The Neuropsychology of Autism and Pervasive Developmental Disorders -- The Extreme Male Brain Theory: An Expert Interview With Simon Baron-Cohen, PhD, MPhil

December 14, 2005

Editor's Note:
The debate over what causes autism spectrum disorders has often focused on thimerosal-containing vaccines or the measles-mumps-rubella vaccine. Simon Baron-Cohen, PhD, MPhil, has a different theory. To find out how this theory fits different types of data and how it is being tested, Randall F. White, MD, FRCPC, Medscape Psychiatry and Mental Health, interviewed Dr. Baron-Cohen, Director, Autism Research Centre; Professor of Developmental Psychopathology, University of Cambridge; Fellow, Trinity College, Cambridge, United Kingdom.

Medscape: You advance the proposition that people with autism have an imbalance in 2 cognitive capabilities, which makes them behave as though they are, at least in the cognitive realm, much more male than female.[1] What are these fundamental differences between men and women that may help us understand the psychology of autism and Asperger's disorder?

Simon Baron-Cohen, PhD, MPhil: In the general population, you find that, on average, males have a stronger drive to systemize and females have a stronger drive to empathize. Those are the 2 cognitive processes we've been focusing on to try to understand autism and why it should be more common among boys than girls. We've found that people on the autistic spectrum show an exaggeration of the male profile.

Medscape: Can you explain what systemizing and empathizing mean in this context?

Dr. Baron-Cohen: Systemizing is the drive to analyze or construct a system. It could be any kind of system -- a mechanical system such as a computer, or a natural system such as the circulatory system of the body. Or it could be an abstract system like mathematics. The key thing is that when you systemize, you identify the rules or the laws that govern that system in order to predict how it will behave.

Empathizing is completely different. It's about being able to imagine what someone else is thinking or feeling, and having an emotional response to the other person's feelings.

Medscape: In your writing, you have used the term "theory of mind."[1] Is that synonymous with empathizing?

Dr. Baron-Cohen: It's part of empathy. Theory of mind is being able to put yourself in somebody else's shoes, being able to imagine what's going on in his or her mind. But imagining someone else's thoughts or feelings is only part of empathy. The other part is having [an appropriate] emotional reaction. The distinction is important because a psychopath might be able to figure out somebody else's thoughts quite accurately but wouldn't necessarily have an appropriate emotional response.

Medscape: In your view, a continuum of empathizing and systemizing exists, but if a person has high capacity in one, does that necessarily mean low capacity in the other?

Dr. Baron-Cohen: Well, when we started working in this area, we set up a model as if these 2 things were independent of each other. But as we've tested more individuals, we've found there's some kind of trade-off, so that people who are good at one tend to be worse at the other.

Medscape: That's in normal people?

Dr. Baron-Cohen: Yes, in the general population, we find that there's a small negative correlation between the 2.[2] That's making us wonder whether they may depend on some common mechanism.

Medscape: A common neural substrate?

Dr. Baron-Cohen: Yes. They might compete in the brain for space; that would be a very simple way that you could end up with that result. Or it could be that they're both under the control of some third factor: it could be genes, it could be hormones.

Medscape: Can you provide some of the evidence that the cognitive profile of autism-spectrum disorders fits into this paradigm? What about the ability of people with such disorders to empathize?

Dr. Baron-Cohen: Many different tests have looked at empathizing. One example would be judging emotional expressions from photographs of faces; that is, what somebody else might be feeling. People in the autistic spectrum tend to find that very difficult.[3] They score significantly lower [than control subjects] on tests of emotion recognition, and that relates again to sex differences in the general population: women tend to do better in such tests than men.

A second example involves listening to speech and detecting if somebody said something hurtful. And again, people in the autistic spectrum find that quite difficult -- to judge if somebody said something hurtful or rude.[4] And in relation to sex differences, girls tend to be ahead of boys in developing this kind of social sensitivity.

Medscape: Can you now discuss the autistic capacity for systemizing?

Dr. Baron-Cohen: There are 2 ways we've thought about this. One is giving tests in, for example, physics, which is an area that requires systemizing -- problems to solve that involve mechanics, for example. We've found that both children and adults with Asperger's disorder score above average on that sort of test.[5]

The other way we've been thinking about it is in the low-functioning individuals with autism (children with learning disabilities and very little language). They have very strong obsessions and do things in a very repetitive way, and those behaviors might also signify strong systemizing. An example might be a child with autism who wants to line things up repetitively or spin the wheels on a toy car for hours. Although they may not be demonstrating any special talent, they nevertheless become totally preoccupied with a very specific local system.

Medscape: In the editorial you recently published in the Archives of Diseases in Childhood, you suggest that the extreme male brain hypothesis can explain why mental retardation is often associated with autistic disorder.[6] Can you discuss that?

Dr. Baron-Cohen: It is a puzzle as to why these things should co-occur. And the first thing to say is that we don't actually know what the rates of mental retardation are in autism. In the old days, the figure given was that 75% of children with autism also had below-average IQ. That's quite high.

These days, the rates of autism are now thought to be 1 in 200 children.[7]

Medscape: In that would you include pervasive developmental disorder not otherwise specified?

Dr. Baron-Cohen: Yes, exactly, and Asperger's and high-functioning autism. So if the whole spectrum is actually more common [than previously thought], it may turn out that the association with mental retardation is not as strong as we previously thought.

One possibility is that the stronger your drive to systemize, the narrower your focus is on the environment. Imagine systemizing on a continuum, and if you're at the highest point of this continuum, it could be that you become so focused on just one system that it delays your learning about the environment more widely. The child with autism who plays the same video over and over again to totally predict the video as a system may become an expert at one thing and fail to learn about even quite ordinary other things, and so may appear to have learning difficulties.

I think the group with what you call mental retardation and what we call in this country learning difficulties is not very well studied. One of the predictions of this new theory is that even those individuals with low IQ who have autism have good systemizing skills. We just need methods to test them.

Medscape: You also suggest that assortative mating may play a role in the cause of autism-spectrum disorders.[6] Can you explain?

Dr. Baron-Cohen: Assortative mating is a familiar concept in genetics. It describes the idea that 2 individuals may be more likely to reproduce together because they share some common characteristics. We know this is true for height -- people tend to choose partners similar in height to themselves. The question is, could there be assortative mating based on systemizing among the parents of children with autism?

We have some preliminary clues that this is, in fact, going on. The first piece of evidence is that both parents of children with autism have very good attention to detail using a test called the embedded figures test.[8] The second clue is that both parents are likely to have fathers who work in the field of engineering, which involves very good systemizing.[9] This suggests that good systemizing may run in both sides of the family, a suggestion of assortative mating in the parents.

The last piece of evidence is a very recent study in which we've been using MRI to look at functional brain activity in mothers and fathers of children with autism and finding that there are areas of the brain in which both parents show what you might call hypermasculinization.[10] Where you expect to find sex differences in the general population in the brain, you find that these mothers and fathers are showing an extreme of the male profile.

I don't think there's much controversy that autism is genetic. There's a lot of consensus around that. The question is which genes are important and what those genes are for.

Medscape: One writer has speculated that the increase in autism- spectrum disorders in Silicon Valley in California is a consequence of computer scientists having children together.[11] What do you think of that notion -- is it a testable hypothesis?

Dr. Baron-Cohen: Yes, I'm aware of it. I think that most of the evidence is anecdotal. People claim that the rate of autism is very high in Silicon Valley, but I haven't seen any good studies looking at it.

It shouldn't be difficult to test, but to do it properly you need to have large samples. We're trying to test it, collecting data in areas where either both parents are strong systemizers, for example, working in the computer industry, or where just one of the parents is working in that field, or where neither parent works in that field. [We will look] at the relative rates of autism among those different combinations of couples.

Medscape: So you are involved in such an epidemiologic study now?

Dr. Baron-Cohen: Yes, and it's just taken quite a lot of time to set up the study, but we hope to have results this year.

Medscape: Is that being done in England?

Dr. Baron-Cohen: It's being done in the United States and in England because to be able to generalize the findings, it helps if you can show it's not specific to one locality.

Medscape: You have undertaken the study of fetal testosterone exposure and its effects on language and social abilities in children. What have you found that corroborates your hypothesis?

Dr. Baron-Cohen: [In] women who are having amniocentesis for clinical reasons, we simply asked for their consent to analyze the amniotic fluid for testosterone, then we followed up the baby after birth to see how hormone levels relate to behavior. We've found that the higher the child's fetal testosterone, the less eye contact the child makes at 12 months of age[12] and the slower it is to develop language at 18 months old.[13]

The same children have been followed up at 4 years old. We've looked at their social skills and also at "narrow interests." The results showed that the higher the fetal testosterone, the more social difficulty the child was having at school and the narrower the child's interests.[14] It relates to whether there might be some relationship between systemizing and empathizing, because narrow interests could be related to systemizing. When you systemize, you really need to zero in on small details, and obviously social skills very much relate to empathizing. What we're finding is that they both seem to have a relationship with fetal testosterone.

Medscape: Fragile X syndrome is a medical condition known to be associated with autistic disorder. Might that association somehow illuminate the neuropsychology of the extreme male brain?

Dr. Baron-Cohen: The X chromosome is an area of a lot of interest. A study found that if you look at girls with Turner's syndrome according to whether they have their X chromosome from the father or the mother, there's a difference in their social skills.[15] As the genetics get better understood, it may well be shown that the X chromosome plays an important part in autism and even in individual differences in empathy. We also know that the X chromosome contains genes that regulate testosterone. I think that the genetic and hormonal explanations may not be mutually exclusive.

Medscape: So is it possible that an extreme male brain may be an end product of several pathways?

Dr. Baron-Cohen: Yes, that sounds possible. We've been investigating one mechanism, which is fetal testosterone, and we've been looking at a relatively small sample of [nonautistic] children, about 70, and just at individual differences. To test if that's got any relevance to autism or an extreme male profile, you'd need a much larger sample. That's what we're doing at the moment, by trying to collect thousands of samples of amniotic fluid so that we can follow up to see who develops autism and if their fetal testosterone was elevated. I want to make it clear that hormones are not necessarily the only factor. They may be interacting with genes, for example.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.