKA, W. Analysis of errors by prefrontal animals on the delayed-response test. In The Frontal Granular Cortex and Behavior, J. M. WARREN and K. AKERT (Editors), pp. 271-294. McGraw-Hill, New York, 1964. 29. KONOW, A. and PKIBRAM, K. H. Error recognition and utilization produced by injury to the frontal cortex in man. Neuropsychologia 8,489%491, 1970. 30. KUBOTA, K., NIKI, H. and GOTO, A. Thalamic unit activity and delayed alternation performance in the monkey. Acta Neurobiol. Exp. 32, 177-192, 1972. 31. LAPLANE, D., LEVASSEUR,M., PILLON, B., DUBOIS, B., BAULAC, M., MAZOYER, B., TRAN DINH, S., SETTE, G., DANZE, F. and BARON, J. C. Obsessive
1396

M.

VERINet ai.

47. NELSON, H. E. A modified card sorting test sensitive to frontal lobe defect. Cortex 12, 313-324, 1976. 48. ~BERG, R. G. E. and DIVAC,I. ‘Cognitive’ functions of the neostriatum. In The Neostriatum, I. D~VACand

R. G. E. C~BERG (Editors), pp. 291-313. Pergamon Press, New York, 1979. 49. PETERS,R. H., ROSVOLD, H. E. and MIRSKY,A. F. Theeffect ofthalamic lesions upon delayed response-type tests in the rhesus monkey. J. camp. Physiol. Psycho/. 49, 11l-l 16, 1956. 50. PETRIDES, M. and MILNER,B. Deficit on subject-ordered tasks after frontal- and temporal-lobe lesions in man. Neuropsychoiogia 20, 249-262, 1982. 51. PRIBRAM,K. H., AHUMADA,A., HARTOG,J. and Ross, L. A progress report on the neurological processes disturbed by frontal lesions in primates. In The Frontal Granular Cortex and Behavior, J. M. WARRENand K. AKERT(Editors), pp. 28-55. McGraw-Hill, New York, 1964. 52. ROSEXKILDE, C. E. Functional heterogeneity of prefrontal cortex in the monkey: A review. Behau. Neural Biol. 25,301-345, 1979. 53. ROSVOLD, H. E. and SZWARCBART, M. K. Neural structures involved in delayed-response performance. In The Frontal Granular Cortex and Behavior, J. M. WARRENand K. AKERT(Editors), pp. l-15. McGraw-Hill, New York, 1964. 54. SHALLICE,T. Specific impairments of planning. Phil. Trans. R. Sot. Land. (Bio[.) 298, 211-226, 1982. 55. SHALLICE, T. The allocation of processing resources: Higher-level control. In From Neuropsychology to Mental Structures, pp. 328-352. Cambridge University Press, Cambridge, 1989. 56. SHIMAMURA, A. P., JANOWSKY, J. S. and SQUIRE,L. R. What is the role of frontal lobe damage in memory disorders? In FrontalLobe Function and Dysfunction, H. S. LEVIN,H. M. EISENBERG and A. L. BENTON(Editors), pp. 173-195. Oxford University Press, New York, 1991. 57. SHULMAN,S. Impaired delayed response from thalamic lesions. Archs Neurol. Chicago 11,477-499, 1964. 58. SQUIRE,L. R. Prefrontal cortex. In Memory and Brain, pp. 224-240. Oxford University Press, Oxford, 1987. 59. STLJSS, D. T. and BENSON,D. F. Neuropsychological studies of the frontal lobes. Psychol. Rev. 9S,3-28,1984. 60. SXJSS, D. T. and BENSON,D. F. The Frontal Lobes. Raven Press, New York, 1986. 61. TALAIRACH,J. and SZIKLA,G. Atlas of Stereotaxic Anatomy of the Telencephalon. Masson, Paris, 1967. 62. TAYLOR,A. E., SAINT-CYR,J. A. and LANG,A. E. Frontal lobe dysfunction in Parkinson’s disease. Brain 109, 845-883, 1986. 63. WANG, P. L. Concept formation and frontal lobe function. In The Frontal Lobes Revisited, E. PERECMAN (Editor), pp. 189-205. The IRBN Press, New York, 1987. 64. WEINBERGER,D. R. Schizophrenia and frontal lobe. Trends Neurosci. 11,367-370, 1988.