Skip navigation

Calls to Action

Q&A with Autism Pioneer Simon Baron-Cohen

The world-renowned researcher discusses his ground-breaking theories and development of autism support services and educational products
April 13, 2015

World-renowned developmental psychologist Simon Baron-Cohen directs the Autism Research Centre at the University of Cambridge, where he also teaches and does research. In addition to more than 400 scientific papers, he has authored many books including Mindblindness and Teaching Children with Autism to Mindread. Most recently, he produced the DVDs Mind Reading and The Transporters to help children with autism learn emotion recognition.

Dr. Baron-Cohen sat down to answer our questions about his career, famous ideas about autism and current projects.


Autism Speaks: What drew you to work with children affected by autism three decades ago?

Dr. Baron-Cohen: Little was known about autism when I worked as a teacher in a small psychiatric unit for kids with autism in north London from 1981 to 1982. Called Family Tree, it was influenced by the psychoanalytic theories of the day [that autism was caused by emotionally detached parenting]. I remember Frances Tustin came to visit. But it was also an open-minded, experimental school. For example, the Nobel Laureate and ethologist Niko Tinbergen also visited.

The staff-pupil ratio was one-to-one, with just six kids and six teachers. Our work with the kids was very interactive. In addition to being a teacher, I also drove the mini-bus and often took the kids out to the park. We had video cameras in every room. At staff meetings, we analyzed the videos and discussed why an interaction with a child had gone well or ended unexpectedly with a tantrum or the child running out of the room. The head teacher, June Felton, was pioneering new educational approaches. We were in unchartered territory.

I found the puzzle of autism intriguing, particularly the dissociation between the children’s frequently precocious non-social intelligence, alongside delays in social understanding.

Inspired by that year of teaching, I decided to pursue doctoral studies with Uta Frith, at the Medical Research Council’s Developmental Psychology Unit, attached to University College London. Dr. Frith had studied for her own PhD with Beate (Ati) Hermelin, a Berlin psychologist who fled from the Nazis in 1939.

Hermelin’s 1970 book Psychological Experiments with Autistic Children exemplified how to analyze human cognition and perception through elegant experiments. This approach is in danger of vanishing from contemporary autism research. Hermelin’s own doctoral supervisor was Neil O’Connor, who had worked with the Soviet neuropsychologist Alexander Luria, who himself had worked with another Soviet social-developmental psychologist Lev Vygotsky, who came from my grandparents’ home of Belarus.

As Uta Frith’s doctoral student, I felt privileged to be part of an academic lineage stretching back to early 20th century Russian psychology.

Autism Speaks: You’re famous for your “theory of mind” research. What are its practical applications?

Dr. Baron-Cohen: During my year as a teacher, I became fascinated by the possibility that children with autism didn’t think about what other people were thinking. For example, a child would come right up close to someone else’s face, unaware that the other person might consider this intrusion odd or unsettling.

One boy, age 14, grabbed my glasses and threw them across the room because he didn’t like gold-framed glasses. My startled look of surprise seemed of no concern to him. A 13-year-old girl wandered into her parents’ dinner party with no clothes on, apparently unaware of what the guests were thinking. This absence of embarrassment in teenagers with autism gave me the clue that their “theory of mind,” or awareness of other people’s thoughts, was not developing typically.

Another child would ask me the same set of questions every day (“Is your birthday on a Tuesday this year?”) despite knowing the answer. She wasn’t using language to impart or get new information, and she wasn’t concerned about what I thought about her question. She was simply repeating statements to confirm the factual patterns that interested her.

My PhD proposal was to test whether this aspect of social cognition – understanding what others were thinking – might be delayed or impaired in autism. This led Uta Frith, my co-supervisor Alan Leslie and me to develop the “Sally Anne Test” of false belief. (See image at right.) In this now widely replicated test, the participant is told a story in which Sally watches Anne put a marble in a basket. Sally leaves, and Anne moves the marble to a box. Sally returns to look for the marble. The participant is then asked where Sally will look for her marble.

The results of this experiment confirmed our theory of an autism-related inability to infer other people’s mental states: Most children with autism predicted that Sally would look for the marble in its new location. That is, they failed to appreciate that Sally didn’t know that Anne had secretly moved the marble. By contrast, most typical 4-year-olds predicted that Sally would look for the marble in its old location – evidence that they appreciated what Sally knew and didn’t know.

You asked about the clinical value of this basic cognitive science. We went on to publish Teaching Children with Autism to Mindread, showing that you can break “theory of mind” down into steps that children with autism can learn. We also produced a DVD called Mindreading, a fun project that involved putting every human emotion – there are 412 of them! – into an electronic encyclopedia and computer game that children and adults with autism can use to learn to recognize what another person might be feeling.

The idea of learning about emotions via a computer may seem counterintuitive to how children typically learn social skills. But computer-based learning may actually be preferable for many children and adults on the autism spectrum. We found that, with practice, they can improve their emotion-recognition skills. The DVD turned out to be an excellent example of creating something practical from a theory.

We created another useful application – the Checklist for Autism in Toddlers (CHAT) based on the precursors of theory of mind – pretend play and joint attention. We see these skills emerge by 14 months of age in typically developing infants. The CHAT screens for autism at 18 months based on the absence of these milestones. In 1992, we tested this tool in the first “baby sibs” study. In 1996, we confirmed the results in a larger population of infants. In doing so, we showed that infants could be diagnosed with autism as young as 18 months of age. At that time this was a big step forward for earlier diagnosis.

We have just revised the CHAT to make it more relevant to the idea of the spectrum, or degrees of autism. The new instrument is called the Q-CHAT (Quantitative Checklist for Autism in Toddlers), and we’ll be publishing the results later this year.

Autism Speaks: You’re also known for your “extreme male brain” theory of autism. Please tell us about its development and your related theory of typical sex differences.

Dr. Baron-Cohen: The extreme-male brain theory of autism comes out of the “empathizing-systemizing theory” of typical sex differences. Prior to this, the field of sex differences research had focused on verbal versus spatial abilities as the only important sex differences. The empathizing-systemizing theory argued that, as a group, girls typically develop empathy at an earlier age than boys do. Empathy includes the ability to recognize other people’s thoughts and feelings and the drive to respond with appropriate emotions. Conversely, research suggests that boys, as a group, typically have a stronger interest in rules and patterns, analyzing and creating rule-based systems.

The extreme-male brain theory simply proposes that people with autism show an extreme of the typical male profile: They have intact – or even above average – systemizing drive, alongside below-average empathy.

It’s important to note that people with autism don’t struggle with all aspects of empathy. Most are caring people. For example, they become upset when they hear or see others suffering. We call this “affective empathy.”

By contrast, people with autism tend to struggle to read, or interpret, other people’s thoughts, intentions and feelings. We call this “cognitive empathy,” or theory of mind.

In 2014 we published a very large study using two scales – the Empathy Quotient and the Systemizing Quotient – with more than 800 adults with autism. We also included a control group of nearly 4,000 typical adults. The results precisely fitted what we would have predicted from our extreme-male brain theory.

This didn’t stop at least one journalist calling the theory “controversial”! I hope one day it won’t be contentious to discuss psychological sex differences. But I imagine it still triggers this reaction because of the valid concern that research into sex differences could be used to stereotype people. If you keep in mind that the theory is about group averages, not about individuals, this reduces the risk of misunderstanding.

The extreme male brain theory is occasionally been misunderstood in another way. Some wrongly interpret it as stating that people with autism are hyper-male in all respects, including in aggression. This misinterpretation makes clear that someone hasn’t actually read the theory. It specifically relates only to empathy and systemizing, not to all sex-linked characteristics.  

On the positive side, the extreme male brain theory has led to some novel investigations into the biological mechanisms that create typical sex differences. These include varying levels of prenatal sex hormones and the role these may play in autism. (Editor’s note: More on Dr. Baron-Cohen’s prenatal hormone research below.)

Autism Speaks: Your playful animated series “The Transporters” was nominated for a British Academy of Film (BAFTA) award in the UK, and clinical trials show it leads to improvements in emotion recognition and understanding. How can readers find out more?

The Transporters DVD is another practical product developed from our theories and research findings. Each animated episode features vehicles including trains, trams, tractors and cable cars with expressive human faces. We know that many kids on the autism spectrum love watching wheels and trains going round and round. They’re drawn to the repetition and predictable systems.

In The Transporters, we grafted the moving images of actors’ faces onto the front of each vehicle. This is important because we know that young children with autism look less at faces and make less eye contact than typically developing children do. So we’re harnessing their strong interest in systems to help them with their area of difficulty in cognitive empathy. With The Transporters, we provide enjoyable exposure to facial expressions, to teach which emotions go with different situations.

In short, The Transporters provides opportunities for social learning in an autism-friendly format. One-hundred percent of the profits from the DVD sales go to autism research through the Autism Research Trust, a charity that’s pleased to provide a special discount to the Autism Speaks community.

Watch a demo episode at

To get your 30 percent discount (valid through the Dec. 30, 2015), enter the discount code LIUB at checkout.

Autism Speaks: You recently published a study showing elevated rates of suicidal thoughts among those affected by Asperger syndrome. And like Autism Speaks, you’ve long championed the need for services and support for adults on the spectrum. What do you see as the most urgent needs?

Dr. Baron-Cohen: In 1999, we opened one of the first clinics for the diagnosis and support of adults with Asperger syndrome. We knew there was a ‘lost generation’ of people born before 1994, when Asperger syndrome was first recognized in the English-speaking medical world.

Editor’s note: In the United States, the once-separate diagnosis of Asperger syndrome was folded into the umbrella diagnosis of “autism spectrum disorder” with the publication of the DSM-5 diagnostic manual.

As you mention, we reported in the Lancet Psychiatry last year that more than half (66 percent) of 374 patients with Asperger syndrome, seen at our clinic, had suicidal thoughts or feelings. More than a third (35 percent) had made plans to commit suicide or had actually attempted to do so. This was an urgent a wake-up call. We need to develop support services for these adults. Their suicidality probably reflects feelings of being marginalized, isolated, lonely and under-valued.

We know, for example, that many adults with autism and Asperger syndrome are unemployed or under-employed. We know the importance of employment in promoting mental health, self-esteem, self-confidence and a sense of inclusion. With small reasonable adjustments to the work place, many people with autism can be an asset at work.

We are pleased to be working with the multinational company SAP to create internships for people with autism who are gifted in computer programming. The aim is that these internships will lead to permanent jobs. For SAP, it’s not just about social responsibility. The company’s leadership understands that people with Asperger syndrome can do the job as well – and sometimes even better – than a neurotypical person.


Video courtesy the University of Cambridge.


For North American autism employment resources, see
Autism Speaks Joins National Funders to Expand Inclusion in Employment.”

Autism Speaks: You’re now studying hormone levels – particularly levels of the steroid testosterone – in the womb. How is this advancing our understanding of the early development of autism?

Dr. Baron-Cohen: We’ve known since the 1940s that, in laboratory animals, high levels of the hormone testosterone masculinize the brain and body. Further, we know that this effect occurs only during a critical period of prenatal development. In humans, it’s between 9 and 24 weeks of gestation. Amniocentesis, a clinical test used to detect fetal chromosomal abnormalities, is typically done during this period of a woman’s pregnancy. This allows us to study hormone levels in the amniotic fluid without the mother having to go through additional testing. We began our fetal testosterone testing in the late 1990s. Since then, we’ve been following these children and their behavioral development.

Consistent with the fetal sex steroid theory, we found that higher levels of fetal testosterone correlated with a child making less eye contact at the child’s first birthday. It also correlated with slower language development at the child’s second birthday. Still later in development, higher fetal testosterone correlated with poorer social skills and lower cognitive empathy, or theory of mind.

Interestingly, our analysis also associated higher fetal testosterone with stronger interests in systems, better attention to detail, narrower interests and a higher number of autistic traits.

So we found associations between high levels of one hormone – testosterone – during a critical window of fetal brain development with both social and non-social aspects of cognition and behavior.

As a group, male fetuses produce twice as much testosterone as females do. Taken together with our findings, this may explain why autism is more common in males.

Our findings are important because fetal testosterone can change brain structure in a permanent way. For example, fetal testosterone slows down the pruning of connections – also known as apoptosis – between nerve brain cells. We know, from studies of donated post-mortem brain tissue, that autism often involves an excess of neuronal connections within distinct areas of the brain. This is called “local over-connectivity.”

Elevated fetal testosterone levels also change gene expression. That is, it has epigenetic effects. So this research may be relevant to helping us understand the complex gene-environment interactions that can produce autism. (Learn more about autism and epigenetics here.)

However, our Cambridge study involved just 600 typically developing children. This was too small a sample to test if autism involves elevated levels of fetal testosterone. To test this important question, we went on to collaborate with the Danish Biobank, which has more than 100,000 samples of stored amniotic fluid from pregnancies since the 1980s. From these, we identified 128 boys who had gone on to develop autism or Asperger syndrome. As we reported in 2014, we found that children on the autism spectrum indeed had higher levels of fetal testosterone.

Next, we want to test if the same elevation is found in girls who go on to develop autism. We already have a clue in that there are elevated rates of polycystic ovary syndrome in women with autism and their mothers. This syndrome is caused by elevated fetal testosterone.

Autism Speaks: So you’ve come a long way from your early psychological studies into theory of mind, to these new molecular studies into prenatal testosterone?

Dr. Baron-Cohen: It’s true. But to understand autism in its entirety, one needs to cross disciplines – to move from perception and cognition to neuroscience and from there into molecular biology, including hormones and genetics. To integrate findings in this way means working in multi-disciplinary teams, and I feel privileged to have such talented collaborators. I’ve also enjoyed complementing basic research with applied, clinical trials that can evaluate what actually helps.

I’m grateful to the Autism Speaks community for supporting autism research at all of these levels and across the world. But there’s no room for complacency. There’s a lot more work to do.