Summary: A new international study resolves long-standing clinical uncertainty by demonstrating that autism can be divided into at least two biologically distinct subtypes based on patterns of brain connectivity measured with functional magnetic resonance imaging (fMRI). This work represents the first systematic cross-species effort to translate fMRI connectivity signatures into specific molecular mechanisms using genetically characterized mouse models.
By analyzing more than 1,900 human brain scans together with data from 20 genetically defined mouse models, the researchers identified a reproducible “hypoconnectivity” subtype linked to synaptic dysfunction and a “hyperconnectivity” subtype associated with immune-related molecular systems. These distinct connectivity profiles provide measurable, brain-based biomarkers that can guide more precise, biology-informed approaches to diagnosing and treating autism spectrum disorder (ASD).
Key Facts
- Parsing autism variability: Autism has long been described as highly heterogeneous in behavior and clinical presentation. This study provides direct biological evidence that some of that variability maps onto fundamentally different neural and molecular mechanisms, supporting the move from one-size-fits-all care toward precision medicine for ASD.
- Cross-species fMRI mapping: Led by Dr. Alessandro Gozzi (IIT) and Dr. Adriana Di Martino (Child Mind Institute), the team used genetic and biochemical characterizations of mouse models as a biological reference to interpret human fMRI connectivity patterns, effectively creating a cross-species “Rosetta Stone” for linking imaging signatures to cellular pathways.
- Hypoconnectivity subtype — synaptic origins: One reproducible subtype shows widespread reductions in functional connectivity between brain regions. Gene expression analyses indicate that the affected areas are enriched for genes involved in synaptic structure and function, implicating synaptic pathways as a driving mechanism.
- Hyperconnectivity subtype — immune associations: The other reproducible subtype displays increased connectivity, with brain regions showing elevated communication. This profile corresponds to transcriptional and immune-related pathways and is associated with moderately higher scores on standardized measures of autism severity.
- Extensive validation across datasets: The study combined fMRI from 20 mouse models with scans from 940 individuals with idiopathic autism and 1,036 neurotypical controls drawn from the Autism Brain Imaging Data Exchange (ABIDE) and other sources. The two dominant subtypes together explained roughly 25% of the individuals studied and were reproducible across many independent research sites.
- Moving beyond behavior alone: Standard behavioral assessments do not capture underlying neural circuitry or the molecular causes of ASD. These brain-based subtypes point toward targeted diagnostic markers and therapeutic strategies that address specific synaptic or immune environments rather than only outward symptoms.
Source: Child Mind Institute
An international team coordinated by the Istituto Italiano di Tecnologia (IIT) in Rovereto, Italy, and the Child Mind Institute in New York, with collaborators from the University of Trento and other institutions, has demonstrated that at least two autism subtypes can be reliably identified by their fMRI connectivity signatures.
In the identified subtypes, brain networks either communicate more than typical (hyperconnectivity) or less than typical (hypoconnectivity). By anchoring human imaging patterns to molecular and cellular findings from mouse models, the study builds tools intended to support personalized diagnosis and care for individuals with autism.
The research article was published in the journal Nature Neuroscience. The study was coordinated by Alessandro Gozzi, PhD, director of the Center for Neuroscience and Cognitive Systems (CNCS) at the IIT, and Adriana Di Martino, MD, founding director of the Autism Center at the Child Mind Institute. This represents the first coordinated attempt to decode human fMRI connectivity by tracing signatures back to defined biological alterations in animal models.
Using functional connectivity maps from 20 mouse models alongside multicenter human fMRI datasets, the researchers established two robust subtypes: hypoconnectivity tied to synaptic pathways and hyperconnectivity linked to immune- and transcriptional-related systems. Collectively, these subtypes captured approximately one quarter of the autistic individuals in the combined human datasets.
“For decades, we’ve observed tremendous variability in how autism manifests, but we lacked direct evidence that these differences reflected distinct underlying biology,” said Dr. Alessandro Gozzi. “Our approach allowed us to isolate specific genetic and immune factors and translate those signatures to human brain scans, showing that different connectivity patterns encode different mechanistic pathways underlying autism.”
By combining imaging with genetic and biochemical analyses in mouse models, the team linked connectivity signatures to cellular dysfunctions. This mapping showed how specific molecular pathways produce distinct connectivity patterns detectable with fMRI and established biological reference patterns that guided subtype detection in humans.
“The mouse models gave us a biological ‘Rosetta Stone,’” said Dr. Adriana Di Martino. “We could see which pathways drive which connectivity signatures and then search for those same patterns in human data.”
Human neuroimaging data were drawn from multicenter resources including the Autism Brain Imaging Data Exchange (ABIDE) and institutional contributions coordinated by the Child Mind Institute. Analyses reproduced hypo- and hyperconnectivity subtypes in human cohorts and confirmed, through gene expression analysis, that hypoconnected brain regions are enriched for synaptic genes while hyperconnected regions are enriched for immune-related genes. These findings were robust across independent datasets, supporting biological consistency.
The two subtypes differ not only in connectivity architecture but also show modest differences on standardized autism assessments, with the hyperconnectivity group tending to score somewhat higher on severity measures. The researchers note that additional subtypes likely exist and that larger datasets and refined analytic tools will be needed to map the full diversity of ASD.
Funding: This international collaboration was coordinated by the Italian Institute of Technology and the Child Mind Institute, with support from the Simons Foundation Autism Research Initiative, the European Research Council (projects #DISCONN and #BRAINAMICS), the Brain and Behavior Foundation, Fondazione Telethon, and the U.S. National Institute of Mental Health.
Key Questions Answered:
A: Autism is not a single condition but a highly heterogeneous set of neurobiological profiles. For decades, diagnosis and treatment relied mainly on observable behavior, which masks distinct underlying brain circuitry and molecular mechanisms. This study shows that different connectivity patterns reflect different biological systems, underscoring the need for personalized, mechanism-based treatments.
A: Mouse models with defined genetic or molecular alterations serve as a controlled biological reference. While fMRI reveals network-level communication in human brains, it cannot directly identify the microscopic molecular causes. The mouse models allowed researchers to link specific genetic and immune changes with connectivity signatures, creating a cross-species map that can be applied to human fMRI data.
A: They differ fundamentally in how brain regions communicate. Hypoconnectivity shows reduced inter-regional communication and is associated with synaptic gene alterations. Hyperconnectivity shows excessive inter-regional communication, aligns with immune-related transcriptional changes, and tends to correlate with somewhat higher autism severity scores.
Editorial Notes:
- This article was edited by a Neuroscience News editor.
- The journal paper was reviewed in full by editorial staff.
- Additional context and synthesis were added by the reporting team.
About this autism research news
Author: Media Office, Child Mind Institute
Source: Child Mind Institute
Contact: Media Office – Child Mind Institute
Image: Image credited to Neuroscience News
Original Research: Closed access. “Autism subtypes identified using cross-species functional connectivity analyses” by Marco Pagani, Valerio Zerbi, Silvia Gini, Filomena Grazia Alvino, Abhishek Banerjee, Andrea Barberis, M. Albert Basson, Yuri Bozzi, Alberto Galbusera, Jacob Ellegood, Michela Fagiolini, Jason P. Lerch, Michela Matteoli, Caterina Montani, Davide Pozzi, Giovanni Provenzano, Maria Luisa Scattoni, Nicole Wenderoth, Ting Xu, Michael V. Lombardo, Michael P. Milham, Adriana Di Martino & Alessandro Gozzi. Nature Neuroscience. DOI: 10.1038/s41593-026-02287-z
Abstract
Autism subtypes identified using cross-species functional connectivity analyses
Phenotypic heterogeneity in autism has long been thought to reflect underlying biological diversity, but direct cross-scale evidence has been limited. Using cross-species functional neuroimaging, the authors show that fMRI connectivity alterations observed across 20 genetic mouse models cluster into hypoconnectivity-dominant and hyperconnectivity-dominant subtypes. These subtypes correspond to distinct biological pathways: hypoconnectivity with synaptic dysfunction and hyperconnectivity with transcriptional and immune-related alterations. Analogous subtypes were identified in a multicenter human fMRI dataset (n = 940 individuals with idiopathic autism; n = 1,036 neurotypical controls). The human subtypes are replicable, show distinct network architectures and behavioral profiles, and recapitulate the synaptic and immune-related pathways found in rodents. This work establishes an empirical framework for targeted, biology-informed subtyping across the autism spectrum.