Researchers identify and characterize two distinct autism spectrum disorder subtypes

Autism spectrum disorder may not be a single, uniform condition after all. A new study published in Translational Psychiatry reports the identification of two distinct autism spectrum disorder subtypes, offering fresh evidence that the diagnosis may encompass biologically and clinically different patterns rather than one continuous profile. The findings could intensify a major shift in autism research: moving away from treating autism as a single category and toward understanding it as a collection of related but distinguishable developmental pathways.

The study, led by H. Zhang, T. Qin, Y. Zhai and colleagues, is described as a multidimensional characterization of autism spectrum disorder. That term is important because autism is defined through a broad combination of traits, including differences in social communication and restricted or repetitive behaviors. However, people who meet the same diagnostic criteria can differ substantially in language development, sensory processing, intellectual abilities, daily functioning, behavioral patterns and co-occurring conditions. A multidimensional approach attempts to examine these features together instead of reducing autism to one overall score.

The researchers’ identification of two subtypes suggests that the apparent diversity within autism may follow an underlying structure. Rather than viewing every individual as occupying a random position on a single spectrum, scientists are increasingly testing whether groups of people share recognizable constellations of traits. Such groupings may reveal that similar diagnostic outcomes can arise through different combinations of developmental characteristics, cognitive profiles or biological mechanisms.

In technical terms, subtype discovery generally involves analyzing multiple variables simultaneously and searching for recurring patterns within complex data. Statistical methods such as clustering, latent class analysis or other unsupervised approaches can separate participants according to similarities across several dimensions. The resulting groups are then evaluated to determine whether they are stable, clinically meaningful and distinguishable from one another. Although the citation alone does not provide the study’s full methodology or the precise features defining the two groups, its central contribution is the reported recognition and multidimensional description of two autism-related profiles.

This kind of classification could eventually influence how clinicians assess autistic people. Current diagnosis relies primarily on observed behavior and developmental history, which are essential but may not fully capture the biological diversity behind those behaviors. If distinct subtypes can be replicated in larger and more diverse populations, assessment could become more precise, helping clinicians anticipate different support needs. One subtype might be associated with greater challenges in one developmental domain, while another could show a different balance of communication, behavior, cognition or adaptive functioning.

The potential impact extends beyond diagnosis. Autism interventions are not one-size-fits-all, and response to therapy can vary widely. A reliable subtype framework could help researchers design more targeted clinical trials by determining whether a treatment benefits a particular profile rather than producing modest average effects across a highly mixed group. This approach, known as stratified or precision medicine, seeks to match interventions to the characteristics most relevant to an individual’s needs. In autism research, it could improve the detection of treatment effects that are currently obscured by biological and behavioral variation.

The findings may also help explain why autism research has sometimes produced inconsistent results. Studies that combine participants with very different developmental profiles may dilute signals in brain imaging, genetics, physiology or behavior. If two groups have different underlying mechanisms, averaging them together can make both appear less distinct. Separating subtypes could allow scientists to investigate whether specific neural circuits, genetic pathways or developmental processes are more strongly associated with one profile than another.

At the same time, the discovery of subtypes should not be interpreted as creating rigid boxes or ranking autistic people by severity or value. Autism is a lifelong neurodevelopmental condition with considerable individual variation, and people may not fit neatly into any classification system. Subtypes are scientific models designed to organize complexity, not definitions of identity. Their usefulness will depend on whether they remain respectful of lived experience, improve support and remain flexible as new evidence emerges.

The next test will be replication. A meaningful subtype system must hold up across different ages, cultures, countries, recruitment strategies and levels of support need. Researchers will also need to determine whether the two profiles remain stable over development or whether people can move between them as language, cognition and adaptive skills change. Future studies may combine clinical measurements with genetics, brain imaging, physiological data and longitudinal follow-up to discover whether the reported groups represent distinct biological pathways or primarily different patterns of behavior.

The study’s message is therefore both provocative and practical: autism may be one diagnosis composed of multiple underlying realities. By examining the condition across several dimensions, Zhang and colleagues have added momentum to efforts to make autism research more precise without losing sight of its human diversity. If the two-subtype model survives further testing, it could become a foundation for more individualized assessment, better-targeted research and support strategies designed around the actual needs of autistic people rather than the averages of a broad diagnostic category.