Researchers led by Italy's Istituto Italiano di Tecnologia and New York's Child Mind Institute have identified two fundamentally distinct autism subtypes—one defined by hyperconnectivity, the other by hypoconnectivity—each rooted in different genetic and immune biology. The discovery, published in Nature Neuroscience, marks the first systematic effort to trace human brain imaging patterns back to their molecular causes, offering a foundation for precision medicine approaches that could reshape how autism is understood and treated.
For decades, clinicians and researchers have observed remarkable variability in how autism presents itself across individuals, yet they lacked direct biological evidence for whether these differences reflected distinct underlying mechanisms or simply surface variation of a single condition. That gap is what Dr. Alessandro Gozzi, director of the Center for Neuroscience and Cognitive Systems at the Italian Institute of Technology, and Dr. Adriana Di Martino, founding director of the Autism Center at the Child Mind Institute, set out to close. Their team used an innovative approach: they analyzed functional brain connectivity patterns in twenty different mouse models alongside brain scans from 940 children and young adults with autism and more than 1,000 neurotypical control subjects.
The pattern that emerged was striking. In the hypoconnectivity subtype, brain areas communicate less than they do in typically developing brains, and genetic analyses revealed these regions were enriched for synaptic genes—those governing how neurons connect and communicate. In the hyperconnectivity subtype, brain areas show excessive communication, and the enriched genes related to immune function, pointing toward inflammatory mechanisms. Together, these two subtypes accounted for approximately 25 percent of the individuals with autism the team examined.
What made this discovery possible was an elegant experimental strategy. The researchers used mouse models as a biological "Rosetta Stone," as Dr. Di Martino described it. By linking specific genetic and biochemical alterations in mice to observable patterns of brain connectivity, they could identify the exact molecular pathways driving each connectivity signature. They then searched for those same signatures in human brain scans drawn from the Autism Brain Imaging Data Exchange (ABIDE)—a pioneering neuroimaging initiative that aggregates datasets from research laboratories worldwide—and the Child Mind Institute's own data.
The findings held up under scrutiny. The hypo- and hyperconnectivity subtypes proved reproducible across dozens of independent research datasets, a critical validation that confirmed their biological consistency rather than artifacts of any single lab. The two subtypes also showed modest but meaningful differences on standardized autism severity assessments, with the hyperconnectivity group scoring moderately higher. Yet the most significant discovery was that brain-based biological markers revealed distinctions current behavioral tests do not fully capture.
The implications extend far beyond academic understanding. By identifying biological subtypes with different mechanistic underpinnings, researchers now have a clearer path toward targeted interventions. A treatment effective for immune-driven hyperconnectivity may have little value for synaptic-driven hypoconnectivity, and vice versa. This precision medicine approach could transform autism care from a one-size-fits-all model into personalized strategies matched to each individual's underlying biology. The work invites the field toward a future where autism diagnosis becomes not just a behavioral assessment, but a biologically informed classification that opens doors to more effective, individualized support.