Three autism subtypes based on single-subject gray matter network revealed by semi-supervised machine learning.
Machine-learning on single brain scans splits autism into three biologically different sub-types linked to IQ and ADOS scores.
01Research in Context
What this study did
Xu et al. (2024) fed each child’s gray-matter brain scan into a machine-learning tool. The program looked at how one brain region talks to another without naming the child.
It then let the data sort itself into groups. The goal was to see if autism looks the same in every brain or if clear sub-types hide inside the diagnosis.
What they found
The computer pulled out three separate autism blocks. Each block had its own wiring pattern and differed on IQ and ADOS scores.
In plain words: one label, three brain stories.
How this fits with other research
Guo et al. (2023) also found autism sub-types, but they used salience-network scans and saw only two groups. The numbers differ, yet both studies prove the same point: autism is not one-size-fits-all.
Lancioni et al. (2011) showed three sensory-based clusters in kids. Guomei et al. now match that three-way split using brain networks instead of parent reports. The match hints the same biology drives both findings.
HByiers et al. (2025) add age to the picture. They show brain micro-structure clusters shift as kids grow. Guomei et al. did not track age, so future work can layer the age factor onto the three gray-matter types.
Why it matters
You can’t scan every client, but you can borrow the idea. Treat the autism label as a starting point, not a finish line. Watch IQ, ADOS, and sensory profiles; they quietly point to which brain subtype a child may belong to. Planning different social, play, or self-help goals for each cluster could speed progress until brain-based screening lands in clinics.
Want CEUs on This Topic?
The ABA Clubhouse has 60+ free CEUs — live every Wednesday. Ethics, supervision & clinical topics.
Join Free →Pull IQ, ADOS, and sensory scores for each learner and sort your caseload into rough high, mid, and low support clusters; trial different reinforcement densities across clusters and track which set responds faster.
02At a glance
03Original abstract
Autism spectrum disorder (ASD) is a heterogeneous, early-onset neurodevelopmental condition characterized by persistent impairments in social interaction and communication. This study aims to delineate ASD subtypes based on individual gray matter brain networks and provide new insights from a graph theory perspective. In this study, we extracted and normalized single-subject gray matter networks and calculated each network's topological properties. The heterogeneity through discriminative analysis (HYDRA) method was utilized to subtype all patients based on network properties. Next, we explored the differences among ASD subtypes in terms of network properties and clinical measures. Our investigation identified three distinct ASD subtypes. In the case-control study, these subtypes exhibited significant differences, particularly in the precentral gyrus, lingual gyrus, and middle frontal gyrus. In the case analysis, significant differences in global and nodal properties were observed between any two subtypes. Clinically, subtype 1 showed lower VIQ and PIQ compared to subtype 3, but exhibited higher scores in ADOS-Communication and ADOS-Total compared to subtype 2. The results highlight the distinct brain network properties and behaviors among different subtypes of male patients with ASD, providing valuable insights into the neural mechanisms underlying ASD heterogeneity.
Autism research : official journal of the International Society for Autism Research, 2024 · doi:10.1002/aur.3183