Subthreshold Autism and the Broader Autism Phenotype

Subthreshold autism refers to a pattern of autistic traits that are clearly present and measurable but fall short of meeting the full diagnostic criteria for autism spectrum disorder. The traits themselves are real and can affect daily life, but they may be too few in number, too mild in severity, or not consistently visible across all settings to warrant a clinical diagnosis.¹ The concept has gained traction in research over the past two decades as evidence mounts that the boundary between “autistic” and “not autistic” is far blurrier than a yes-or-no diagnosis suggests, and that people living below that diagnostic line can still face genuine difficulties.

What “Subthreshold” Actually Means

A person might show clear signs of social-communication difficulty and some rigid or repetitive behaviors, but not enough of each to check every required box in the diagnostic manual. Or they might have the right number of symptoms at home but appear to function well at school or work, creating a mismatch between settings that technically disqualifies them. Sometimes the issue is that different observers disagree: a parent sees clear difficulties, but a teacher does not, and the diagnostic criteria expect consistency across sources.² Any of these gaps can push someone below the threshold while leaving them with challenges that feel very real in everyday life.

The diagnostic threshold in modern psychiatry is designed to emphasize specificity over sensitivity. That means the system is built to avoid false positives rather than to catch everyone who might benefit from support. Individual autistic-like symptoms, such as difficulty reading social cues or a strong need for routine, can show up in many different mental health conditions and even in people with no diagnosis at all.³ This is part of why the line between “has autism” and “has some autistic traits” is genuinely hard to draw, and why the concept of subthreshold autism keeps coming up in both clinical practice and research.

The Broader Autism Phenotype in Families

One of the strongest pieces of evidence that autistic traits exist on a continuum comes from family studies. Parents and siblings of people diagnosed with autism are more likely to show mild versions of the same social, communication, and behavioral features, even when they have no diagnosis themselves. Researchers call this the broader autism phenotype, or BAP.⁴ The most commonly reported pattern is subtle difficulty with social interaction and communication that mirrors what you see in autism, just turned down in intensity.

Siblings of autistic children have been studied extensively for these traits. Reviews spanning over two decades of research consistently find that these siblings are more likely to show differences in social functioning, language use, and cognitive style compared to siblings of children without autism.⁵ These are not children with undiagnosed autism; they are children who function within the typical range but sit higher on the trait distribution.

Genetic research has reinforced this picture. The same common genetic variants that increase the statistical risk of an autism diagnosis also appear to influence everyday autistic-like traits in the general population. A study comparing genetic risk scores found significant overlap between the variants linked to diagnosed autism and those associated with traits like rigidity, attention to detail, and certain childhood behavioral patterns.⁶ In mothers of autistic individuals, higher genetic loading for autism-related variants was associated with more pronounced broader autism phenotype traits, suggesting a shared genetic liability that runs through families.⁷

How Subthreshold Traits Affect Thinking and Social Cognition

People with elevated autistic traits who do not have a diagnosis still show measurable differences in several cognitive areas. The pattern is not random; it tends to track the same profile seen in diagnosed autism, just to a lesser degree. This is one of the reasons researchers take subthreshold traits seriously rather than treating them as noise.

Theory of mind, the ability to infer what someone else is thinking or feeling, is one of the clearest areas of difference. Siblings of autistic children show reduced theory of mind abilities regardless of whether they meet the criteria for any clinical label themselves.⁸ In adults from the general population, high levels of autistic traits are specifically linked to difficulty with cognitive theory of mind, the ability to figure out what someone else believes or intends, while affective empathy, feeling what someone else feels, tends to remain intact.⁹ That distinction matters: a person with subthreshold traits may genuinely care about others’ emotions but struggle to read the social cues that indicate what those emotions are.

Executive function shows a similar pattern. People with higher autistic traits tend to have more difficulty with planning, mental flexibility, and managing competing demands, even after accounting for any co-occurring attention deficit symptoms.¹⁰ The areas that seem to be spared, like basic inhibitory control and keeping physical materials organized, are the same areas that tend to be relatively preserved in diagnosed autism. This mirroring is part of what supports the idea that subthreshold traits are a lighter version of the same underlying profile rather than something fundamentally different.

Cognitive flexibility, the ability to switch between different mental tasks or adjust when the rules change, is another area where people with high autistic traits perform notably worse. This holds on both standard neuropsychological tests and on tasks that require switching between different emotional categories.¹¹ In practical terms, this can show up as difficulty adapting to unexpected changes in plans, trouble shifting conversational topics gracefully, or a strong preference for routine that causes real distress when routines break down.

Mental Health Below the Diagnostic Line

One of the most clinically important findings about subthreshold autism is that it carries real mental health risk. People with elevated autistic traits who do not meet the full criteria for autism still report significantly more mood problems, anxiety, and depressive symptoms than those with low autistic traits.¹² This puts them in a kind of clinical gap: they have enough trait burden to face real psychological consequences, but not enough to qualify for the diagnosis that might connect them to specialized support.

The pattern shows up early in life. In a community-based study of Japanese preschoolers, children with mild to moderate autistic traits were significantly more likely to score above clinical thresholds for emotional and behavioral problems compared to their peers with fewer traits.¹³ These were not children on the verge of an autism diagnosis; they were children with subclinical trait levels who nonetheless had a harder time emotionally and behaviorally. Subthreshold autistic traits have also been reported in roughly one in ten children and adolescents with obsessive-compulsive disorder, and that overlap was not explained by OCD severity.¹⁴

The Eating Disorder Connection

A growing body of research has documented elevated autistic traits in people with eating disorders, particularly anorexia nervosa. The overlap is striking enough that some researchers have proposed anorexia may represent a female-presenting phenotype of the autism spectrum.¹⁵ A systematic review of the published literature found an overrepresentation of autistic symptoms in people with anorexia nervosa.¹⁶

When researchers assess people with various eating disorders using tools designed to capture the full range of autistic traits (including subthreshold features), those with eating disorders consistently score higher than healthy controls across most domains.¹⁷ Sensory sensitivity appears to be one mechanism connecting the two conditions. In a study of adult women with eating disorders, the link between sensory sensitivity and disordered eating behavior was significantly mediated by the presence of autistic traits, meaning that higher sensory sensitivity was associated with more autistic features, which were in turn associated with more eating-related difficulties.¹⁸

This matters practically because subthreshold autistic traits in a person with an eating disorder can complicate treatment. Rigid thinking patterns, sensory aversions to food textures, and difficulty with the kind of flexible goal-shifting that cognitive behavioral therapy relies on may all be intensified when a patient has a significant autistic trait burden, even one that falls short of a full autism diagnosis.

Loneliness and Social Life

Perhaps the most consistent finding in this area is that subthreshold autistic traits are linked to loneliness and social isolation, even in people who would never be flagged for a clinical assessment. In college students, a stronger broader autism phenotype, characterized by rigidity, preference for sameness, and high attention to detail, was associated with significantly more loneliness as well as fewer and shorter friendships.¹⁹

The mechanism appears to run through social connectedness. Young adults with higher autistic trait scores report feeling less socially connected, which relates to more loneliness, which in turn predicts more anxiety and depression.²⁰ More recent work has extended this, finding that people with higher autistic traits not only experience more loneliness but also report greater susceptibility to social pain from rejection, which was in turn associated with more unusual perceptual experiences.²¹

In a study of performing arts professionals, those with elevated subthreshold autistic traits reported lower self-efficacy, lower quality of life across all domains measured, and higher levels of anxiety and depression.²² This is a population that is generally social and expressive by occupational demand, which underscores that the effects of subthreshold traits are not limited to people who already lead isolated lives.

Camouflaging and Its Costs

Many people with subthreshold autistic traits learn to mask or camouflage their difficulties in social settings. They consciously mimic social behaviors, rehearse conversational scripts, or force eye contact in ways that look effortless from the outside but are exhausting to maintain. Research on medical students found that social camouflaging behaviors were linked to burnout, which in turn was associated with worse mental well-being. The chain ran from autistic traits to camouflaging to burnout to reduced well-being, with anxiety and emotional reactivity acting as additional pathways in the same model.²³

Camouflaging helps explain why some people with subthreshold traits can appear to function well in structured settings like work or school while still struggling intensely in private. The effort required to “pass” as neurotypical drains cognitive and emotional resources, leaving less for everything else. Clinicians who only see the polished performance in a structured assessment may never detect the traits at all, which is part of why subthreshold presentations slip through diagnostic cracks so easily.

How Traits Shift Across the Lifespan

Autistic traits are not static. In a large cross-sectional study of nearly 3,000 adults aged 18 to 85, total broader autism phenotype scores were higher in younger adults and lower in older ones. The age-related decline was most dramatic for pragmatic language difficulties, the kind of social communication challenges like taking turns in conversation, sticking to a topic, or adjusting language to the audience. Aloofness showed a smaller decline with age, while rigidity was not significantly associated with age at all.²⁴ The pattern is consistent with the idea that some social communication skills improve with life experience, while preference for routine and structure tends to persist.

That said, what older adults with high autistic traits gain in social polish they may lose in cognitive performance. In a study of over 22,000 adults aged 50 to 80, those with high autistic traits showed more difficulty on tasks measuring memory, processing speed, and executive function compared to those with low traits.²⁵ This suggests that while the social dimension of the broader autism phenotype may soften with age, the cognitive dimension may become more prominent, or at least more detectable, as people get older and the normal age-related decline in these abilities leaves less margin for compensation.

Substance Use Patterns

The relationship between subthreshold autistic traits and substance use is not straightforward, and it defies the stereotype that autistic features are always protective. In a large twin study of Australian adults, higher autistic trait scores were associated with elevated rates of regular smoking, cannabis use, and nicotine and cannabis use disorders. The picture for alcohol was more nuanced: people with high autistic traits were actually less likely to drink to the point of intoxication. But among those who did initiate regular drinking, they were at elevated risk of developing alcohol dependence.²⁶ That pattern, lower casual use but higher risk of problematic use once started, is an unusual combination that clinicians might not anticipate.

Case report literature has also raised the possibility that unrecognized subthreshold autistic features in adults with other psychiatric conditions may play a role in complex addiction profiles. One detailed report described a patient with bipolar disorder and multiple substance and behavioral addictions whose clinical picture became far more coherent once subthreshold autism spectrum features were recognized and factored into the treatment plan.²⁷ The argument is not that autistic traits cause addiction, but that missing them in a clinical picture can lead to treatment approaches that do not address the full set of vulnerabilities.

Measuring Subthreshold Traits

Several questionnaires have been developed specifically to capture autistic features below the diagnostic threshold, rather than just detecting who does and does not meet criteria for autism. The Subthreshold Autism Trait Questionnaire (SATQ) is a 24-item self-report tool that was designed to measure these traits in the general population and has shown good ability to distinguish between people with diagnosed autism and typical university students.²⁸

The Adult Autism Subthreshold Spectrum (AdAS Spectrum) scale takes a broader approach, assessing the full range of autistic-like features across seven domains. It was originally validated in clinical groups but has since been tested in non-clinical samples, where it showed strong reliability and correlated in the expected direction with other established measures of autistic traits and empathy.²⁹ The tool has been used to identify optimal score thresholds, with one study suggesting a cutoff that distinguishes people with significant autistic traits from those without.³⁰

These tools are useful for research and screening, but they are not diagnostic instruments. A high score suggests someone has a lot of autistic-like traits; it does not by itself confirm or rule out a clinical diagnosis. Still, they have been valuable in demonstrating that autistic traits are distributed continuously in the population rather than clustering neatly into “has autism” and “does not have autism” categories.

Cross-Cultural Differences in How Traits Are Recognized

One complicating factor is that the way autistic traits manifest and are recognized varies across cultures. Research using the Autism-Spectrum Quotient in Italian and British samples has found that the broader autism phenotype can be reliably measured cross-culturally: parents of autistic children in Sicily scored higher than control parents on the same subscales as parents in English-speaking samples, and the tool identified a similar gradient of trait severity across families.³¹

But the picture is not perfectly consistent everywhere. When researchers tested whether the structure of autism symptoms as defined by DSM-5 held up across Finnish and British samples, the model fit well in the UK for people with broader autism phenotype traits, but poorly in Finland.³² The researchers noted that cross-cultural variability appeared to be greatest for milder autistic characteristics, precisely the range where subthreshold presentations live. This suggests that the threshold between “autistic enough for diagnosis” and “subthreshold” is partly a cultural construction, influenced by local norms for social behavior, communication style, and what counts as unusual.

The Evolutionary Question

If the genes associated with autism are common in the population and autism itself carries reproductive costs, why have those genes persisted? One theoretical proposal is that the same genetic variants that produce autism at one extreme of the distribution provide social and cognitive advantages to people in the middle and lower portions of the trait range. Under this model, the genes do not exist “despite” autism but “because” the traits they produce benefit most of the people who carry them. The postulated mechanism involves epistasis, where the effect of one gene depends on the context of other genes, creating an oscillation in outcomes across the trait distribution.³³ This remains a hypothesis rather than established fact, but it offers an intriguing framework for why autistic traits are so widely distributed in the general population rather than being confined to the small percentage of people who meet diagnostic criteria.

The practical implication of this perspective is that subthreshold autistic traits are not inherently problems to be solved. Attention to detail, deep focus on specific interests, preference for structure, and a drive for logical consistency can be assets in many environments. The difficulty arises when those traits interact with social demands, sensory environments, or mental health vulnerabilities in ways that cause suffering, and when the person experiencing that suffering does not have a diagnostic label that would connect them to help.