The ability to impute mental states with content to other people, that is a theory of mind (Premack & Woodruff, 1978), is present in normal children from at least four years of age (Wellman, 1990).
Theory of mind and emotion recognition can be seen as complementary activities (Steerneman, 1994 a,b) . Emotion recognition can be regarded as the ability to interpret one's own inner situation or that of others in a particular context.
The ability to recognise emotions can be seen, in fact, as a basic element of a theory of mind, a precondition of the development of a theory of mind. The understanding of emotion depends on an understanding of other mental states, particularly desires and beliefs (Harris, 1989).
Before concluding that the cognitive deficit is specific to the autistic child's theory of mind, it would be necessary to show that other aspects of social cognition are unimpaired in autism (Baron-Cohen ,1989; Happé, 1994). Social cognition can be defined as "the child's intuitive or logical representation of others, that is, how he characterizes others and makes inferences about their covert, inner psychological experiences" (Shantz, 1975, p.71).
Some aspects of social cognition are indeed unimpaired: object and person permanence, first-order belief, understandig the basic distinction between mental and physical phenomena, perceptual role-taking, self-recognition and peer-recognition. Leslie (1987) has proposed an explanation for the little insight of autistic children into the desires, beliefs and intentions that underlie people's actions and emotions. He suggests that autistic children have great difficulty with the "decoupling" from reality and non-truth. This decoupling or disengagement is needed both to understand mental states and to carry out pretence or symbolic play. There is evidence to suggest that individuals with autism can learn to develop some aspects of symbolic play (Jordan, Powell 1995).
Baron-Cohen (1989) suggests a specific developmental delay. We support Baron-Cohen's delay-hypothesis. There may be indeed a delay in the development of a theory of mind in autism, such that all children with autism are late in developing this ability but it may emerge or more specífic: it can be emerged in some cases.
We observed positive results in a few experiments in which imitation and perception skills of autistic children who participated in a social cognition training increased (Steerneman & Huskens, 1996; Steerneman, Muris, Merckelbach & Willems, 1996; Steerneman, in prep.).
Analysing these exmperiments we found two relevant aspects:
But we also found that these children appear to be more verbal: they have a higher verbal IQ and/or a higher verbal mental age and they are better at understanding thoughts behind words (see Happé 1994).
Verbal mental age is a necessary, but not a sufficient condition for the development of a theory of mind.
We think that a theory of mind develops in complex interactions between biological factors (e.g. intelligence, verbal mental age, severity of the disorder) and experiences in and with the environment.
We think that i''s possible to train or stimulate the development of a theory of mind.
Social insight and understanding of the social environment are often considered as important abilities which underlie socially skilled behavior. Despite this, few attempts have been made to develop training approaches in which social cognition, or more specifically, theory of mind, forms the starting point.
Steerneman (1994) has recently introduced an intervention program, that is a social cognition training, which is directly based on theory of mind ideas. In this workshop we report the use of the program with a group of children with PDDNOS and adults with autism.
In two earlier pilot-studies we measured the effect of the program.
In one experiment (Steerneman & Huskens, 1996) in 1992 we trained 5 autistic children (mean age: 8; mean intelligence (WISC-R IQ): 75) with the social cognition intervention program. We used a case control design and the training group did show (after 21 sessions on a weekly basis) significant improvement of social functioning compared with the (matched) control group.
There were indications found that the improvement in social behavior was caused by an improvement of social cognitive performance, especially in emotion recognition.
In another experiment (Steerneman, Jackson, Pelzer & Muris, 1996) in 1994 we trained 8 non-autistic children with social handicaps (mean age: 7; mean intelligence (WISC-R IQ): 95).
The training took place on a weekly basis and a period of seven month. In total the training group participated in 21 sessions. Apart from the training group two control groups (matched with the training group) also took part in the study.
Again, the training group showed clear improvements in social functioning. The study suggests that there is a positive relationship between social cognition in terms of theory of mind and social behavior and/or social skills.
Training: A theory of mind approach
The theoretical ideas concerning social cognítive development
offer a framework for the development of the social cognition
training; i.e. where spontaneous development is impaired, the
(guided) social cognition training takes over.
The program is directly based on theory of mind principles. Special
attention is focused upon a variety of learning situations: recognition
and assessment of the feelings of others, assessment of the context
in which particular information can be provided or not, assessment
of others'thoughts about oneself, recognition and understanding
of the motives and intentions underlying the behavior of others,
and understanding deception and humor. Central issue is the capacity
of the individual with autism to analyze behaviors of others by
recognizing mental states, such as desires and beliefs, which
underlie normal behavior (Wellman, 1990). The mental state "belief"
is a basic mental state that possesses intentionality. Wimmer
and Perner (1985) distinguish different dimensions of belief.
These dimensions are crucial and essential where analysing theory
of mind is concerned.
Wimmer and Perner (1983, see also Perner and Wimmer, 1985) distinguish between first order and second order belief. First order belief refers to causal and behavioral thinking, i.e. recognising that John knows something because he has observed it happen. Second order belief is more mentalistic in nature and concerns the childls capacity to think about thinking. At about three or four year children usually develop the capacity to use first order belief (e.g. Michael thinks that it is Sophie's birthday) . At about six years, social cognition is usually sufficiently developed to enable children to form second order beliefs (e.g. Michael thinks that Sophie thinks that he's angry with her) . They also argue that understanding of false belief develops at about the same time. Deceit and deception are central features of false belief: the child has to understand that a person who has been deceived will have a different view of a situation than his own. In a false belief situation one person has acces to specific information while another, whose reactions have to be forecast, possesses either partial or even mistaken information. Baron-Cohen, Frith and Leslie (see Frith 1989) investigated autistic children's conception of another person's mind and their ability to attribute mental states to other people.
They found that the majority (but not all) of the autistic children
were not able to understand that another person had been deceived.
In the famous Sally & Ann and pencilbox or false-belief experiments
the subjects had to discriminate between, on the one hand, the
actual location of an object and, on the other hand, the place
where a story character mistakenly believed that object to be.
The autistic children chose the real position of the object when
asked where the character thought it was. But failure on such
false belief tasks does not necessarily indicate that these children
do not understand or employ mental representatíons (Wellman
1990).
In testing and training a child's theory of mind it is essential
that we clearly distinguish all three different levels of belief:
first & second order and false-belief.
The training program sets out to operationalize the several dimensions
of belief (Perner & Wimmer, 1985; Wimmer & Perner, 1983).
Thinking about thinking and understanding of emotions are mediated
by understanding mental states such as desires and beliefs. This
principle is basic to the planning of the social cognition training.
The program is aimed at individuals with pervasive developmental
disorders and a developmental age of at least four years of age.
The training consists of 20 sessions (73 activities). First
attention is given to skills such as making someone's acquaintance,
describing oneself and listening to each other. Thereafter, detailed
information is given to perception and imitation: that is, visual
and aural perception, and verbal and motor imitation. A theory
of mind is approached using tasks aimed at learning to recognize
the difference between fantasy and reality, learning to assess
a social situation and learning to recognize the intentions of
others. Then, detailed attention is given to recognition of emotions,
that is happy, afraid, angry and sorrowful. Finally, attention
is given to the theory of mind itself. This involves activities
designed to develop self-related ideas, pretence, taking perspective
in the thoughts and feelings of others, as well as activities
to stimulate imagination and understanding of humor. The activities
(i.e. 20 sessions) become progressively more difficult as the
training proceeds. The number of activities occurring in a session
can be increased or decreased depending on factors such as age,
developmental level, the severity of the disorder. This flexibility
indicates that the approach can be tailored to the needs of the
group.
The effects of the social cognition training
Method
Subjects
An experiment has been designed to measure the effects of the
social cognition training. Participants are mentally retarded
(young) adults with autism and PDD-NOS (i.e. pervasive developmental
disorder not otherwise specified) (group 1-3) and children with
PDD-NOS (group 4).
Characteristics of these groups:
- group 1: 2 males, 2 females, mean age 40 (range 28-46), mean IQ (WISC-R) 52 (range 45-58);
- group 2: 5 males, 1 female, mean age 18 (range 15-20), mean IQ (WISC-R) 74 (range 61 -86);
- group 3: 3 males, 1 female, mean age 28 (range 24-31), mean IQ (WISC-R) 66 (range 64-67);
- group 4: 3 males, 2 females, mean age 11 (range 10-12), mean
IQ (WISC-R) 89 (range 70-116).
The majority of the mentally retarded adults is living in group homes. The rest of the adults and all the children are living with their parents.
The start of the study was in August 1995. In April 1996 the
study will be finished.
The study
A multiple baseline design across groups was used. There are,
as mentioned before, three groups of mentally retarded (young)
adults with PDD, one group with children with PDD-NOS, and a control
group. Every group consists of four to six participants. The
control group consists of mentally retarded young adults with
autism (all males, mean age 16 (range 14-18), mean IQ (WISC-R)
69 (range 60-78).
Measures
Tasks
We used social cognition tasks or Theory of Mind (TOM) tasks with
all participants during four testperiods: a pretest, two tests
in the course of the training, and a posttest. We used the "theory
of mind screening scale" (Steerneman, 1994b). This scale
offers an operationalization of the theory of mind concept and
incorporates a variety of tasks regarding beliefs and desires.
Attention is focused upon assessment and recognition of the thoughts
and feelings of others (first-order belief), assessment of situations
in which information can be given or not, recognition of deceit
and deception (false-belief) and the recognition of intentions
(second order belief).
Behavior recording
Using the video camera we recorded the social behavior of at least
one participant of each group except for the PDD-NOS group. The
recording took place during a social situation in the natural
setting, that is at dinner or coffee time. Once a week we recorded
the social behavior of the participants using partial interval
recording (5"5"), in that the observer has to observe
5 sec and has to record 5 sec. Each recording session lasted
25 min. The initiative of the interaction, affect and direction
were recorded. A primary and a secondary observer received guidelines
for recording. Guidelines and response definitions were discussed
and recording was practised. Reliability checks were conducted
in at least 30% of the recordings. At the time of printing this
manuscript observations are still taking place. Because of this
we cannot give reliability coefficients yet.
Procedure
The three mentally retarded training groups and the PDD-NOS group
received a one-hour training every week. During the training
two trainers and a coach were present. To be certain that effects
measured in the pilot study are training effects we controled
for bias of attention. That is, maybe the factor attention influences
the results. The control group received no training, but met
every week for one hour. During this weekly meeting hour an adult
and a coach were present.
It is hypothesized that:
a. the posttest taskscores of the training groups will be statistically significant increased in
comparison with the pretest scores;
b. the posttest taskscores of the control group will not be statistically sígnificant increased
in comparison with the pretest scores;
c. the training groups will show statistically significant increases in social behavior during
and after the training period in comparison with the baseline;
d. the control groups will not show statistically significant increases in social behavior
during and after the training period in comparison
with the baseline.
Results (1)
(1) Because observations and training of groups are still taking
place at the time of printing, the results are only preliminary
and concern only hypothesis a. and b. We hope to have more results
at the time the workshop takes place.
As can be seen in Table 1, training group 3 and the PDD-NOS training
group have a high score at the pretest. Both training group 1
and 2 gained higher scores at the posttest than at the pretest.
The PDD-NOS group gained also a higher score at test 3 than at
the pretest.
Using SPSSX, a MANOVA for repeated measures revealed significant differences between groups at the pretest (F (4,18)= 4.69 p = .009).
Then, using a t-test the differences between pretest and posttest scores for each group were examined.
The differences between these scores were not statistically significant for the first three mentally retarded training groups (respectively t = -1.05, df = 3, p = .372; t = -1.17, df = 3,
p = .325; t = .33, df = 5, p = .756).
Although the PDD-NOS group had not received a posttest yet, we
examined the difference in scores between the pretest and test
3. The score on test 3 was statistically significantly higher
than the score on the pretest (t = -4.79, df = 4,
p = .009).
All the four testperiods consisted of five different tasks; respectively
pretence, false belief, perspective-taking, emotion-recognition,
and perceptual role-taking.
In Table 2 the pretence taskscores of each group are shown. It
is reparable that all groups, except for the PDD-NOS group, receive
the highest scores on the pretest. A t-test revealed that
group 2 (t = 3.30, df = 3, p = .046) received statistically
significantly lower scores on the posttest than on the pretest.
In Table 3 the false-belief scores are shown for each group.
The three mentally retarded training groups received higher posttest
scores than pretest scores. The difference was statistically
significant for group 2 (t = -5.20, df = 3, p
= .014). However, the controlgroup received in test 2 also higher
scores than in the pretest. This difference was not statistically
significant (t= -1.70, df = 3, p = .188).
In Table 4 the perspective-taking scores are shown for each group. Again, except for the PDD-NOS group, all groups gained scores at the different testperiods. The gains in scores in group 1 and 2 were statistically significant (respectively, t =-5.33, df = 3, p= .013 and t = -4.53, df = 3, p = .020). The gain in score in the controlgroup was not statistically
significant (t =- 1.76, df = 3, p = .177).
The emotion-recognition taskscores are shown in Table 5. As can be seen in Table 5 the groups 2 and 3 and the PDD-NOS group had some gain over the testperiod.
However, these gains were not statistically significant (t
=1.65, df = 3, p=.198 for group 2; t= -.21, df =
5, p = .842 for group 3; t= 2.45, df = 4, p = .070).
In Table 6 the perceptual-role-taking scores are shown. As can
be seen in table 6 the three mentally retarded training groups
have lower scores at the posttest in comparison with the pretest.
This difference in pretest-posttest scores was statistically
significant for group 3; t = 3.87, df = 5, p
=.012.
In summary, the following prudent conclusions can be made:
1. The overall pretest-posttest taskscores of the three mentally retarded training groups did not statistically significantly differ from each other.
2. The difference between pretest and test 3 of the PDD-NOS group was statistícally significant, in that the overall scores on test 3 were higher. This gain was caused by a gain in emotion-recognition and perceptual role-taking.
3. Group 2 received statistically significantly lower scores on the posttest pretence taskscores.
4. False-belief scores of Group 2 were statistically significantly higher in the posttest.
5. Group 1 and 2 received statistically significantly higher perspective-taking scores in the posttest.
6. No statistically significant dífferences were found between pretest-posttest scores on emotion recognition.
7. Perceptual role-taking for group 3 was statistically significantly lower in the posttest.
8. In the testscores of the control-group no statistically significant
differences were found over two testperiods.
Table 1
Overall taskscores per group
| Group 1 | ||||
| Group 2 | ||||
| Group 3 | ||||
| PDD-NOS | ||||
| Control-group |
(*) Note: At the time of printing not all groups were tested.
Max score = l00.
Table 2
Pretence taskscores per group
| Group 1 | ||||
| Group 2 | ||||
| Group 3 | ||||
| PDD-NOS | ||||
| Control-group |
(*) Note: At the time of printing not all groups were tested.
Max score = 20.
Table 3
False belief-scores per group
| Group 1 | ||||
| Group 2 | ||||
| Group 3 | ||||
| PDD-NOS | ||||
| Control-group |
(*) Note: At the time of printing not all groups were tested.
Max score = 20.
Table 4
Perspective-taking scores per group
| Group 1 | ||||
| Group 2 | ||||
| Group 3 | ||||
| PDD-NOS | ||||
| Control-group |
(*) Note: At the time of printing not all groups were tested.
Max score = 20.
Table 5
Emotion-recognition scores per group
| Group 1 | ||||
| Group 2 | ||||
| Group 3 | ||||
| PDD-NOS | ||||
| Control-group |
(*) Note: At the time of printing not all groups were tested.
Max score = 20.
Table 6
Perceptual role-taking scores per group
| Group 1 | ||||
| Group 2 | ||||
| Group 3 | ||||
| PDD-NOS | ||||
| Control-group |
(*) Note: At the time of printing not all groups were tested.
Max score = 20.
References
Baron-Cohen, S. & Howlin, P. (1992). The theory of mind deficit in autism: some questions for teaching and diagnosis. In: S. Baron-Cohen, A.S. Tager-Flusberg & D. Cohen (eds). Understanding other minds: perspectives from autism.
Oxford: Basil Blackwell
Baron-Cohen, S. (1989). The autistic child's theory of mind: A case of specific developmental delay.
Journal of Child Psychology and Psychiatry, 30, 285-297.
Frith, U. (1989). Autism. Explaining the Enigma. Oxford:
Basil Blackwell
Happé, F. (1994). Autism: an introduction to psychological
theory. London: UCL Press.
Harris, P. (1989). Children and Emotion. The development of psychological understanding.
Oxford: Basil Blackwell.
Hobson, R. (1986). The autistic child's appraisal of expression
of emotion. Journal of Child Psychology and Psychiatry, 27,
321-342
Jordan, R. & Powell, S. (1995). Understandíng and
teaching children with autism. New York: John Wíley.
Leslie, A. (1987). Pretense and representation: the origins of
"theory of mind". Psychological Review, 94,
412-426.
Perner, J. & Wímmer, H. (1985). "John thinks
that Mary thinks that...". Attribution of second-order beliefs
with 5-10 year old children. Journal of Experimental Child
Psychology, 39, 437-471.
Premack, D. & Woodruff, G. (1978). Does the chimpanzee have
a theory of mind ? The Behavioral and Brain Sciences, 4,
515-526.
Schopler, E. & Mesibov, G. (eds). (1986). Social behavior
in autism. New York, Plenum Press.
Shantz, C. (1975). The development of social cognition.
In: E. Hetherington (ed). Review of child development research
(vol. 5). Chicago: Uníversity Press.
Sigman, M., Mundy, P., Sherman, T. & Ungerer, J. (1986).
Social interactions of autistic mentally retarded and normal children
and their caregivers. Journal of Child Psychology and Psychiatry,
27, 647-656.
Steerneman, P. (1994a). Denken over denken en leren begrijpen
van emoties. Groepsbehandeling van kinderen. [Thinking about
thinking and understandíng of emotions. Group treatment
of chíldren]. Leuven: Garant.
Steerneman, P. (1994b). Theory of mind-screening-schaal.
(Theory-of-mind-screening-scale). Leuven: Garant.
Steerneman, P., Jackson, S., Pelzer, H. & Muris, P. (1996).
Children wíth social handicaps: an intervention programme
using a theory of mind approach. Journal of Clinical Child
Psychology and Psychíatry. In press.
Steerneman, P. & Huskens, B. (1996). Die Entwicklung eines
sozialen Kognitionstrainings für autistische Kinder (The
development of a social cognition training for autistic children).
Zeitschrift für Psychologie in Erziehung und Unterricht.
In press.
Steerneman, P., Muris, P., Merckelbach, H. & Willems, H. (1996).
Assessment of development and abnormal behavior in children with
pervasive developmental disorders. Evidence for the reliability
and validity of the revised psychoeducational profile (PEP-R).
Journal of Autism and Developmental
Disorders. In press.
Steerneman, P. Imitation, perception, verbal mental age and theory
of mind. Manuscript in preparation.
Wellman, H. (1990). The child's theory of mind. London:
Mit Press.
Wimmer, H. & Perner, J. (1983). Beliefs about beliefs. Representatíon and constraining functioning of wrong beliefs in young children's understanding of deception.
Cognition, 13, 103-128.