Two distinct groups, two different brain patterns
For generations, autism has been understood as a spectrum — vast, varied, and resistant to easy categorization. Now, researchers drawing on a large federal data archive have identified two distinct subtypes, grounded not in behavior alone but in the measurable terrain of language, intellect, adaptive functioning, and the brain's own connectivity patterns. The finding invites a deeper question: if the same diagnostic label can describe such different neurobiological realities, what does precision in care and understanding truly require of us?
- A long-standing tension in autism research — that a single diagnosis may obscure profoundly different conditions — has now found measurable, biological footing.
- Individuals with lower language and intellectual functioning show far greater divergence from neurotypical brain connectivity patterns than those with higher functioning, suggesting the gap between subtypes is not merely behavioral but neurological.
- The study sidesteps the slow machinery of new recruitment by mining an existing NIMH data archive, allowing patterns across many individuals to surface with unusual clarity.
- Clinicians and researchers now have a more precise framework for distinguishing subtypes — one that could reshape how interventions, educational strategies, and support services are matched to individuals.
- Open questions remain urgent: Are these subtypes stable across a lifetime? What genetic or developmental forces drive the connectivity differences? And will they hold when tested in new populations?
Researchers working with National Institute of Mental Health data have identified two distinct autism subtypes by examining language ability, intellectual functioning, and adaptive skills as they appear in late childhood and adulthood. The finding carries weight because it suggests that autism's internal divisions are not merely descriptive — they are written into the brain itself.
By drawing on an existing large-scale data archive, the team was able to map patterns across many individuals without new recruitment. The result was a clear division: those who struggle significantly with language, intellectual tasks, and adaptive functioning form one coherent group, while those with stronger abilities in these areas form another. When researchers examined functional MRI data along the brain's sensorimotor-association axis — the pathways connecting basic sensory and motor processing with higher-order thinking — the two groups diverged sharply. The lower-functioning subtype showed far greater differences from neurotypical connectivity patterns, suggesting distinct underlying neurobiology, not just different outcomes.
The work lands inside a broader push in autism research toward more granular, biologically grounded categorization. If subtypes can be reliably distinguished by cognitive and adaptive measures and validated against brain imaging, clinicians gain a more precise vocabulary — and potentially a more targeted toolkit for intervention and support.
Still, the study opens as many doors as it closes. What drives the connectivity differences? Do the subtypes remain stable across a lifetime? The research adds a meaningful data point to an emerging picture of autism as a condition of distinct biological realities — and nudges the field further from one-size-fits-all thinking toward something more honest about the variation it has always contained.
Researchers working with data from the National Institute of Mental Health have identified two distinct autism subtypes by examining how language ability, intellectual functioning, and adaptive skills show up in late childhood and adulthood. The finding matters because it suggests that autism is not a single condition but rather a spectrum with measurable, meaningful divisions—and those divisions show up not just in how people function day-to-day, but in the actual wiring of their brains.
The study drew from a large archive of existing data, allowing researchers to map patterns across many individuals without requiring new recruitment or testing. What they found was clean enough to be useful: people with autism who struggle significantly with language, intellectual tasks, and adaptive functioning form one coherent group, while those with stronger abilities in these areas form another. This distinction is not arbitrary. When the researchers looked at functional MRI scans and examined connectivity along the brain's sensorimotor-association axis—the neural pathways that link basic sensory and motor processing with higher-order thinking—they saw a clear difference between the two groups.
Specifically, people with autism who have lower language and intellectual functioning showed much greater differences in sensorimotor-association connectivity compared to neurotypical brains than did people with autism who have higher functioning in those same domains. In other words, the brain connectivity patterns diverge more sharply from typical development in the lower-functioning group. This suggests that the two subtypes may involve different underlying neurobiology, not just different outcomes.
The work builds on a conversation that has been gaining momentum in autism research: the recognition that the field needs better, more granular ways of categorizing autism beyond the current diagnostic framework. Researchers have been pushing for subtypes that capture real biological and functional differences, not just behavioral checklists. This study takes a step in that direction by grounding the subtypes in measurable cognitive and adaptive abilities, then validating them against brain imaging data.
The implications are practical. If autism subtypes can be reliably distinguished by language, intellectual, and adaptive functioning—and if those subtypes correlate with distinct brain connectivity patterns—then clinicians and researchers have a more precise language for talking about what they are seeing. Treatment approaches, educational strategies, and support services could potentially be tailored more effectively. A person identified as belonging to one subtype might benefit from different interventions than someone in the other group.
The study also opens questions. What drives the difference in sensorimotor-association connectivity between the two groups? Is it genetic, developmental, environmental, or some combination? Do the subtypes remain stable across the lifespan, or do people move between them? And perhaps most importantly: do these subtypes hold up when tested in new populations and with different measurement approaches?
For now, the research adds a data point to an emerging picture of autism as a condition with distinct biological subtypes. The work is part of a broader shift toward precision medicine in neurodevelopmental research—moving away from one-size-fits-all thinking and toward models that acknowledge real variation in how autism manifests in the brain and in lived experience.
The Hearth Conversation Another angle on the story
Why does it matter that researchers found two subtypes instead of just describing a spectrum from low to high functioning?
Because subtypes suggest different underlying biology. If two groups have different brain connectivity patterns, they might need different approaches to support or treatment. It's the difference between saying "people vary" and saying "there are actually distinct groups here."
But couldn't this just be the low end and high end of a single spectrum? Why call them subtypes?
That's the right question. The key is that the connectivity differences don't just scale smoothly. The low-functioning group shows a much sharper divergence from typical brain patterns than the high-functioning group does. That suggests something more categorical than just "more or less."
How did they avoid just rediscovering what everyone already knows—that some autistic people have more support needs than others?
They grounded it in brain imaging. Anyone can observe functioning differences. But when those differences correlate with specific patterns of neural connectivity, you have something more mechanistic. You're not just describing behavior; you're pointing to a biological basis.
What happens next? Does this change how autism is diagnosed or treated?
Not immediately. This is foundational work. But if these subtypes hold up in other datasets and other populations, then yes—clinicians could use language and adaptive functioning as markers to predict brain connectivity patterns, which could guide intervention choices.
Is there a risk this divides autistic people into categories that feel stigmatizing?
That's a real concern. The language of "low" and "high" functioning has always been fraught. But the researchers are trying to be precise about what they're measuring—not worth or potential, just specific cognitive and adaptive domains. The hope is that precision leads to better support, not judgment.