AN ECOLOGICAL APPROACH TO PLANNING

DYSFUNCTION: SCRIPT EXECUTION

Mathilde Chevignard1, Bernard Pillon2, Pascale Pradat-Diehl1, Chantal Taillefer1,
Sylvie Rousseau1, Catherine Le Bras2 and Bruno Dubois2
(1Service de Rééducation Neurologique, 2INSERM EPI 007 and Fédération de Neurologie,
Hôpital de la Salpêtrière, Paris, France)

ABSTRACT

Planning, which concerns many activities in everyday life, is a two-stage process. The
first one predetermines a course of actions aimed at achieving some specific goals. It is
founded on managerial knowledge or overlearned sequences of events and may be tested
by script generation. The second stage entails monitoring and guiding the execution of the
plan to a successful conclusion. It must take into account environmental contingencies and
may be tested by script execution. If the frontal lobes intervene not only in managerial
knowledge (Grafman, 1989) but also in binding the plan with contextual environment
(Damasio, Tranel and Damasio, 1991; Shallice and Burgess, 1991), script execution would
be more sensitive than script generation to planning deficits. To test this hypothesis, script
execution and script generation were compared in 11 patients with a dysexecutive syndrome
and 10 matched controls, using three scripts of daily life activities: (1) ‘shopping for
groceries’; (2) ‘cooking’; (3) ‘answering a letter and finding the way to post the reply’.
Two way ANOVAs showed more errors in execution than in generation, more errors in
patients than in controls, and a greater difference between execution and generation in
patients than in controls. Furthermore, ‘context neglect’ and ‘environmental adherence’
were the two types of errors that best differentiated patients from controls. Finally, the total
number of errors in execution correlated with the score on behavioral questionnaires
answered by occupational therapists. These results confirm our hypothesis and suggest that
script execution may be a valid ecological approach to estimate the severity of deficits in
daily life activities.

Key words: script generation and execution, planning, dysexecutive syndrome,

ecological approach, frontal lobe lesions, closed brain injury

INTRODUCTION

Planning is the ability to organize behavior in relation to a goal that must be

achieved through a series of intermediate steps (Luria, 1966; Owen, 1997). This
ability is currently assessed with laboratory tests, such as maze tracing (Porteus,
1959; Karnath, Wallesch and Zimmerman, 1991) and problem resolution (Luria,
1966; Lhermitte, Derouesné and Signoret, 1972; Klosowska, 1976), including the
Tower of London Task (Shallice, 1982, 1988). The performance of patients with
frontal lobe lesions on such tests is regularly impaired (Owen, Downes,
Sahakian et al., 1990; Morris, Miotto, Feigenbaum et al., 1997), underlining the
role of the dorsolateral prefrontal cortex in cognitive organization (Stuss, Eskes
and Foster, 1994). This cognitive role has recently been confirmed by PET
studies of normal subjects resolving Tower Tasks (Baker, Rogers, Owen et al.,

Cortex, (2000) 36, 649-669

650 Mathilde Chevignard and Others

1996; Owen, Doyon, Petrides et al., 1996). Planning is also important for
behavioral adaptation in daily life. For example, following frontal lobe surgery a
woman failed to prepare an entire family meal, although she was capable of
cooking the individual dishes (Penfield and Evans, 1935). Since this description,
several neuropsychological studies showed that daily life activities might be
impaired despite normal cognitive performance in the laboratory (Eslinger and
Damasio, 1985; Shallice and Burgess, 1991). As underlined by the latter, the
difference might be due to the need of maintaining the plan over longer periods
of time in the face of competing alternatives in daily life situations.
However, assessing planning in daily life is difficult. Due to a reduction in
insight associated with frontal lobe lesions (Stuss, 1991; Prigatano, 1991),
patients generally underestimate their functional impairment in everyday
situations. Relative’s or caregiver’s rating on behavioral scales (Lhermitte, Pillon
and Serdaru, 1986) has been found to correlate with the metabolism of specific
regions of the frontal cortex, namely orbitofrontal areas (Sarazin, Pillon,
Giannakopoulos et al., 1998). Rating depends, however, on subjective feelings
and many caregivers also tend to underestimate the patient’s deficits. Therefore,
more objective evaluations have been proposed, such as the Six Element or the
Multiple Errands Tests (Shallice and Burgess, 1991), the Executive Function
Route Finding Test (Boyd and Sautter, 1993) or the Behavioural Assessment of
the Dysexecutive Syndrome (Wilson, Evans, Emslie et al., 1998). Although
ecologically valid, these tests show a great variability of performance even in
normal controls, some of them being as severely impaired as patients with
frontal-lobe lesions (Levine, Stuss, Milberg et al., 1998). The impaired
performance of normal subjects on these tasks underlines the importance of
using more familiar plans of actions in order to limit the inter-individual
variability of normal control subjects and allow a better differentiation between
patients and controls. Naturalistic actions, even as simple as preparing a slice of
toast with butter and jam or packing a lunchbox, may be impaired in patients
with brain lesions, at least in cases of closed head injury (Schwartz,
Montgomery, Buxbaum et al., 1998).
One recent cognitive model postulates that planning deficits are related to
impairments in a domain specific to managerial knowledge involved in the
representation and manipulation of script-like structures (Grafman, 1989).
Managerial Knowledge Units would be overlearned sequences of events with a
beginning and a end and a hierarchical organization going from more abstract
(eating to a restaurant) to more concrete levels (paying the bill). Frontal lobe
lesions would specifically affect this kind of representations. In agreement with
this hypothesis, script generation and sorting has been shown to be impaired in
patients with frontal-lobe lesions (Le Gall, Aubin, Alain et al., 1993; Godbout
and Doyon, 1995). In previous studies, we found that these patients made errors
in ordering actions in the correct temporal sequence, failed to close scripts and
remain within the stated boundaries, and made deviant estimates of action
importance (Sirigu, Zalla, Pillon et al., 1995, 1996).
Script generation, which is mainly based on the ability to access script-
related content, only involves the first stage of planning, i.e. the
predetermination of a course of action aimed at achieving some goal. The
 Script execution 651

second stage entails self-monitoring and guiding the execution of the plan to a
successful conclusion (Hayes-Roth and Hayes-Roth, 1979). Two qualitatively
distinct processes could determine which particular schema of action would be
activated at a given moment of the script execution: an automatic process, the
Contention Scheduling, which would operate in familiar situations, and a
controlled process, the Supervisory Attentional System, which is thought to
modulate operations when situations are unfamiliar (Shallice, 1988). The two
processes could allow integration between plan execution and environmental
contingencies by operations such as marker creation, marker triggering and
evaluation of the consequences of action. Only the Supervisory Attentional
System would depend on the prefrontal cortex (Damasio, Tranel and Damasio,
1991; Shallice and Burgess, 1991; Verin, Partiot, Pillon et al., 1993; Dubois,
Levy, Verin et al., 1995).
If the frontal lobes are indeed involved in both managerial knowledge and
monitoring of action, script execution would be more sensitive than script
generation to planning dysfunction. The first aim of the study was to test this
hypothesis by comparing execution and generation of relatively familiar scripts,
such as ‘shopping for groceries’, ‘cooking’ and ‘answering a letter and finding
the way to post the reply’. The second aim of the study was to evaluate the
validity of such script execution as an ecological approach to deficits of daily
life activities and to test the hypothesis that it provides a more reliable indication
of dysexecutive deficits in real life than performance in cognitive tests.

MATERIALS AND METHODS

Subjects

Patients

Given the aims of the study, criteria for inclusion were: (1) the evidence of executive
deficits in daily life activities shown by a score > 25 on the dysexecutive questionnaire of
Wilson et al. (1998), according to the rating of either the caregiver or the occupational
therapist; and (2) the existence of structural damage within the frontal lobes on CT scanner
and/or MRI (see Figure 1 and Table I). Exclusion criteria were: (1) pre-existing psychiatric
or neurological disorders; (2) intellectual deterioration (performance < centile 50 on the PM
38 Raven’s Progressive Matrices); (3) motor or linguistic deficits, sufficiently severe to
interfere with the tasks. Eleven patients (9 men, 2 women) were selected: two had focal
brain damage restricted to the frontal lobes (left dorsolateral prefrontal cortex haemorrhage
in one patient, orbitofrontal lesion resulting from the surgical removal of a meningioma in
the other); nine suffered from severe post-traumatic brain injury. Age ranged from 23 to 60
years [mean (SD) = 35.4 (12.3)], and educational level from 8 to 19 years of scolarity [13.8
(3.4)]. Duration of disease ranged from 3 to 80 months [27.3 (12.3)]. The performance of
patients in laboratory tests aimed at assessing executive functions are shown in Table II.
For the 9 patients with post-traumatic brain injury, the initial Glasgow score was 6.6 (1.4;
range: 3-8), coma duration was 16.6 (12.8) days, and post-traumatic amnesia duration was
76.3 (60.9) days.
All patients showed severe difficulties in daily life. Two of them had returned to
employment (OM worked part time in a sheltered employment and JFC worked full time
with his former employer), but both lost their job a few months later. One other subject
had attempted to return to work, though without medical team approval, but was soon fired.
Several caregivers reported accidents in everyday life, some of them being obviously
652 Mathilde Chevignard and Others

Fig. 1 – Lesion location and extent.

Script execution 653
654 Mathilde Chevignard and Others

TABLE I
Localisation of Cerebral Lesions

Patients Duration Initial Coma PTA Localisation of cerebral lesions

of disease Glasgow duration duration on CT scanner or MRI
(months) score (days) (days)
OM 31 5 7 120 Left prefrontal cortical and subcortical;
bilateral frontal subcortical
EG 5,5 8 7 7 Bilateral anterior prefrontal cortex
DG 24,5 5 21 111 Bilateral prefrontal cortical and subcortical,
left temporal, left internal capsule
JFC 1,5 8 6 15 Left prefrontal and temporal cortex
AG 81 8 45 180 Right prefrontal cortical and subcortical
LD 18,5 5 21 135 Bilateral (left > right) prefrontal cortical and
subcortical
SL 41 7 20 60 Bilateral prefrontal cortical and subcortical;
right internal capsule
PR 8,5 5 14 90 Bilateral prefrontal cortical; left temporal
MT 2 6 8 20 Left prefrontal cortical and subcortical;
left internal capsule
MFG 39 Left dorsolateral prefrontal cortex
MC 36 Bilateral (right > left) orbitofrontal

TABLE II
Neuropsychological Characteristcs of Patients

Mean Standard p
deviation
Wisconsin CST
Number of criteria 5.7 0.9
Number of errors 3.5 3.4
Number of perseverations 1.2 1.7
Tower of London Task
3N-Number of moves 3.5 1.1
3N-Time (sec.) 16.2 19.4
5N-Number of moves 7.7 2.7
5N-Time (sec.) 26.8 11.6
5I(+)-Number of moves 7.5 2.3
5I(+)-Time (sec.) 32.6 21.2
5I(-)-Number of moves 10 5.5
5I(-)-Time (sec.) 42 37.6
Six Elements Test
Global score 880 176.9
Number of errors 6.3 3.3 *
Trail Making Test
Form A 54.5 24.2 *
Form B 118.2 67.6 *
B–A 63.6 47.6
Verbal fluency
Category (animals) 20.7 5.1
Literal (P, F, L) 28.5 8.1
Grober and Buschke Verbal Learning Test
Total free recall 30.1 7.1 *
Total recall 47.2 0.9
Sensitivity to cues 95%
Delayed free recall 10.7 3.7 *
Delayed total recall 15.5 0.7
* p < 0.05 compared to norms.
 Script execution 655

abnormal. One patient (DG), previously an engineering student, suffered burns while trying
to make holes in plastic bottles with a screwdriver previously heated on the stove; he then
took a shower ‘for 21 minutes’ and went to bed without calling a doctor, despite having
third degree burns on all the upper part of his body. Although the caregivers were able to
signal behavioral disorders, they currently underestimated their importance and potential
consequences: the score of the patients on the dysexecutive questionnaire (Wilson et al.,
1998) was 26.9 (12.9) when rated by the caregivers versus 35.3 (13.0) when rated by
occupational therapists.

Controls

The control group comprised ten subjects with no history of neurological or psychiatric
disease. They were matched for sex (8 males, 2 females), age [31.6 (9.4) years] and
education level [15.0 (2.1) years] with the frontal patients.
All subjects gave informed consent.

Script Presentation

We chose scripts among current activities of daily life. They had to be complex enough
to be sensitive to impaired dysexecutive functions and to be feasible in a rehabilitation
center, over a relatively short time period. Three scripts were studied: (1) shopping for
groceries; (2) preparing two recipes of different levels of difficulty (scrambled eggs for two
and a chocolate cake); a written recipe was only proposed for the cake, as cooking
scrambled eggs is easy, even without a wide experience of cooking; and (3) answering a
letter and finding one’s way to a given post-box to post the reply. The familiarity of each
subject with the scripts was previously controlled using a 3-point scale (0-2): never (0),
sometimes (1), often put into practice (2). There was no significant difference in script
familiarity between patients and controls. For each script, execution and verbal generation
of actions were compared and the order of presentation was randomized.

Script Execution

The ‘shopping for groceries’ and ‘cooking’ scripts were performed on the same day.
On a shelf, several items were displayed always in the same location: (1) salt, pepper and
oil (usually found in a standard kitchen); (2) all the utensils necessary for the cooking
activity; and (3) some items unnecessary for this particular activity, but normally present in
a kitchen, that were used as distracters. On the same shelf, there was a closed cookbook:
the cake recipe was on page five. One of the examiners (MC) defined each activity with a
statement describing its goal and the script’s starting point and ending point. The examiner
made sure the subject had correctly understood the instructions and told him he could
whenever necessary consult them on a cue card, which remained available throughout the
task. For ‘shopping for groceries’, the instructions were: “You must go shopping and buy
everything you need to cook scrambled eggs for two persons and bake a chocolate cake.
The recipe for the cake is in this book. None of the ingredients are available here (except
salt, pepper and oil). The supermarket is not far from the hospital. We will take you there.
You must do your shopping as if you were alone. Tell us when you have finished”. For the
‘cooking’ activity, the instructions were: “You must bake a chocolate cake and cook
scrambled eggs for two persons. The cake recipe is available in this book. All the utensils
you need are available here (the examiner points to the shelf). You should use the
ingredients you have just bought. I cannot help you. You must act as if you were alone.
Tell me when you have finished.”
The ‘answering a letter’ script was usually performed on another day. The activity
started in the examiner’s office. The items displayed on the desk were always the same and
in the same location. They comprised various types of paper, envelopes, stamps and a pen.
The instructions were: “You must answer this letter fixing an appointment for a medical
examination. You must then go and post it in the letterbox outside the rheumatology
656 Mathilde Chevignard and Others

department. This place is unfamiliar to you. You must find your own way there. I cannot
help you. You must act as if you were alone. Tell me when you have finished.”
For each activity, two examiners were present. They followed the patient everywhere
and both wrote down everything he did or said. They intervened only when the situation
might had become dangerous. The supermarket manager was aware of the study and gave
his consent.

Script Generation

As in script execution, the examiner first defined each activity with a statement
describing the goal and the script’s starting point and ending point. Instructions were given
verbally and were displayed on a cue card, which remained available throughout the task.
Before the subject started, the examiner made sure he had correctly understood the
instructions. For each script, the instructions stated that the subject had to write as many
actions as he could think of, and that the examiner could not help him at all.

Script Analysis

Script Execution

Once the two examiners had written down everything that happened during the
execution of each script, they were able to produce a comprehensive list of actions that
occurred, regardless of the group. Error analysis was first tried using the framework of the
cognitive model proposed by Lezak, Le Gall and Aubin (1994) according to which an error
of execution may reflect a dysfunction in ‘volition’, ‘planning’, ‘goal direction’ or ‘task
monitoring’. Many errors could be interpreted, however, as a dysfunction in more than one
level of the cognitive model. Consequently, five types of descriptive errors were first
defined according to an inter-judge agreement: (1) omission: any action or sequence of
actions necessary for goal reaching which is omitted or incompletely performed, such as
going to the supermarket without taking money; (2) addition: any action or sequence of
actions unnecessary for the completion of the task which is performed, such as buying
ingredients not necessary for the cooking scripts; (3) inversion-substitution: any action
performed out of the appropriate temporal sequence or any object that is misused or
inappropriate to the sub-goal, such as putting the cake in the baker before lighting it or
putting a salad bowl in the baker; (4) estimation error: poor estimation of the quantity of
ingredients, of the size of an object, of space or time, such as putting too much sugar or
chocolate in the cake; (5) commentary: any question, remark or joke addressed to the
examiners, although the patient had been instructed to act as if he were alone, such as
asking how to do something. At this first step of analysis, each error was counted and
included in only one category of error. Then, a second step analysis was made, taking into
account neuropsychological mechanisms and allowing to define a six-category classification:
(1) context neglect: failure to respect instructions, poor evaluation of environment, such as
not taking the recipes into account or searching for spoons or tools in drawers instead of
using those being on the working space; (2) control errors: inefficient monitoring of action,
such as letting the cake to be burnt; (3) environmental adherence: inappropriate action
induced by the presence of an object, such as washing plates being in the sink and not
required by the scripts; (4) distractibility: behavioural sequence not contributing to goal
achievement, such as exploring the kitchen or staying a long time without doing anything;
(5) dependency: any question about the way to perform action, such as asking what there
was to do; and (6) behavioural disorder: any socially inappropriate or dangerous behaviour,
such as making jokes about people in the shop or taking the hot scrambled eggs with the
fingers. In this second step analysis, some errors could be also classified under two different
categories of errors, or not classified, which accounts for the fact that the total number of
errors was not the same at the descriptive and the neuropsychological levels. A perfect
inter-raters agreement was required to count or classify an error. These analyses were used
for ‘shopping’, ‘cooking’ and ‘answering a letter’ scripts. As the way finding part of the
 Script execution 657

‘letter’ script is similar to the Executive Function Route-Finding Test of Boyd and Sautter
(1993), it was analyzed according to both our and to their criteria, allowing the comparison
between the two methods.

Script Generation

In order to compare the performance in script execution and generation, we classified

the errors produced in script generation with the same framework as the execution errors.
The existence of errors of ‘context neglect’, ‘estimation’ or ‘dependency’ may be surprising
in script generation. Some errors, however, could be observed. In generation, as in
execution, the recipes necessary to know which ingredient to buy could be neglected, the
quantity of ingredients to use could be poorly estimated, and questions about the way to
perform could be asked in spite of the consign to work alone. To allow comparison with
previous studies, the centrality of each evoked action and mean evocation time were also
analyzed. An action was considered as central when evoked by at least 70% of the controls.

Neuropsychological Evaluation

The cognitive assessment used the following laboratory tests: Modified Wisconsin Card
Sorting Test (Nelson, 1976), Trail Making Test (Reitan, 1958), category (animal names in
60 sec.) and phonemic (words beginning with “P”, “F” and “L” in 60 sec.) verbal fluency
(Benton, 1968), Grober and Buschke Verbal Learning Test (1987), planning tasks such as
the Tower of London Task (Shallice, 1982) and the Six Elements Task (Shallice and
Burgess, 1991; Garnier, Enot-Joyeux, Jokic et al., 1998; Wilson et al., 1998).
The behavioral assessment included two questionnaires constructed to sample the range
of dysexecutive problems commonly encountered in daily life: the Behavioral Scale
(Lhermitte et al., 1986; Sarazin et al., 1998) and the Dysexecutive Questionnaire (Wilson,
Alderman, Burgess et al., 1996; Wilson et al.,1998). The Behavioral Scale was designed to
evaluate, by comparison with the premorbid state of the patient, the severity of frontal lobe-
related behavioral abnormalities such as apathy, restlessness, stereotypy, impulsiveness,
indifference, euphoria, disinterestedness, cheerfulness, dependence on social environment,
indifference to rules, decrease in attention, dependence on stimuli from the physical
environment, programming disorders, personality changes, and impairment of intellectual
and emotional control. The number of true/false answers to five items (giving a score from
0, absence of impairment, to 5, severe impairment) assessed the importance of each of these
sixteen features (eighty items in all, presented in random order). The Behavioral Scale has
been shown to be sensitive to frontal lobe lesions (Lhermitte et al., 1986) and scores
obtained by patients with such lesions were significantly correlated with regional glucose
consumption in orbitoventral areas, as determined using PET (Sarazin et al., 1998). The 20-
item Dysexecutive Questionnaire samples four broad areas of dysexecutive changes:
emotional or personality, motivational, behavioral and cognitive changes. Each item is
scored on a five-point (0-4) Likert scale, ranging from never to very often. Both
questionnaires have two versions: one to be administered to the patient and the other to the
caregiver or relative who is in close contact with the patient.

Data Analysis

Statistical analyses were performed with ANOVAs using group as between factor and
experimental condition (script execution and script generation) as within factor. Given
variance heterogeneity, the group and condition effects were controlled with non-parametric
analyses, which gave similar results. We looked for a correlation between scores in the
experimental tasks and the performance in the laboratory evaluation of executive functions
on one hand, the scores on the behavioural questionnaires on the other hand, using the
Spearman rank correlation test.
658 Mathilde Chevignard and Others

RESULTS

Script Execution
Analysis of Errors
All control subjects made errors during script execution. The mean (SD)
number of errors was 7.6 (2.5) for ‘shopping’, 16.9 (7.8) for ‘cooking’ and 3.0
(1.8) for ‘answering the letter’. The most frequent types of errors were
‘commentaries’ and ‘context neglect’, but nearly all categories of errors were
represented (see Tables III, IV and V).
TABLE III
Comparison of Errors Made by Controls and Patients with a Dysexecutive Syndrome
in the ‘Shopping’ Script

Controls Patients p

Shopping Generation Execution Generation Execution Group Exec/Gen Interaction

Descriptive level
Omission 1.7 (1.6) 1.4 (1.7) 7.1 (2.0) 7.7 (3.6) ***
Addition 0.0 (0.0) 2.0 (2.1) 1.5 (4.8) 6.6 (6.3) **
Inversion 0.1 (0.3) 0.8 (0.9) 0.3 (0.6) 2.1 (2.2) **
Estimation 0.0 (0.0) 0.7 (0.8) 0.6 (1.3) 1.8 (1.7) * **
Commentary 2.0 (1.5) 2.7 (2.8) 3.9 (3.8) 6.1 (3.6) *
Total 3.8 (2.5) 7.6 (2.5) 13.5 (8.3) 24.4 (11.9) *** **
Neuropsychological level
Context neglect 1.6 (1.2) 3.9 (1.6) 3.6 (4.7) 13.5 (5.3) *** *** ***
Control 0.0 (0.0) 0.2 (0.4) 0.0 (0.0) 1.4 (1.1) ** *** **
Adherence 0.0 (0.0) 1.0 (0.9) 0.4 (0.9) 4.3 (2.7) ** *** **
Distractibility 0.0 (0.0) 1.3 (1.6) 0.0 (0.0) 4.6 (5.4) **
Dependency 1.5 (1.2) 1.8 (1.1) 1.2 (1.0) 4.5 (2.2) * *** **
Behaviour 0.0 (0.0) 0.7 (0.9) 0.2 (0.4) 2.5 (2.8) **
Note. Results are expressed as mean (SD). * p < 0.05; ** p < 0.01; *** p < 0.001.

TABLE IV
Comparison of Errors Made by Controls and Patients with a Dysexecutive Syndrome
in the ‘Cooking’ Script

Controls Patients p

Cooking Generation Execution Generation Execution Group Exec/Gen Interaction

Descriptive level
Omission 4.2 (5.0) 3.8 (1.0) 12.1 (4.9) 23.4 (16.6) *** * *
Addition 0.0 (0.0) 4.0 (2.7) 0.4 (0.7) 25.1 (20.3) ** *** **
Inversion 0.1 (0.3) 1.5 (1.1) 0.6 (1.0) 16.1 (8.7) *** *** ***
Estimation 0.0 (0.0) 1.0 (1.2) 0. 3 (0.5) 7. 5 (5.1) *** *** **
Commentary 0.9 (1.4) 6.6 (4.9) 5.5 (4.1) 17.3 (14.5) ** **
Total 5.2 (5.2) 16.9 (7.8) 18.8 (8.1) 89.3 (55.6) *** *** **
Neuropsychological level
Context neglect 0.8 (0.9) 9.6 (4.9) 3.1 (2.8) 64.8 (42.1) *** *** ***
Control 0.0 (0.0) 0.4 (0.7) 0.0 (0.0) 6.3 (3.6) *** *** ***
Adherence 0.0 (0.0) 1.0 (1.3) 0.1 (0.3) 16.4 (11.4) *** *** ***
Distractibility 0.0 (0.0) 1.6 (1.4) 0.0 (0.0) 6.4 (5.5) * *** *
Dependency 0.3 (0.7) 4.0 (4.2) 1.8 (2.6) 10.1 (7.9) * ***
Behaviour 0.0 (0.0) 1.7 (1.9) 0.3 (0.9) 11.1 (9.1) ** *** **
Note. Results are expressed as mean (SD). * p < 0.05; ** p < 0.01; *** p < 0.001.
 Script execution 659

TABLE V
Comparison of Errors Made by Controls and Patients with a Dysexecutive Syndrome
in the ‘Letter’ Script

Controls Patients p

Shopping Generation Execution Generation Execution Group Exec/Gen Interaction

Descriptive level
Omission 2.2 (1.7) 0.5 (0.5) 8.1 (2.6) 3.3 (4.2) *** ***
Addition 0.0 (0.0) 0.7 (0.8) 0.3 (0.9) 2.0 (3.0) *
Inversion 0.0 (0.0) 0.2 (0.4) 0.2 (0.4) 2.1 (1.5) *** *** **
Estimation 0.0 (0.0) 0.0 (0.0) 0. 0 (0.0) 0.2 (0.4)
Commentary 1.1 (1.3) 1.6 (1.9) 3.6 (2.5) 6.5 (7.6) *
Total 3.2 (2.3) 3.0 (1.8) 12.2 (4.2) 14.1 (9.7) ***
Neuropsychological level
Context neglect 0.6 (0.8) 1.6 (0.8) 1.3 (1.1) 9.1 (5.8) *** *** **
Control 0.0 (0.0) 0.1 (0.3) 0.0 (0.0) 0.0 (0.0)
Adherence 0.0 (0.0) 0.1 (0.3) 0.4 (0.7) 2.0 (1.2) *** *** ***
Distractibility 0.0 (0.0) 0.2 (0.4) 0.0 (0.0) 1. 9 (3.5)
Dependency 0.4 (0.5) 0.6 (0.7) 0.7 (0.9) 2.9 (1.9) ** *** **
Behaviour 0.2 (0.6) 0.0 (0.0) 0.2 (0.4) 2.2 (5.3)
Note. Results are expressed as mean (SD). * p < 0.05; ** p < 0.01; *** p < 0.001.

Patients made a significantly (p < 0.01) higher number of errors for the three
scripts: 24.4 (11.9) for ‘shopping’, 89.3 (55.6) for ‘cooking’, and 14.1 (9.7) for
‘answering the letter’. The difference between patients and controls was
significant for nearly all categories of errors, particularly for ‘cooking’ (Figure
2). However, ‘omissions’, ‘context neglect’ and ‘environmental adherence’ best
differentiated patients from controls for each of the three scripts.
The total number of errors in the execution of each of the three scripts was
significantly correlated with that of the two others: rho = 0.7 (p < 0.03) for
‘shopping’ and ‘cooking’; rho = 0.7 (p < 0.03) for ‘shopping’ and ‘answering
the letter’; rho = 0.65 (p = 0.04) for ‘cooking’ and ‘answering the letter’.

Mean Script Execution Time

Script execution time, from the beginning to the end of the script, was
significantly longer for the patient group than for controls: 20.5 (8.5) versus 10.5
(2.3) minutes for ‘shopping’ (p = 0.0019); 81.2 (33.5) versus 51 (9.9) minutes
for ‘cooking’ (p = 0.013); and 17.9 (11.0) versus 10.2 (3.9) minutes for
‘answering the letter’ (p = 0.05).

Money Spent in the Supermarket

There was no significant difference between the amount of money spent

by the patients and the controls [54.9 (13.6) versus 55.2 (10.1) French
francs] during shopping. Omissions were, however, more frequent in patients
than in controls, suggesting that controls devoted more attention to prices than
patients.
660 Mathilde Chevignard and Others

Controls

Patients

Fig. 2 – Execution of the ‘cooking’ script: performance of controls and patients with a
dysexecutive syndrome. 1: number total of errors. 2: types of errors at the descriptive level. 3: types
of errors at the neuropsychological level. Results are expressed as mean (SD). * p < 0.05;
** p < 0.01; *** p < 0.001.

Executive Function Route-Finding Test

‘Posting the letter’ (Figure 3) was also analyzed according to the Executive
Function Route Finding Test criteria (Boyd and Sautter, 1993). They include: (1)
task understanding or the ability to grasp the nature of an open-ended task; (2)
information seeking or the strategies used to find the destination; (3) retaining
directions or the manner of retaining important information; (4) error detection
or the extent to which discrepancies between directions and performance were
noted; (5) error correction or level of independence once difficulties are
encountered; (6) on-task behavior or the extent to which digressions of the task
interfere with performance. Likert rating for each criteria ranges from one
(extensive dependence on the examiner) to 4 (independent completion). An
average score (from 0 to 4) is obtained by dividing the total score by the number
of criteria. Performance by patients was significantly impaired when compared
to that of controls for the mean total score achieved [3.4/4 (0.6) versus 3.9/4
(0.1), p = 0.01] and for the mean execution time [15.4 (6.5) minutes versus 8.5
(2.9), p = 0.006]. When the patient deviates from the path approaching the goal
 Script execution 661

Patient D.L.
Control

Fig. 3 – Route-finding in the Pitié-Salpêtrière Hôspital.

and passes up an opportunity for correction, he is given a non-specific cue to

alert him to monitor performance (What do you need to do now?). If the patient
fails to attempt correction, a specific cue provides information on how to
execute the task. The number of non-specific cues provided by the examiner
during the task tended to differ between the two groups [2.5 (4.0) in patients
versus 0.2 (0.4) in controls, p = 0.09], but not that of specific cues [0.3 (0.5) in
patients versus 0.1 (0.3) in controls, p = 0.34].
662 Mathilde Chevignard and Others

The results of this analysis and of the ‘posting a letter’ error analysis were
strongly correlated: rho = – 0.79 (p = 0.0127) for the mean total score, and
rho = 0.79 (p = 0.0127) for the mean execution time; correlation approached
significance for the number of non-specific cues: rho = 0.58 (p = 0.068).

Script Generation

Analysis of Errors

For controls, the mean (SD) number of errors was 4.5 (2.5) for ‘shopping’,
7.3 (5.2) for ‘cooking’ and 3.4 (2.3) for ‘answering the letter’. The most
frequent types of errors were ‘omissions’, ‘commentaries’, ‘context neglect’ and
‘environmental adherence’ (see Tables III, IV and V).
Patients made a significantly (p < 0.01) higher number of errors for the three
scripts: 14.5 (8.3) for ‘shopping’, 22.2 (8.1) for ‘cooking’, and 13.1 (4.2) for
‘answering the letter’. The types of errors that best differentiated patients from
controls were ‘omissions’ and ‘commentaries’.

Centrality of Evoked Actions

The patients evoked significantly (p < 0.01) fewer central actions than
controls in each script. However, the percentage of total evoked actions that
were central did not differ between patients and controls.

Mean Evocation Time

The mean evocation time by action, time to generate a script from the
beginning to the end of the script divided by the number of actions generated,
was significantly (p < 0.01) increased in patients for the ‘shopping’ and
‘cooking’ scripts, but not for the ‘letter’ script.

Script Execution versus Script Generation

ANOVA with group as a between factor and condition (execution versus

generation) as a within factor was performed for each script. This analysis
confirmed the existence of a group effect: patients made significantly more
errors than controls whatever the script, condition or category of error (Tables
III, IV and V). It also showed a condition effect: the number of errors was
higher in execution than in generation, whatever the script or type of error
(Figure 4). Lastly, it underlined a significant interaction between group and
condition: the difference between execution and generation was greater in
patients than in controls, whatever the script, and this was particularly evident
for ‘inversion-substitution’ errors at the descriptive level, for ‘context neglect’
and ‘environmental adherence’ errors at the neuropsychological level.
No significant correlation was found between the total number of errors in
script execution and the total number of errors in script generation, whatever the
script or group considered. For controls the coefficients of correlation were
 Script execution 663

Fig. 4 – ‘Cooking’ script in patients with a dysexecutive syndrome: comparison of errors in

generation and in execution. 1: number total of errors. 2: types of errors at the descriptive level.
3: types of errors at the neuropsychological level. Results are expressed as mean (SD). * p < 0.05;
** p < 0.01; *** p < 0.001.
664 Mathilde Chevignard and Others

rho = – 0.05 (p = 0.88) for ‘shopping’, rho = 0.30 (p = 0.37) for ‘cooking’, and
rho = 0.50 (p = 0.13) for ‘answering the letter’. For patients the coefficients of
correlation were rho = 0.51 (p = 0.10) for ‘shopping’, rho = 0.20 (p = 0.53) for
‘cooking’, and rho = 0.12 (p = 0.67) for ‘answering the letter’.

Correlation between Script Execution and Performance of Tests of Executive

Functions

For patients, there was no significant correlation between the total number of
errors in the different scripts executed on one hand and the performance on tests
of executive functions on the other hand.

Correlation between Script Execution and Scores on the Behavioral Scales

For patients, there was no significant correlation between the various

numbers of errors in script execution and the global score on the Behavioral
Scale or the Dysexecutive Questionnaire answered either by the patient or by the
caregiver. For instance, for the total number of errors in execution of the
‘cooking’ script, the coefficient of correlation with the Behavioural Scale was
rho = 0.17 (p = 0.59) when answered by the patient and rho = 0.42 (p = 0.18)
when answered by the caregiver; the coefficient of correlation with the
Dysexecutive Questionnaire was rho = 0.05 (p = 0.88) when answered by the
patient, and rho = 0.29 (p = 0.36) when answered by the caregiver. By contrast,
the same coefficients of correlation became significant if the questionnaires were
answered by occupational therapists well acquainted with the patients. For the
Behavioral Scale, the coefficient of correlation was rho = 0.77 (p = 0.02) with
the total number of errors in script execution, rho = 0.77 (p = 0.02) with the
total number of errors in ‘cooking’, rho = 0.55 (p < 0.10) with the total number
of errors in ‘shopping’, and rho = 0.78 (p < 0.02) with the total number of errors
in ‘answering the letter’. For the Dysexecutive Questionnaire, the coefficient of
correlation was rho = 0.76 (p = 0.02) with the total number of errors in script
execution, rho = 0.74 (p < 0.03) with the total number of errors in ‘cooking’,
rho = 0.57 (p < 0.10) with the total number of errors in ‘shopping’, and rho =
0.82 (p = 0.01) with the total number of errors in ‘answering the letter’.

DISCUSSION

The first aim of the study was to compare execution and generation of
relatively familiar scripts. Script execution, which includes monitoring of action,
may be a better indicator of a planning dysfunction than script generation, which
investigates a specific category of knowledge and measures goal-related action
fluency. Script generation was, however, significantly impaired in patients
compared to controls, confirming previous studies implicating the frontal lobes
in managerial knowledge (Le Gall et al., 1993; Godbout and Doyon, 1995). As
in our previous study (Sirigu et al., 1995), two different cognitive modes of
representing actions could be hypothesised from an analysis of patients’
 Script execution 665

performance: (1) an associative level related to semantic knowledge and

centrality of actions, which was preserved in patients and could be under the
control of the temporal lobes; (2) a pragmatic level, using the goal of the plan
as a binding element between the actions, which was impaired in patients and
could require the involvement of the prefrontal cortex. The nature of pragmatic
errors, however, differed between the two studies: temporal ordering, included in
‘inversion-substitution’ errors, was not significantly impaired in the present
study. By contrast, evocation of actions was slowed and patients made more
‘omissions’ and ‘commentaries’. Differences in aetiology and location of lesions
may influence the nature of errors. In the previous study, the lesions were
mainly focal vascular and surgical lesions of the dorsolateral areas of the
prefrontal cortex. In the present study, the lesions were mainly traumatic. Being
more diffuse, the latter lesions could be responsible for cognitive slowing and
‘omission’ errors, because activation deficits and limitation of attentional
resources are reported to be more severe in diffuse than in focal lesions
(Schwartz et al., 1998). Involving orbital areas, the latter lesions could induce a
higher occurrence of ‘commentaries’, a sign of disinhibition associated with
orbital dysfunction (Luria, 1966; Eslinger and Damasio, 1985). Despite these
discrepancies, this study confirms the sensitivity of script generation to frontal
lobe damage, which disrupts the ability to mentally activate the correct sequence
of actions needed to achieve a goal, even for relatively familiar actions.
As planning impairment might result not only from a deficit of internal
representation, but also from a binding deficit between the plan and contextual
information, we wanted to study the same patients in the real execution of the
same scripts. It might be postulated that the performance would improve in a
contextual situation that could drive the patient and facilitate the binding
between actions. Conversely, our hypothesis was that this condition would
increase the deficit, since monitoring processes would be more highly solicited
by multiple sub-goal plans in open-ended situations, which would require the
performance of actions in parallel and suppression of distracters. Indeed, script
execution was more sensitive than script generation, both for controls and for
patients with dysexecutive deficits. This underlines the fact that planning should
not be restricted to managerial knowledge, but also includes the monitoring of
action (Hayes-Roth and Hayes-Roth, 1979). It might also suggest that planning
is not just a top-down predetermined sequential process (Grafman, 1989),
including cognitive (Shallice and Burgess, 1991) and affective (Damasio et al.,
1991) markers and able to reactivate after a delay previously generated
intentions; it would also be a bottom-up opportunistic process interacting with
the contextual environment to suggest decisions whenever promising
opportunities arise (Hayes-Roth and Hayes-Roth, 1979). This capacity for on-
line adjustment is critical for behavioral adaptation.
The difference in number of errors between execution and generation was
higher in patients than in controls, in agreement with our hypothesis that the
frontal lobes intervene not only in script representation, but also in the binding
between the plan of action and contextual environment. The fact that ‘context
neglect’ and ‘environmental adherence’ were errors that best differentiated
patients from controls at the neuropsychological level further supports this
666 Mathilde Chevignard and Others

interpretation. ‘Context neglect’ is commonly observed in laboratory testing of

patients with frontal lobe lesions. It accounts for disturbances in the ability to
analyze data, the first stage of problem resolution, that may be compensated for
by cues that direct attention to the pertinent information (Luria, 1966; Fasotti,
1992; Aubin, Le Gall and Etcharry-Bouyx, 1997). This is consistent with the
frontal lobes being an area of sensorial integration (Fuster, 1989; Pandya and
Yeterian, 1996) and of control of attention (Mesulam, 1990; Godefroy, Lhuillier
and Rousseaux, 1996). ‘Environmental adherence’ has also been described in
patients with frontal lobe lesions. It is responsible for pathological behaviors of
imitation, prehension and utilization commonly found in these patients
(Lhermitte et al., 1986). It expresses the release of parietal lobe activities from
frontal lobe inhibition, given that the dorsolateral prefrontal cortex and the
parietal cortex are involved in the same lateral system responsive to external
information (Barbas and Pandya, 1989). The weakening control of the
Supervisory Attentional System would predispose to environmental capture
(Shallice, Burgess, Schon et al., 1989; Miotto and Morris, 1998).
The ‘cooking’ script was the most sensitive to executive dysfunction (see
Tables III, IV and V). The total number of errors in execution of this script
increased from 16.9 in controls to 89.3 in patients and all types of errors were
significantly increased in patients compared to controls. In this script, in contrast
with the other two, some actions had to be performed in parallel: for example,
the scrambled eggs had to be prepared while the cake was being baked.
Therefore, the planning deficit might be increased by competing actions at a
decision node, i.e. when the subject has to set priorities in the face of two or
more competing tasks (Shallice and Burgess, 1991; Della Malva, Stuss, D’Alton
et al., 1993). As the increase in number of errors was particularly dramatic for
‘context neglect’, it may be suggested that contextual information may be
especially at risk when attentional resources are already required for the
selection of the appropriate action.
The deficit in binding plan and contextual information explains why the
execution of familiar multiple-action scripts in an opened and interactive
environment was more sensitive to a dysexecutive syndrome than laboratory
tests or script generation. The impaired performance in neuropsychological tests
was particularly mild in this study, even for planning tests such as the Tower of
London or the Six Element Tasks (see Table II). The reliability of the script
execution procedure is underlined by the strong correlation observed between the
performance on the three scripts on one hand, and between ‘posting the letter’
and the Executive Function Route-Finding Test on the other hand. Despite the
ecological value of script execution, the performance of patients was not
correlated with the answers of patients or caregivers on the behavioral scales.
Loss of insight is well known for patients with a dysexecutive syndrome who
commonly underestimate their deficits (Stuss, 1991; Prigatano, 1991). It is likely
the consequence of impaired autonoetic consciousness – the capacity to mentally
represent and become aware of subjective experiences in the past, present and
future (Wheeler, Stuss and Tulving, 1997) – and related reality distortion
(Johnson, 1997). Answers given by caregivers are usually more reliable (Sarazin
et al., 1998; Wilson et al., 1998), although an underestimation was also observed
 Script execution 667

in our study. The long duration of the disease may explain this underestimation:
(1) the caregiver implicitly adapts his behaviour to that of the patient and
progressively monitors the patient behavior, acting more or less as his own
frontal lobe; (2) the patient recovers a sufficient degree of autonomy to stay
alone at home with less control from the caregiver; (3) the patient’s errors
progressively become more similar to those produced by normal subjects and the
caregivers consider more the final result than the way of organizing the different
steps. Such an underestimation was not detected with more objective observers,
since significant coefficients of correlation were found with the scores answered
by the occupational therapist on the two behavioral scales which were aimed at
assessing the severity of frontal lobe-related behavioral abnormalities (Lhermitte
et al., 1986) and broad areas of dysexecutive changes (Wilson et al., 1998).
In conclusion, a continuous interaction with the environment may help to
shape an action plan, but also provoke action slips due to competing alternatives
for control and selection of action (Reason, 1979; Shallice, 1988). Action slips are
observed even in normal subjects. However, the impaired monitoring of action
associated with frontal lobe lesion or dysfunction results in both ‘context neglect’,
inducing a poorer action plan, and ‘environmental adherence’, which increases the
number of action slips, explaining why script execution is more sensitive than
script generation or laboratory tests. Therefore, script execution seems to be the
best predictor of the behavioral adaptation of patients with dysexecutive deficits
to daily life activities, which require a plan with multiple sub-goals to be executed
in open situations and its execution to be adapted to contextual information.
Although the method may appear time-consuming, it has the advantage of giving
a more exact appreciation of the patient’s deficits outside the laboratory and of
orienting rehabilitation in relation to the actual deficits of each patient.
Acknowledgements. This work was supported by a grant from the “Assistance Publique
– Hôpitaux de Paris” and a grant from the “Programme de Recherche Incitatif de l’IFR de
Neurosciences du CHU Pitié-Salpêtrière (Paris, France)”.

REFERENCES

AUBIN, G., LE GALL, D., and ETCHARRY-BOUYX, F. Niveaux cognitifs de la résolution de problèmes
numériques: À propos d’une observation de syndrome frontal. Revue Neuropsychologique, 7: 411-
433, 1997.
BAKER, S.C., ROGERS, R.D., OWEN, A.M., FRITH, C.D., DOLAN, R.J., FRACKOWIAK, R.S., and ROBBINS,
T.W. Neural systems engaged by planning: A PET study of the Tower of London task.
Neuropsychologia, 34: 515-526, 1996.
BARBAS, H., and PANDYA, D.N. Architecture and intrinsic connections of the prefrontal cortex in the
rhesus monkey. Journal of Comparative Neurology, 286: 353-375, 1989.
BENTON, A.L. Differential behavioural effects in frontal lobe disease. Neuropsychologia, 6: 53-60, 1968.
BOYD, T.M., and SAUTTER, S.W. Route-finding: A measure of everyday executive functioning in the
head-injured adult. Applications of Cognitive Psychology, 7: 171-181, 1993.
DAMASIO, A.R., TRANEL, D., and DAMASIO, H. Somatic markers and the guidance of behavior: Theory
and preliminary testing. In H.S. Levin, H.M. Eisenberg and A.L. Benton (Eds.), Frontal Lobe
Function and Dysfunction. Oxford: Oxford University Press, 1991, pp. 217-229.
DELLA MALVA, C.L., STUSS, D.T., D’ALTON, J., and WILLMER, J. Capture errors and sequencing after
frontal brain lesions. Neuropsychologia, 31: 363-372, 1993.
DUBOIS, B., LEVY, R., VÉRIN, M., TEIXEIRA-FERREIRA, C., and PILLON, B. Experimental approach to
prefrontal functions in Humans. In J. Grafman, F. Boller and K.J. Holyoak (Eds.), Structure and
Functions of the Human Prefrontal Cortex. Annals of the New York Academy of Sciences 1995;
179, pp. 41-60.
668 Mathilde Chevignard and Others

ESLINGER, P.J., and DAMASIO, A.R. Severe disturbance of higher cognition after bilateral frontal lobe
ablation: Patient EVR. Neurology, 35: 1731-1741, 1985.
FASSOTI, L. Arithmetical Word Problem Solving after Frontal Lobe Damage. A Cognitive
Neuropsychological Approach. Amsterdam: Swets & Zeitlinger, 1992.
FUSTER, J.M. The Prefrontal Cortex. 2nd ed. New York: Raven Press, 1989.
GARNIER, C., ENOT-JOYEUX, F., JOKIC, C., LE THIEC, F., DESGRANGES , B., and EUSTACHE, F. Une
évaluation des fonctions exécutives chez les traumatisés crâniens: L’adaptation du test des six
éléments. Revue de Neuropsychologie, 8: 385-414, 1998.
GODBOUT, L., and DOYON, J. Mental representation of knowledge following frontal-lobe or postrolandic
lesions. Neuropsychologia, 33: 1671-1696, 1995.
GODEFROY, O., LHUILLIER, C., and ROUSSEAUX, M. Non-spatial attention disorders in patients with
prefrontal or posterior brain damage. Brain, 119: 191-202, 1996.
GRAFMAN, J. Plans, actions and mental sets: Managerial knowledge units in the frontal lobes. In E.
Perecman (Ed.), Integrating Theory and Practice in Clinical Neuropsychology. Hillsdale, NJ:
Erlbaum, 1989, pp. 93-138.
GROBER, E., and BUSCHKE, H. Genuine memory deficits in dementia. Developmental Neuropsychology,
3: 13-36, 1987.
JOHNSON, M.K. Source monitoring and memory distortion. Philosophical Transactions of the Royal
Society of London, B 352: 1733-1745, 1997.
HAYES-ROTH, B., and HAYES-ROTH, F. A cognitive model of planning. Cognitive Science, 3: 245-262,
1979.
KARNATH, H.O., WALLESCH, C.W., and ZIMMERMAN, P. Mental planning and anticipatory processes with
acute and chronic frontal lobe lesions: A comparison of maze performance in routine and non
routine situations. Neuropsychologia, 29: 271-290, 1991.
KLOSOWSKA, D. Relation between ability to program actions and location of brain damage. Polish
Psychology Bulletin, 7: 245-255, 1976.
LE GALL, D., AUBIN, G., ALAIN, P.H., and EMILE, J. Script et syndrome frontal: À propos de deux
observations. Revue Neuropsychologique, 3: 87-110, 1993.
LEZAC, M.D., LE GALL, D., and AUBIN, G. Evaluation des fonctions exécutives lors des atteintes des
lobes frontaux. Revue Neuropsychologique, 4: 327-343, 1994.
LEVINE, B., STUSS, D.T., MILBERG, W.P., ALEXANDER, M.P., SCHWARTZ, M., and MACDONALD, R. The
effects of focal and diffuse brain damage on strategy application: Evidence from focal lesions,
traumatic brain injury and normal aging. JINS, 4: 247-264, 1998.
LHERMITTE, F., DÉROUESNÉ, J, and SIGNORET, J.L. Analyse neuropsychologique du syndrome frontal.
Revue Neurologique, 127: 415-440, 1972.
LHERMITTE, F., PILLON, B., and SERDARU, M. Human autonomy and the frontal lobes. Part I: Imitation
and utilization behaviours: A neuropsychological study of 75 patients. Annnals of Neurology, 19:
326-34, 1986.
LURIA, A.R. Higher Cortical Functions in Man. New York, NY: Basic Books Inc., 1966.
MESULAM, M.M. Large-scale neurocognitive networks and distributed processing for attention, language
and memory. Annals of Neurology, 28: 597-613, 1990.
MIOTTO, E.C., and MORRIS, R.G. Virtual planning in patients with frontal lobe lesions. Cortex, 34: 639-
657, 1998.
MORRIS, R.G., MIOTTO, E.C., FEIGENBAUM, J.D., BULLOCK, P., and POLKEY, C.E. The effect of goal-
subgoal conflict on planning ability after frontal and temporal lobe lesions in humans.
Neuropsychologia, 35: 1147-1157, 1997.
NELSON, H.E. A modified Card Sorting Test sensitive to frontal lobe defect. Cortex, 12: 313-324, 1976.
OWEN, A., DOWNES, J.D., SAHAKIAN, B.J., POLKEY, C.E., ROBBINS, T.W. Planning and spatial working
memory following frontal lobe lesions in man. Neuropsychologia, 28: 1021-1034, 1990.
OWEN, A.M., DOYON, J., PETRIDES, M., and EVANS, A.C. Planning and spatial working memory: A
positron emission tomography study in humans. European Journal of Neuroscience, 8: 353-364,
1996.
OWEN, A. Cognitive planning in humans: Neuropsychological, neuroanatomical and
neuropharmacological perspectives. Progress in Neurobiology, 53: 431-450, 1997.
PANDYA, D.N., and YETERIAN, E.H. Morphological correlations of human and monkey frontal lobe. In
A.R. Damasio, H. Damasio and Y. Christen (Eds.), Neurobiology of Decision-Making. Berlin:
Springer-Verlag, 1996, pp. 13-46.
PENFIELD, W., and EVANS, J. The frontal lobe in man: A clinical study of maximum removals. Brain, 58:
115-133, 1935.
PORTEUS, S.D. The Maze Test and Clinical Psychology. Palo Alto, CA: Pacific Books, 1959.
PRIGATANO, G.P. Disturbances of self-awareness of deficit after traumatic brain injury. In G.P. Prigatano
and D.L. Schacter (Eds.), Awareness of Deficit after Brain Injury. Clinical and Theoretical Issues.
Oxford: Oxford University Press, 1991, pp. 111-126.
REASON, J.T. Actions not as planned: The price of automatisation. In G. Underwood and R. Stevens
 Script execution 669

(Eds.) Aspects of Consciousness, vol 1. London: Academic Press, 1979, pp. 67-89.
REITAN, R.M. Validity of the trail making test as an indication of organic brain damage. Perceptual
Motor Skills, 8: 271-276, 1958.
SARAZIN, M., PILLON, B., GIANNAKOPOULOS, P., RANCUREL, G., SAMSON, Y., and DUBOIS, B. Clinico-
metabolic dissociation of cognitive functions and social behaviours in frontal-lobe lesions.
Neurology, 51: 142-148, 1998.
SCHWARTZ, M.F., MONTGOMERY, M.W., BUXBAUM, L.J., LEE, S.S., CAREW, T.G., COSLETT, B.H., et al.
Naturalistic action impairment in closed head injury. Neuropsychology, 12: 13-28, 1998.
SHALLICE, T. Specific impairments of planning. Philosophical Transaction of the Royal Society of
London B 298: 199-209, 1982.
SHALLICE, T. The Allocation of Processing Resources: Higher Level Control. From Neuropsychology to
Mental Structures. Cambridge: Cambridge University Press, 1988, pp. 328-352.
SHALLICE, T., and BURGESS, P. Deficits in strategy application following frontal lobe damage in man.
Brain, 114: 727-742, 1991.
SHALLICE, T., BURGESS, P.W., SCHON, F., and BAXTER, D.M. The origins of utilization behavour. Brain,
112: 1587-1598, 1989.
SIRIGU, A., ZALLA, T., PILLON, B., GRAFMAN, J., AGID, Y., and DUBOIS, B. Selective impairments in
managerial knowledge following prefrontal cortex damage. Cortex, 31: 301-316, 1995.
SIRIGU, A., ZALLA, T., PILLON, B., GRAFMAN, J., AGID, Y., and DUBOIS, B. Selective deficits in script
sequence and boundary analysis following lesions in the prefrontal cortex. Cortex, 32: 297-310,
1996.
STUSS, D.T. Disturbances of self-awareness after frontal system damage. In G.P. Prigatano and D.L.
Schacter (Eds.), Awareness of Deficit after Brain Injury. Clinical and Theoretical Issues. Oxford:
Oxford University Press, 1991, pp. 63-83.
STUSS, D.T., ESKES, G.A., and FOSTER, J.K. Experimental neuropsychological studies of frontal lobe
functions. In F. Boller and J. Grafman (Eds.), Handbook of Neuropsychology, vol 9. Amsterdam:
Elsevier Science Publishers, 1994, pp. 149-185.
VÉRIN, M., PARTIOT, A., PILLON, B., MALAPANI, C., AGID, Y., and DUBOIS, B. Delayed response tasks and
prefrontal lesions in man: Evidence for self-generated patterns of behaviour with poor
environmental modulation. Neuropsychologia, 31: 1379-1396, 1993.
WHEELER, M.A., STUSS, D.T., and TULVING, E. Toward a theory of episodic memory: The frontal lobes
and autonoetic consciousness. Psychology Bulletin, 121: 331-354, 1997.
WILSON, B.A., ALDERMAN, N., BURGESS, P., EMSLIE, H., and EVANS, J. Behavioural Assessment of the
Dysexecutive Syndrome. Bury St Edmunds, Suffolk: Thames Valley Test Company, 1996.
WILSON, B.A., EVANS, J.J., EMSLIE, H., ALDERMAN, M., and BURGESS, P. The development of an
ecologically valid test for assessing patients with a dysexecutive syndrome. Neuropsychological
Rehabilitation, 8: 213-228, 1998.

Bernard Pillon, Centre de Neuropsychologie, Fédération de Neurologie, Hôpital de la Salpêtrière, 47 Boulevard de l’Hôpital, 75651 Paris
cedex 13, France. E-mail: bernard.pillon@psl.ap-hop-paris.fr

(Received 13 July 1999; accepted 21 March 2000)