Autism: What's Sex Got to Do With It?

On the day after the hottest day in British history, in a small, stuffy office at the University of Cambridge, Simon Baron-Cohen's empathy is being put to the test. He has already sent his secretary to hunt down one of the few available fans. Crawling under his desk to plug it in, he has abandoned whatever donnish dignity he might have preferred to convey. He has shut the door, in spite of the heat, to keep hallway noise from troubling my tape recorder, and we have begun our interview—only now we are both hearing the relentless hum of the fan. Whuppa whuppa whuppa whuppa. "It's OK," I lie. Baron-Cohen says, "I'm not comfortable with this. You've come such a long way." He gets up and switches off the fan. An autistic person would not have done that. In spite of my sidelong glances at the tape recorder and my furtive efforts to nudge it closer to him, an autistic person would not have been able to see through my polite fib, put himself in my shoes and decipher my concerns. And, anyway, he wouldn't have shared my concerns. Autistic people, says Baron-Cohen, a psychologist who has studied and treated them for 20 years, lack empathy.

But they have a surfeit of something different—what Baron-Cohen calls "systemizing ability." They are lousy at understanding people but relatively good, he says, at making sense of the world. Some of them have a disablingly low IQ, and in such cases the systemizing may take the form of a seemingly purposeless obsession—they may stare for hours, say, at the veins of a leaf, or they may memorize train schedules or license plates. But in others, such as a mathematician Baron-Cohen knows at Cambridge who has been diagnosed with Asperger's syndrome—a disorder at the high-functioning end of the autism spectrum—that same systemizing ability can lead to work that is rewarded with fame. (Asperger's is a mild form of autism in which individuals are able to function normally, but have difficulty reading the emotions of others.)

Low-empathizing, high-systemizing: That, in a nutshell, is Baron-Cohen's theory of what characterizes autism. Those traits span the autism spectrum, from people who are mute and unable to function to people who find a niche in society. Moreover, Baron-Cohen's theory embeds this autism spectrum firmly in a much larger two-dimensional continuum—one that includes all of us. The essential difference between men and women, according to Baron-Cohen, is that women are better at empathizing and men at systemizing—on average, he stresses. There are plenty of male brains in female bodies, and vice versa. There are even female autistics, but there are many more male ones: In Baron-Cohen's theory, autism is a case of the "extreme male brain."

In the back of Baron-Cohen's book, The Essential Difference: The Truth About the Male and Female Brain, you can fill out questionnaires that allow you to determine your Empathy Quotient (EQ) and Systemizing Quotient (SQ). Baron-Cohen himself can't take the empathizing and systemizing tests, because he wrote them. But from all appearances he may be one of those fortunate individuals with a brain that is equally balanced between male and female. People who know him place him far up on the empathizing axis. "When you go into a meeting with him, you always feel good afterward," says one graduate student. Says another, "On the one hand, he'll coach us very closely, but on the other, he leaves us lots of space to do what we like." Yet Baron-Cohen is pushing a theory that attempts to capture the full diversity of human brain types in a single X-Y graph—and if that isn't male systemizing, what is? "We all have some autistic traits," he says. "It's just a matter of degree."

"I am interested in knowing the path a river takes from its source to the sea. Strongly agree? Slightly agree? Slightly disagree? Strongly disagree?"—from the Systemizing Quotient questionnaire

Baron-Cohen—born in 1959—grew up in Golder's Green, a middle-class and strongly orthodox Jewish neighborhood in North London. His father worked in the family menswear business; his mother taught dance. His first cousin, Sacha Baron-Cohen, is Ali G, the notorious assault comedian and on-air deflator of pompous windbags. Simon, in contrast, seems like he would be polite even to windbags. He is around six feet tall, with narrow, sloping shoulders and short, sandy hair that is beginning to show a male pattern; on the day we met he wore a blue short-sleeve shirt over khaki pants and sensible black shoes. The photo on his book jacket shows him without his wire-rim glasses, but he looks more natural with them on. His voice is mild and measured. Nothing in his bland and tidy little office—a Cezanne print, a few framed book covers—provides any obvious clues to where he is coming from.

Baron-Cohen himself offers one: He grew up with an older sister who is severely disabled, both mentally and physically. Today she lives in an institution, is confined to a wheelchair and has a very low IQ. "Yet despite that," says Baron-Cohen, "as soon as you walk into the room, she makes eye contact, her face lights up. Even though she has no language, you feel like you're connecting to another person."

In other words, she is the opposite of autistic. Autism is perfectly compatible with a high IQ—yet some degree of social disconnectedness, of extreme self-centeredness, has been a core feature of the disorder ever since it was first described in the 1940s and given a name derived from the Greek word for self. Baron-Cohen first encountered it when, fresh out of Oxford with an undergraduate degree in developmental psychology, he went to work teaching autistic children one-to-one at a small school in London. It was then he realized that autism is fascinating as well as sad. "I was struck by this dissociation between intelligence and social development," he says. "It became glaringly obvious that they are two different things."

Thanks in part to Baron-Cohen, that understanding of autism is now widely shared—which is one reason the number of children diagnosed as autistic has risen so dramatically in the past decade. Autism was once almost invariably associated with a below-normal IQ, and its prevalence was said to be around 4 in 10,000. Nowadays, it is ten times that. Many children are diagnosed with an autism-spectrum disorder, many of them at the high-functioning Asperger's end. With the explosion in diagnoses there has been an explosion in research. Geneticists are looking for genes linked to autism, which surely exist; the disease has been known to run in families. Neuroscientists are looking for the anatomical or physiological irregularities in the brain that must result from the anomalous genes.

Baron-Cohen is engaged in genetics and neurobiology, too, as codirector of the Autism Research Centre at Cambridge University. But his background is in cognitive psychology; he seeks to identify the basic mental processes that are common to all cases of autism and that link autistic behavior to its biological roots. In 1985, while still a graduate student at University College London, he made a breakthrough discovery of one such process. With his advisers Uta Frith and Alan Leslie, he presented autistic children with dolls named Sally and Anne, and the following story: Sally puts a marble in her basket and leaves the room. Anne takes the marble and hides it in her own box. Sally comes back and looks for her marble—where does she look?

A normal 4-year-old child says that Sally will look for the marble where she left it, in her basket. The child may even giggle at the joke on Sally. A kid with Down's syndrome will get it right too. But autistic children don't get it right. They say Sally will look in Anne's box—because after all, that's where the marble really is. They have no notion, Baron-Cohen discovered, of where Sally might think the marble is. They lack a "theory of mind"—abstract jargon for the simple realization, which the normal child comes to at around age 4, that other people have thoughts and intentions that may differ from his own. And that figuring those thoughts out helps him to understand what those people say and do.

Baron-Cohen later coined a term for this deficit: "mindblindness." In 1989, Uta Frith proposed that autistic people's inability to derive a theory of mind from their experience of the world was just one aspect of a broader deficit: the inability to draw together information so as to derive coherent and meaningful ideas. Frith's weak central coherence theory explained why people with autism remember strings of nonsense words almost as well as they do sentences, or why they do jigsaw puzzles without the picture: They just don't seek the pattern in a mass of details. "Their information-processing systems, like their very beings, are characterized by detachment," Frith wrote. A rival theory, which has proponents today, attributes the narrow interest in details, as well as other symptoms of autism, to executive dysfunction, a very broad inability to plan, to control impulses and to switch attention as needed to solve a problem.

Neither weak central coherence nor executive dysfunction, though, explain why some autistic people do so well. And in the 1990s, after Baron-Cohen had moved to Cambridge and begun seeing adult Asperger's patients, including many high achievers, at his own clinic, he became increasingly aware of that gap. Furthermore, he says, nobody seemed to be addressing another key fact: Autism affects far more boys than girls. At the Asperger's end of the spectrum, the ratio is about 10 to 1. The sex difference, says Baron-Cohen, is "one puzzle that has been completely ignored for over 50 years. I think it's a very big clue. It's got to be sex-linked."

"When I read the newspaper, I am drawn to tables of information, such as football scores or stock-market indices. Strongly agree? Slightly agree? Slightly disagree? Strongly disagree?"—from the Systemizing Quotient questionnaire

In The Essential Difference, before getting to his extreme-male-brain theory of autism, Baron-Cohen combs the psychological literature for evidence that normal sex differences in empathizing and systemizing are real and rooted in biology. He expected this claim to be controversial and was surprised and a little irritated when he read, "Didn't we always know this? Didn't our grandmothers tell us this?" Proving with scientific data that sex differences in behavior are innate is notoriously difficult. But Baron-Cohen, understandably enough, spares his popular audience the data. Indeed, the conclusions alone do have a familiar ring. Girls like dolls, boys like trucks. Girls like to gossip, boys like to roughhouse. Girls are more verbal, boys are more spatial, right through the SATs. Girls attack one another indirectly and verbally (which requires them to know how their victim feels). Boys are direct and physical, and when they reach manhood they are far more likely to commit murder—the ultimate in lack of empathy," as Baron-Cohen puts it.

On the other hand, men are also far more likely to be mathematicians, physicists or engineers, as well as to be better at throwing or catching balls. Those things are all examples of systemizing, according to Baron-Cohen, by which he means "the drive to understand a system and to build one." He defines a system as anything that takes an input and transforms it into an output according to some rule. For instance, a baseball's trajectory depends in a predictable way on where the pitcher places his fingers—so it's a system. Baron-Cohen's empathizing-systemizing dichotomy is far broader than the spatial-verbal one that has long been a feature of sex-difference research.

He has done as much as anyone lately to push the evidence for sex differences in behavior right back to the womb. In one study, for example, his graduate student Jennifer Connellan gave 1-day-old babies a chance to show a preference for looking at Connellan's face, at a distance of eight inches, or at a ball of the same size mounted on a stick. Connellan's face moved naturally, the ball more mechanically.

She found that 19 of 44 boys looked at least 10 seconds longer at the ball than at the face, while 11 preferred the face and 14 had no preference. In contrast, 21 of 58 girls preferred the face, while only 10 preferred the ball and 27—the largest group—had no preference. Earlier studies had suggested that women tend to make more eye contact and are better at decoding the language of the eyes. This study, Connellan and Baron-Cohen concluded rather daringly, demonstrates "beyond reasonable doubt" that "female superiority in social ability" is "in part, biological in origin."

What's more, says Baron-Cohen, that superiority may have something to do with how much testosterone a female fetus is exposed to—which is much less than a male fetus with functioning testes. At a Cambridge hospital, Baron-Cohen and his colleagues have access to a bank of frozen amniotic fluid samples, taken from women who, in 1996 and 1997, underwent amniocentesis before giving birth. When the children were 12 months old, Baron-Cohen and Svetlana Lutchmaya videotaped 70 of them (1 at a time) as they played on the floor of his tiny office and counted how many times in 20 minutes each child looked up at Mom. Later, when the kids reached 18 and 24 months of age, the researchers mailed questionnaires to their parents, asking them to evaluate their child's vocabulary. Meanwhile, the amniotic fluid revealed how much testosterone each child had been exposed to late in the first trimester, a critical time for brain development.

"When we got these results, I had one of those strange feelings, like a shiver down my spine," Baron-Cohen writes in his book. "A few drops more of this little chemical could affect your sociability or your language ability. I found it extraordinary." It would indeed be extraordinary if it were that simple. But to prove that more fetal testosterone (FT) is what makes boys less verbal and less interested in faces, you need to exclude the possibility that some other biological difference between the sexes is responsible. You need to show, for instance, not just that male fetuses have more testosterone than female fetuses and that boys turn out less verbal than girls but that the correlation holds within a single sex—that a boy with more testosterone will tend to have a less-evolved vocabulary than a boy with less. Baron-Cohen doesn't yet have that evidence. In their research papers, he and Lutchmaya state that they found a within-sex correlation with fetal testosterone only in one case: Boys with less FT—but not girls—were more apt to look up at Mom.

Mild-mannered and understated as he is in person, Baron-Cohen is willing in print to draw big conclusions from small studies—but he knows he needs larger studies to confirm the findings. One of his current graduate students, Rebecca Knickmeyer, is now laboriously tracking down 3,000 children who correspond to 3,000 amniotic fluid samples in that Cambridge hospital freezer. If she succeeds, she'll have a large enough group to say something firmer about fetal testosterone and social development—and in particular, about fetal testosterone and autism. A group that large should include around 15 children with autism. Baron-Cohen's working hypothesis is that they will have had the highest exposure to fetal testosterone of all.

But fetal testosterone is just one possible biological mechanism for generating an extreme male brain; Baron-Cohen's autism theory doesn't depend on it. That theory is a psychological one, and the evidence for it is psychological. On a wide variety of tests that distinguish normal females from normal males, he says—from eye contact to language development to understanding facial expressions to intuitive physics—autistics of both genders lie beyond normal males, on the other side of the spectrum from females.

That holds in particular for Baron-Cohen's EQ and SQ questionnaires. Though he sometimes presents them as measures of empathizing and systemizing ability, it is perhaps more accurate to see them as measures of interest. The SQ questionnaire, for example, doesn't determine whether you are actually good at math or even at keeping football statistics and stock quotes in your head—only whether you say you are interested in those things. But it and the EQ do separate the girls from the boys from the autistics on an X-Y graph—at least in the relatively small studies that Baron-Cohen has done so far.

"I tend to notice details that others do not. Strongly agree? Slightly agree? Slightly disagree? Strongly disagree?"—from the Autism Spectrum Quotient questionnaire

People have realized for decades that autism entails a deficit in empathy; that's not what's new about Baron-Cohen's theory. He and a young clinical psychologist named Ofer Golan have come up with a novel way of helping autistic people through a computer program. On a CD-ROM, trained actors demonstrate the facial expressions and vocal inflections that correspond to 412 distinct emotions or mental states, arranged under 24 headings, such as "sneaky" or "happy." The idea is that people with autism can bone up on their mind-reading skills without the stress of having to attend a group-therapy session. ("One of the things they report is that they feel flooded," says Golan.) When they do well on a quiz, the software rewards them with images of things they like—classifiable things, moving things, mechanical things or, ideally, things that are all of the above. "Stars, butterflies, things that move under a microscope," says Golan. "And trains."

What's novel about Baron-Cohen's theory of autism is how it portrays these characteristic obsessions. Autistic people and their families face enormous problems. Besides the universal social impairment, many of them suffer a devastating array of symptoms—mental, neurological, gastrointestinal—that may have nothing to do with their autism per se but nonetheless go along with it. What Baron-Cohen's theory says is that autistic people also have something positive: They're good at something. They're obsessed with systems, and they're good at systemizing, even when they don't happen to be mathematics professors or savants.

"You know," Baron-Cohen says, looking around his office for a ready example, "you and I just say, 'It's hot, we need a fan,' and turn it on. That isn't systemizing. A child with autism would look at the fan, and very likely would become fascinated by the rotation. What happens when light hits the blades, the kinds of reflections you get. So the child ends up staring at the fan for hours every day, because it is a form of mechanical motion that is systemizable—and that obsession gets described as purposeless. I actually think the child is doing something very intelligent."

Which is more or less the feeling I have about Baron-Cohen as I take leave of his office and his obsession: Right or wrong, his approach to understanding autistic people and how they fit in with the rest of us is intelligent—and empathetic.