Unsupervised data-driven stratification of mentalizing heterogeneity in autism

Area of Science:

• Neuroscience
• Genetics
• Psychology

Background:

• Autism spectrum conditions (ASC) exhibit significant heterogeneity, complicating clinical and translational research.
• Novel methods are required to dissect this heterogeneity for improved precision.

Purpose of the Study:

• To apply a systems biology clustering approach to identify discrete subgroups within the adult ASC population.
• To investigate the replicability of these subgroups across independent datasets.

Main Methods:

• Utilized a clustering approach from genomics and systems biology on two large cognitive datasets (n=694, n=249).
• Assessed performance on the Reading the Mind in the Eyes Test (RMET) to evaluate mentalizing abilities in identified subgroups.

Main Results:

• Identified 5 replicable ASC subgroups with distinct patterns of performance on the RMET.
• Three subgroups (45-62% of ASC adults) demonstrated significant impairments in mentalizing abilities (Cohen's d = -1.03 to -11.21).
• Other identified subgroups showed no significant impairments.

Conclusions:

• These findings reveal robust natural subdivisions within the ASC population.
• Subgroup identification can facilitate more individualized research inferences.
• This approach may accelerate progress towards precision medicine in autism research.