Autism in Women: Why It Goes Undiagnosed and What to Do About It

The widely cited autism prevalence ratio is approximately 4:1 male to female in diagnosed populations. But when researchers examine prevalence estimates using broader assessment criteria, community samples, and tools specifically designed to detect autism in women, the actual ratio converges on something closer to 2:1 or even 1:1 in some studies.

The gap between the diagnosed ratio and the estimated actual ratio represents a substantial number of autistic women who are either undiagnosed or misdiagnosed. This is not a small discrepancy — it may represent millions of women in the U.S. alone who are carrying an autism diagnosis labeled as something else, or no diagnosis at all.

Why does this happen? The diagnostic tools, the diagnostic criteria, and the clinical prototype for autism were all built on observations of autistic males — predominantly white autistic boys. The DSM-5 criteria reflect that prototype. When evaluators apply male-coded criteria to female patients, they systematically underidentify autism in women and girls.

Camouflaging — also called masking — is the active suppression and modification of autistic traits in order to appear neurotypical. It is more prevalent, more intensive, and more costly in autistic women than in autistic men, and it is the primary reason why female autism goes undetected.

Camouflaging involves:

• Studying social scripts — carefully observing neurotypical peers and memorizing the rules of social interaction as a cognitive exercise rather than experiencing them intuitively. Many autistic women describe learning to make eye contact by memorizing the “look away every few seconds” rule, or studying how friends gesture and mirror those patterns.
• Suppressing stimming in public — autistic people stim (self-stimulatory behavior: rocking, hand-flapping, repetitive movement or sound) as a sensory regulation mechanism. Women learn early that stimming provokes social attention, and internalize the suppression of stimming as a social survival skill. This suppression is effortful and metabolically costly.
• Performing social engagement — asking scripted follow-up questions, matching the emotional register of the room, maintaining performed interest in conversations that are cognitively and sensory-overloading. For autistic women who are highly socially motivated, this performance can be convincing enough that no one — including therapists — suspects autism.

The cumulative cost of masking is severe. Research using the CAT-Q (Camouflaging Autistic Traits Questionnaire) has found that higher camouflaging scores are associated with significantly worse mental health outcomes — depression, anxiety, burnout, and suicidal ideation. The performance of neurotypicality is physiologically exhausting, and it delays diagnosis by an average of 10–20 years in women.

The DSM-5 criteria for autism spectrum disorder were developed primarily from observations of autistic males. Several features of how autism presents in women systematically escape those criteria:

Intense Interests That Look Socially Acceptable

The DSM specifies “highly restricted, fixated interests that are abnormal in intensity or focus.” In autistic boys, these interests have historically presented in ways that seem obviously unusual to evaluators: encyclopedic knowledge of train schedules, specific numbers, obscure computer systems. In autistic women, the same intensity of interest applies to socially approved topics — animals, psychology, literature, specific musical artists, particular historical periods. These interests may be socially indistinguishable from neurotypical passion, so evaluators don't flag them.

Social Motivation vs. Social Ease

A common misconception is that autistic people don't want social connection. Autistic women, in particular, often very much want connection — and they invest enormous cognitive effort in pursuing it. What they experience is not a lack of social motivation but social exhaustion, confusion about unspoken rules, and the sense that they are perpetually slightly out of step with the social environment in ways they can't fully identify. Providers who see this social motivation assume the patient can't be autistic.

Internalized vs. Externalized Behavioral Differences

Autistic boys often present with more externalized behavioral differences that are visible in school settings and prompt referrals: disruptive behavior, obvious rule-resistance, visible distress. Autistic girls more often internalize — their distress appears as anxiety, shutdowns, and compliance at school followed by meltdowns at home where the mask comes off. School-based screening catches the externalized presentation; the internalized presentation goes home with a note that says “such a sweet, quiet girl.”

Before a late autism diagnosis, the typical female patient has accumulated a diagnostic history that looks like this: anxiety disorder in adolescence, depression in young adulthood, possibly a BPD label at some point (emotional meltdowns misread as emotional dysregulation rather than sensory overwhelm), possibly an eating disorder diagnosis (rigid food rules and sensory food aversions misread as anorexia). Each diagnosis addresses one part of the presentation while missing the organizing framework.

The individual diagnoses are not always wrong — autistic women do have higher rates of anxiety, depression, and eating disorders. But treating the downstream symptoms without identifying the upstream neurodevelopmental architecture is like treating the manifestations of a structural foundation problem one room at a time. The anxiety that comes from chronic sensory overload and masking will not fully respond to SSRI therapy and CBT the way anxiety arising from a different source would.

The co-occurrence of autism and ADHD — sometimes called AuDHD — is estimated at 30–80%, with the wide range reflecting differences in study methodology and diagnostic thresholds. Both conditions involve executive function differences; both involve masking in social contexts; both are significantly underdiagnosed in women.

AuDHD is a clinically distinct presentation with important treatment implications. Autistic individuals have heightened sensory sensitivity and may be more vulnerable to the sensory side effects of stimulant medication — increased heart rate, heightened sensory sensitivity, anxiety amplification. The executive function deficits in AuDHD are partially overlapping and partially distinct from those in ADHD alone, which means stimulants improve some aspects (attention, impulse control) while leaving other aspects (cognitive rigidity, transition difficulty, processing overwhelm) unaddressed.

ADHD also compounds the masking load. Autistic ADHD women are managing the social performance demands of autism masking while simultaneously managing the executive function deficits of ADHD — and using the executive system that is already compromised to do the masking. This is the structural setup for autistic ADHD burnout, which can look nearly indistinguishable from treatment-resistant depression.

Sensory Processing in Autistic Women

Sensory processing differences are a core feature of autism that are frequently reframed as anxiety or OCD in women. Sensory avoidance — strong reactions to clothing textures, food textures, sounds, lights — may be diagnosed as OCD or health anxiety when the evaluator doesn't recognize it as sensory defensiveness. Meltdowns triggered by sensory overwhelm — brief, intense periods of emotional and behavioral dysregulation — are often misread as BPD-associated emotional dysregulation.

The autistic sensory experience is not anxiety about sensations — it is a fundamentally different sensory processing architecture in which the nervous system treats ordinary stimuli as genuinely overwhelming. This distinction matters for treatment: sensory accommodations (quieter environments, sensory tools, predictable sensory schedules) address the cause; SSRI therapy may help with co-occurring anxiety but does not change the underlying sensory processing.

The Hormonal Intersection: Estrogen and Masking Capacity

Estrogen has direct effects on social cognition. It upregulates oxytocin receptor expression, facilitates social reading and mirror neuron activity, and provides what might be called neurological scaffolding for social competence. For autistic women, estrogen — particularly the relatively stable estrogen of reproductive years — partially compensates for autistic social processing differences and supports masking capacity.

This means that autistic women often find masking harder during the premenstrual phase, when estrogen drops sharply. Premenstrual unmasking — appearing more openly autistic, having more difficulty with social performance, sensory hypersensitivity amplified — is frequently misdiagnosed as PMDD.

It also explains why perimenopause is a common late-diagnosis trigger. As estrogen declines in perimenopause (average onset around age 47), the neurological scaffolding supporting masking progressively weakens. Women who have successfully masked their autism for decades find the performance increasingly difficult and begin experiencing the social and sensory world more like an unmasked autistic person. This is when many women receive their first autism evaluation — and, sometimes, their first autism diagnosis — in their late 40s or 50s.

A late autism diagnosis almost always produces a grief cycle, and this is normal and appropriate. The typical sequence:

Relief — finally, a framework that explains the entire pattern. The years of not fitting in, the exhaustion no one else seemed to share, the diagnoses that were partially right but never quite correct. Many women describe this as one of the most profound relief experiences of their lives.

Grief — and then, grief. Grief for the unlived life: the career paths abandoned because the sensory or social environment was unmanageable, the relationships that failed in ways that a diagnosis might have helped navigate differently, the years spent blaming themselves for difficulties that were neurological. This grief is real and deserves space — it is not a sign that the diagnosis was unwelcome.

Post-diagnosis, the clinical work shifts: identity consolidation, learning what accommodations actually help, understanding which previous diagnoses are independent and which are downstream of autism, and — critically — beginning the process of unmasking, which means learning to exist more authentically without the performance overhead that masking requires.

Standard autism assessment tools — the ADOS-2 (Autism Diagnostic Observation Schedule) and the ADI-R (Autism Diagnostic Interview, Revised) — were designed primarily with male presentations in mind and have known sensitivity limitations for autistic women. The CAT-Q (Camouflaging Autistic Traits Questionnaire) was developed specifically to measure camouflaging in adults and is more sensitive to the female presentation.

When seeking an evaluation, here are productive questions to ask the evaluator:

• “Are you familiar with the female autism phenotype and the camouflaging literature? What tools do you use that are sensitive to the female presentation?”
• “I may be camouflaging during our session. How do you account for the performance quality of the evaluation in your assessment?”
• “I have a diagnosis of [anxiety / depression / PMDD]. I'm wondering whether some of those presentations might be downstream of an unrecognized autism diagnosis. Are you able to hold that differential?”
• “My intense interests are in [animals / literature / psychology]. How do you assess restricted interests in adults whose interests are socially conventional?”