The clinical problem the study addresses is both common and consequential. Psychiatric conditions such as anxiety, depression, attention-deficit/hyperactivity disorder, and post-traumatic stress disorder occur at elevated rates in autism spectrum disorder, yet they are chronically underdiagnosed in this population. Standard diagnostic interviews depend on the patient’s ability to articulate internal states, which is precisely what many autistic individuals, particularly those who are minimally verbal, find difficult or impossible. Clinicians therefore fall back on inference: they watch for aberrant behaviors and guess at the underlying distress. The trouble is that the same behavior can carry different meanings in different people. Aggression in a typically developing child might signal oppositional defiance, while the identical behavior in a minimally verbal autistic teenager might reflect undiagnosed obsessive-compulsive symptoms or unexpressed physical discomfort. Without a systematic way to link behaviors to symptom domains, inference becomes guesswork, and guesswork leads to missed diagnoses, inappropriate treatments, and prolonged suffering.

To build that systematic link, Strobl turned to a large pooled dataset drawn from three studies archived in the NIMH Data Archive. The combined sample included assessments on two complementary instruments. The first was the Child and Adolescent Symptom Inventory-5, which quantifies severity across twelve psychiatric symptom domains, including ADHD-hyperactive symptoms, oppositional defiant disorder, generalized anxiety, social anxiety, specific phobia, separation anxiety, obsessions, somatization, depression, post-traumatic stress, and schizophrenia-related severity. The second was the Aberrant Behavior Checklist, a 58-item rating scale originally developed to measure treatment effects in developmental disabilities, which captures behaviors such as irritability, lethargy, stereotypies, hyperactivity, and inappropriate speech. Caregiver ratings on both instruments were available for the same individuals, allowing the analysis to ask, item by item, which behaviors travel together with which symptoms.

The methodological core of the paper is a technique called Differentially Supervised Varimax, an extension of classical factor analytic rotation. Traditional varimax rotation, introduced by Kaiser in 1958, seeks uncorrelated components that each explain a maximal share of variance in a set of observed variables. The new approach modifies this objective: instead of maximizing variance explained in the behavior items alone, it searches for behavior profiles—weighted combinations of the 58 checklist items—that maximally differentiate symptom severity across groups. In other words, the algorithm is supervised by the diagnostic information: it learns which patterns of aberrant behavior track symptom dimensions most sharply within the autistic group, within the typically developing comparison group, and within subgroups of autistic individuals split by verbal ability into minimally verbal and fluently verbal categories. After adjusting for age and sex, group differences in the strength of these behavior-symptom associations were tested with permutation testing, a resampling procedure that builds an empirical null distribution, and controlled for false discoveries across the many simultaneous comparisons.

The results reveal a set of behavior profiles with strikingly specific psychiatric correlates in autism. A disruptive and hyperactive profile—combining irritability, agitation, and excessive motor activity—related more strongly in autistic individuals than in typically developing individuals to ADHD-hyperactive severity and oppositional defiant severity, but also to generalized anxiety, social anxiety, and specific phobia. This finding suggests that what looks like pure externalizing behavior in an autistic child may frequently be the behavioral surface of internalizing conditions, a possibility that standard diagnostic heuristics often overlook. A second profile, characterized by emotional reactivity—intense, rapidly shifting emotional responses to minor triggers—aligned with obsessive symptoms, somatization, and depressive severity. A third, marked by social withdrawal and self-injurious behavior, tracked social anxiety and separation anxiety specifically. Preoccupation, a fourth profile involving intense, circumscribed fixations, corresponded to schizophrenia-related severity, while a hypoactive and depressed-mood profile aligned with post-traumatic stress and specific phobia.

Perhaps the most clinically provocative result concerns minimally verbal autistic individuals. In this subgroup, a profile of passive withdrawal and hypoactivity—quiet disengagement, low energy, reduced initiation—related more strongly to obsessions, somatization, and depressive severity than it did in fluently verbal autistic individuals. The implication is sobering: the quiet, shut-down autistic patient who barely speaks may be carrying a heavy burden of obsessive thoughts, physical distress, and depression that is entirely invisible to a clinician expecting depression to look like tearfulness or verbalized hopelessness. Because minimally verbal individuals represent what researchers have called the neglected end of the spectrum, this finding provides an evidence-based starting point for detecting their suffering through behavior alone.

The study’s authors emphasize that the value of these profiles lies in their practical use as diagnostic triage tools. The proposed workflow is straightforward: a clinician observes a patient’s predominant aberrant behavior pattern, matches it to one of the identified profiles, and then prioritizes targeted history-taking, structured caregiver interviewing, and symptom measures in the linked domains. Prominent withdrawal accompanied by self-injury should trigger focused screening for social and separation anxiety. Marked emotional reactivity should prompt assessment for obsessive symptoms, somatic complaints, and depression. A disruptive, hyperactive presentation in an autistic patient should raise suspicion not only of ADHD and oppositional behavior but also of anxiety disorders. This does not replace clinical judgment; it channels it, converting a broad and unmanageable differential diagnosis into a focused set of hypotheses that can be confirmed or refuted with appropriate instruments and caregiver reports.

The findings also carry theoretical weight for the science of psychopathology in autism. A long-standing debate concerns whether standard psychiatric constructs—depression, anxiety, PTSD—mean the same thing in autistic and non-autistic people, or whether measurement invariance breaks down across the spectrum. Prior work has shown mixed results, with some studies finding equivalent factor structures for anxiety and depression measures and others finding distortions. The present results suggest a middle path: the symptom constructs themselves may be broadly applicable, but the behavioral expressions that signal them differ systematically between autistic and typically developing individuals, and again between minimally verbal and fluently verbal autistic subgroups. This heterogeneity in expression, rather than in underlying experience, may explain why autistic people are so often misdiagnosed or undiagnosed, and why caregiver report remains an indispensable but imperfect window into internal states.

The technical approach also contributes to a broader movement in precision psychiatry, in which machine learning and multivariate statistics are used to find objective markers that stratify patients and guide treatment. Where biomarker efforts have pursued neural signatures and genetic risk scores, this study demonstrates that low-cost, widely collected behavioral ratings can be reanalyzed to yield clinically actionable structure. The differentially supervised rotation method is general: the same framework could, in principle, be applied to other populations in which self-report is unreliable, including individuals with intellectual disability, dementia, or severe psychiatric illness. The pooled NIMH Data Archive design, combining multiple studies with harmonized instruments, strengthens the generalizability of the profiles, though the authors note that the associations are cross-sectional and do not establish that behaviors cause symptoms or vice versa.

For families, educators, and clinicians, the message is that behavior is communication, and that the code can be learned. A child who withdraws and hurts herself is not simply being defiant or self-stimulating; she may be signaling social and separation anxiety that treatment could relieve. A child whose emotions explode at small provocations may be struggling with obsessions, bodily distress, or depression. A quiet, passive child on the spectrum may be the one in the most psychological pain. By matching observed behavior patterns to empirically derived profiles and then probing the linked symptom domains, caregivers and clinicians can move from reactive crisis management to proactive detection of mental distress in people who cannot describe it in words—a shift that could meaningfully change outcomes across the autism spectrum.

Subject of Research: Behavior profiles that differentially indicate psychiatric symptom dimensions in autistic versus typically developing individuals and across verbal ability subgroups

Article Title: Unique Behavior Profiles That Specify Mental Distress in Autism

Article References: Strobl, E. V. (2026). Unique Behavior Profiles That Specify Mental Distress in Autism. Journal of Autism and Developmental Disorders. https://doi.org/10.1007/s10803-026-07518-7

Image Credits: AI Generated

DOI: 10.1007/s10803-026-07518-7

Keywords: autism spectrum disorder, aberrant behavior checklist, mental health, anxiety, depression, minimally verbal, self-injury, psychiatric comorbidity, factor analysis, precision psychiatry, caregiver report, NIMH Data Archive