Baron - Cohen Revised Eyes Test 2001

Front 1

Background of study

Back 1

One feature of autism is the lack theory of mind. This was first tested by Baron-Cohen et al. in 1985 using a procedure called the sally anne test. In 1997 Baron-Cohen devised a theory of mind test for adults called the eyes test. A number of methodological weaknesses arising from this test were resolved in the revised versions.

Front 2

Identify additional aims of the study

Back 2

- To test a group of adults with AS\HFA on the revised version of the test. This was in order to check if the deficit in this group of patients that had been found on the original version and related tests could be replicated.

- To test if in a sample of normal adults, an inverse correlation would be found between performance in the eyes test and the autism spectrum quotient.

- To test if the sex difference (female superiority) found on the first version of the test replicated.

Front 3

Describe the original test

Back 3

In this test, the participant in presented with a series of 25 photographs of the eye region of the face of different actors and actresses, and is asked to choose which of the 2 words best describes what the person in the photograph is thinking or feeling

Front 4

Describe the results of the original test

Back 4

- Adult males in the general population scored a mean of 18.8, whereas women scored a mean of 21.8 out of 25

- Adults with AS\HFA scored 16.3 out of 25

- Adults with Tourettes syndrome scored 20.4 out of 25. They did not differ significantly on this test from the general population

Front 5

Define the term “Theory of mind”

Back 5

It is the ability to attribute mental states to oneself or another person, and his mental ability is the main way in which we make sense of or predict another person’s behaviour.

Front 6

What were the problems with the original eyes test and how were they solved in the revised version?

Back 6

- There was a forced choice of only 2 words for each set of eyes, meaning the answer could be a 50/50 guess.

- Four words were added in the revised test, one should therefore score 17 or above out of 25 to be significantly above chance. This meant that the range of scores in which the test can reveal individual differences whilst still being above chance is only 9 points.

- Parents of children with AS/HFA scored below the general population as well (fathers scoring on average 17.3 and mothers 18.9) even though they did not have the condition. This highlights that the test has too narrow a range of scores to be able to distinguish between someone with the lesser variant/broader phenotype (e.g in a first degree relative of someone with autism) and someone with the condition itself.

- The test did not differentiate widely enough as scores covered a very narrow band.

- The revised test had 40 (reduced to 36) not 25 items.Ceiling effects (too many at the top end of the mark range) were observed with too many people scoring too highly. Ceiling effects are obviously undesirable because one loses power to detect individual differences. Having 36 items and 4 words aimed to remove the ceiling effect.

- Some test items were too easy, again causing ceiling effects. This revised version had fewer easy items.Some items were guessed by checking gaze direction (e.g noticing or ignoring). These items did not assess mental states so were excluded from the revised version.

- The original version had more female faces than male. In the revised version there were equal numbers.

- In the original the semantic word and foil (the other word) were semantic opposites, such as sympathetic and unsympathetic. Again this could be too each and contribute to ceiling effects. If three words are not semantic opposites then the level of difficulty increases and ceiling effects should be removed.

- The words used might not have been understood by all participants (i.e comprehension problems) In the revised version a glossary of terms was included, which participants could refer to.

Front 7

Identify the five predictions of the study

Back 7

1. AS/HFA will score significantly lower on the eyes test than other groups, but be unimpaired on the gender control judgements

2. AS/HFA will score significantly higher on the AQ test than other groups.

3. Normal females (groups 2 and 3) will score higher than males on the eyes test

4. Normal males will score higher than females on the AQScores on the

5. AQ and eyes will be inversely correlated

Front 8

Describe participants

Back 8

Group 1: Adults with high functioning autism or AS. Recruited from an autistic society magazine advertisement. All male. N=15. IQ score mean= 115

Group 2: Normal adults (the control group) selected from community classes in Exeter or public libraries users in Cambridge. From a range of occupational classes and educational levels. N=122

Group 3: Normal adult students. Cambridge undergraduates 71 in science 32 in other subjects. With much higher IQ than average. N=103 ( male=53, female=50)

Group 4: Normal adults, IQ matched (mean = 116) controls with group 1, randomly selected from the population, N=14.

Front 9

Describe the eyes test

Back 9

Consisted of 36 black and white photographs of different male and female eye regions taken from a magazine.

The images were all the same size (15X10 cm).

Each photograph has four words that describe the mental state of the person.

A participant is presented with a set of eyes and 4 words and is asked ‘which word best describes what this person is feeling or thinking’ for example if the target word was ‘serious’ the foils might be ‘ashamed’, ‘alarmed’ and bewildered’.

One answer is correct and the other 3 are incorrect.

After completing all 36, the total number of the overall score is added to give an overall score.

A glossary of all mental state terms was included

Front 10

How were the target and foil words for the eyes test selected?

Back 10

The words for each set of eyes (target and foils) were in initially chosen by 2 of the authors and judged by 4 male and 4 female member team. At least 5 of the 8 judges had to agree that a particular word was the correct one. As a check, at least 50% of participants from group 2 and 3 had to choose the correct word. Four items did not achieve this level of consistency and so were dropped. Note that the original test had 40 items, reduced to 36. This means that there was a check (of validity) to see if the correct word really was describing the mental state.

Front 11

Describe the AQ test

Back 11

The AQ test measure the degree to which any individual adult of normal IQ posses traits related to the autistic spectrum.

A self report questionnaire. The higher the score the more autistic trait a person possesses.

Consisted of 50 statements to which a participant chooses one of 4 answers, which are always ‘definitely agree’, ‘slightly agree’, ‘slightly disagree’, and ‘definitely disagree’.

In reality there is no difference between ‘definitely’ and ‘slightly’ in scoring, simply 1 point is given to ‘agree’ on half of the items and point to ‘disagree’ on the other half.

The total score is out of 50AQ test is both reliable and valid

Front 12

Design, Apparatus and Procedure

Back 12

Design:Groups 1 and 4 were matched but each group was independent.

Apparatus: The AQ, the eyes test and a quiet room in Cambridge/exeter.

Procedure:

All 4 groups were given the eyes test to complete in a quiet room in Cambridge or Exeter.Participants in groups 1, 3 and 4 were given the AQ.

Participants in AS/HFA group were also asked to judge the gender of each person in each photo,, as a control task , given anticipated impairments on mental state recognition.

Participants were asked at the outset to read through the glossary and indicate any word meanings they were unsure of.

Front 13

Describe data collected

Back 13

Quantitative data were gathered because both the eyes test and the AQ test gave numerical scores. Participants were not asked ant ‘why do you think..’ questions or open-ended questions.

Front 14

Explain findings and how they supported the 5 predictions

Back 14

The distribution of the eyes test scores showed no score lower than 17 and none higher than 35 (scale of 0-36), and most participants were normally distributed with the modal score being 24. A one way ANOVA was used to compare the 4 groups.

The mean eyes test score was lowest for group 1 at 21.9 and this group scored significantly lower than all of the other groups, means of 26.2 for the general population group, 28 for the students, and 30.9 for the IQ matched controls. So prediction 1 is supported.

Sex differences on the eyes test were examined in group 2 and 3 using ANOVA group X sex and females did score higher than males but this difference was not significant. Prediction 3 is supported.

For the AQ text, group 1 (mean of 34.4) scored significantly higher than groups 3 and 4 ( means of 18.3 and 18.9). Prediction 2 is supported.

Sex differences on the AQ test were examined in groups 2 and 3 and males (19.5) did score higher than the females (16.6) but this difference was not significant. Prediction 4 is supported.

There was a significant (p=0.004) inverse (negative) correlation of -0.53 between the AQ and the eyes test scores. Prediction 5 is supported.

Front 15

Describe the stages of attribution of theory of mind and what stage does the eyes test include?

Back 15

The two stages of attribution of theory of mind are:

First stage: Attribution of the relevant mental state (e,g compassion)

Second stage: Inferring the content of that mental state (e.g. compassion for her mother’s loss).They eyes test only involves the first stage.

Front 16

Strengths of the study

Back 16

experimental validity: measures an autistic trait not a normal one.

Question criteria

High control over extraneous variables

Reliable and replicable

Front 17

Weaknesses of the study

Back 17

Low ecological validity: because there are only static pictures of just the eyes region. It is also an abnormal thing for people to do.

Front 18

Other results

Back 18

The results basically replicated those of the 1997 original test, the adults with AS\HFA performed poorly compared with normal adults

The revised test improved the power to detect subtle individual differences in social sensitivity in otherwise normally intelligent adults.

Front 19

Other applications of study

Back 19

- The revised eyes test may be relevant to clinical groups beyond those on the autistic spectrum (e.g., brain damaged patients following amygdalectomy or prefrontal cortical lesions). The researchers also developed a child version of the test.

- The AQ is not diagnostic but may serve as a useful instrument of quantifying the extent of an individual’s “caseness” in terms of AS\HFA, measuring personality traits.

Front 20

Outline an advantage and a disadvantage for using static pictures in the study

Back 20

Disasdv: Even with the new modifications the pictures are static whereas the real world never is and it’s of the eye regions only.

Adv: Static stimuli however make the test quick and easy to use, since it can be administered as a “pencil and paper” test.

Front 21

Why is it important to develop tests that are sensitive to subtle cognitive dysfunction?

Back 21

In neurodevelopmental conditions such as AS or HFA, deficits in social cognition may persist across the lifespan. There may however be camouflaged as a result of learning compensatory strategies. Without a subtle and sensitive test, the investigator may erroneously conclude that the patient is recovered or normal.

Front 22

Conclusions

Back 22

This study replicates the results found in the original eyes test - that AS/HFA participants are significantly impaired compared with non-AS/HFA participants

The modifications made in this test, compared with the original version, improved the test ‘in that the same weaknesses were not observed’

This test helps to validate the eyes test as a useful tool for identifying impairments related to AS/HFA.

All the initial aims were met and all the predictions were confirmed.