Why Are Autism Spectrum Conditions More Prevalent in Males?

Four to one. That's the male to female ratio for autism diagnoses. If you're in a room of one hundred autistic people, roughly eighty of them are male. The question that ratio forces is deceptively simple: is that real, or is it an artifact of how we look at autism? And if it's real, what does it tell us about what autism actually is? Baron-Cohen and colleagues spent years working through that very question, and the answer reaches deep into fetal biology, cognitive science, and genetics. Let's start with the artifact problem, because it's significant. Baron-Cohen and colleagues lay out the ways clinical sampling inflates male to female ratios across neurodevelopmental conditions. Attention-deficit hyperactivity disorder, or ADHD, is the clearest example: clinic samples report ratios as high as ten to one, but community samples pull that down to between two and four to one, and adult studies often find no significant sex difference at all. Something similar happens with dyslexia and language impairment. Early reports suggested a strong male excess, but careful epidemiological work often finds near-equal numbers once you remove referral bias. For autism specifically, girls may present in ways that get interpreted as other conditions. The authors mention anorexia and borderline personality disorder as possible misclassifications. Girls also appear better at learning social conformity, imitating expected behavior well enough that clinicians miss the core features. Additionally, the standard diagnostic tools, the Autism Diagnostic Observation Schedule and the Autism Diagnostic Interview-Revised, were built largely on male samples. This means they may simply be less sensitive to how autism presents in females. So some of the four to one ratio is due to measurement. But Baron-Cohen and colleagues are clear that even after accounting for all of that, a genuine male skew remains. The artifact story doesn't close the case. To explain why that residual skew exists, Baron-Cohen built a theoretical framework called the Empathizing-Systemizing theory, or E-S theory. Empathizing is the drive to identify another person's thoughts and feelings and to respond with an appropriate emotion. Systemizing is the drive to analyze or construct rule-based systems—predicting how a mechanism works or extracting the logic from a pattern. Crucially, these are framed as drives, not just skills. They describe what you're motivated to seek out. The population-level data on these drives show consistent sex differences. On the Empathy Quotient, typical females score higher than typical males. On the Systemizing Quotient, typical males score higher than typical females. Those are averages across large groups, not descriptions of individuals, but the pattern is consistent across studies. The E-S framework maps these two dimensions onto different brain types, and the Extreme Male Brain hypothesis proposes that autism represents the far end of the male typical profile: very low empathizing combined with very high systemizing. That's not just a theoretical claim; the empirical data support it. On the Empathy Quotient, autistic individuals score lower than typical males, who score lower than typical females. On the Systemizing Quotient, autistic individuals score higher than typical males, who score higher than typical females. The autistic profile isn't a different kind of cognition; it's the male typical profile pushed to its extreme. The cognitive consequences are specific, resulting in superior performance on tasks that reward local, rule-based processing, like the Embedded Figures Task, along with genuine difficulty with social inference. The neural evidence aligns with these findings. The amygdala is, on average, larger in typical males than in females, and in autism, the amygdala is enlarged beyond typical male values. Brain size shows a parallel pattern: infant males tend to have larger brains than females, and children with autism show even larger brains early in life, around the time when diagnosis typically occurs. These aren't isolated findings. They are consistent with the view that the sex typical differences in brain structure are exaggerated in autism, rather than inverted or randomized. So, the pattern is clear. What causes it? The leading candidate is fetal testosterone. Males experience a surge in testosterone between approximately weeks eight and twenty-four of gestation—levels that, in the paper's phrasing, reach almost pubertal values. This surge is thought to play an organizing role in brain development, shaping the structure and connectivity of regions involved in social cognition, empathy, and systemizing. Because you can't manipulate fetal hormones experimentally in humans, researchers have turned to amniocentesis—routine clinical sampling of amniotic fluid during midpregnancy—as a way to measure what the fetus was actually exposed to. The Cambridge Fetal Testosterone Project enrolled around six hundred thirty-five children and followed them from infancy into middle childhood. The findings are striking in their consistency. Higher fetal testosterone was associated with less frequent eye contact at twelve months, smaller vocabulary at eighteen and twenty-four months, poorer quality social relationships at forty-eight months, and more restricted interests at forty-eight months. By ninety-six months, higher fetal testosterone predicted lower empathy scores and higher systemizing scores. Three studies confirmed a direct positive correlation between amniotic testosterone levels and the number of autistic traits a child shows in toddlerhood and later childhood. That's a longitudinal chain running from a measurement taken in the womb to cognitive and social outcomes a decade later. It's rare to have that kind of temporal depth in developmental research, and it's one of the reasons the fetal testosterone theory has accumulated the most converging evidence of any biological account. Genetic data add another layer. Baron-Cohen and colleagues summarize associations between autistic traits and ten genes involved in sex steroid synthesis, transport, or metabolism. Among them are the CYP17A1 gene, the CYP19A1 gene, the ESR1 gene, and the androgen receptor. Post-mortem brain tissue from autistic individuals shows decreased expression of aromatase and RORA, both enzymes involved in converting androgens to estrogens. There's also the digit ratio proxy: the ratio of the second to fourth finger length reflects prenatal androgen exposure, and people with autism tend to show lower ratios, consistent with higher fetal testosterone. Additionally, females with congenital adrenal hyperplasia—a condition of elevated prenatal androgens—score higher on Autism-Spectrum Quotient measures than controls, representing a natural clinical test of the androgen hypothesis. The fetal testosterone account is strong, but Baron-Cohen and colleagues are careful to present the competing theories fairly. The X-chromosome theory points out that the X chromosome contains more brain-expressed genes than any other chromosome, and that X-linked causes explain some known forms of autism. For instance, Fragile X produces autism in forty-six percent of males who carry the full mutation. Sex chromosome aneuploidies like Turner syndrome and XYY also show elevated autistic traits. However, broad linkage and association studies haven't found consistent X-linked loci for autism, and copy-number variation scans show X mutations in only a small minority of cases. The Y-chromosome theory is thinner. The SRY gene, which initiates male sex determination, is expressed in the hypothalamus and in frontal and temporal brain regions, and in vitro data suggest it can influence the transcription of tyrosine hydroxylase, a step in dopamine synthesis. But the human genetic evidence is sparse, and the data are too limited to draw firm conclusions. The reduced penetrance model takes a different angle: the same de novo genetic mutations that cause autism in males may cause it at lower rates in females. That framing has some statistical support from family data, but most studies find a roughly equal sex ratio among autistic individuals who carry de novo copy-number variants, which would be strange if the female protective effect were strong. What's important here is that these theories are not actually in competition. X and Y chromosome genes could be regulated by fetal testosterone, or could affect how sensitive the developing brain is to it. Multiple mechanisms could be operating simultaneously, each contributing a portion of the male excess. Autism is multi-factorial; that's not a hedge—it's the most accurate description of a condition with diverse presentations, heterogeneous genetics, and a complex developmental timeline. What comes next, according to Baron-Cohen and colleagues, is scale. The Cambridge cohort of six hundred thirty-five children is too small to test whether fetal testosterone is specifically elevated in those who go on to receive a clinical diagnosis. Biobanks like the Danish Biobank, which holds tens of thousands of amniotic samples, could power that test. Large sex-stratified genetic studies would help disentangle chromosome and hormone effects. Improved understanding of the female protective effect—whatever it turns out to be—could directly enhance how clinicians identify autistic females who are currently being missed because the tools weren't built with them in mind. The four to one ratio isn't just a statistic about prevalence; it's a research target, and there's real ground to gain. This lecture was created by ennepō. Go to https://ennepo.ai to Discover, Create and Follow the latest research in your field. Read when you can. Listen when you want to.