NormMoX improves autism diagnosis by explaining brain differences
NormMoX: A Normative Mixture of Experts for Explainable Autism Classification from Structural and Functional MRI
Machine Learning
Summary
Diagnosing autism with brain scans is hard because existing methods struggle to explain why they make their decisions. The authors created NormMoX, a system that compares each person's brain to a typical brain based on age and sex, making it easier to see where and how a brain is unusual. Their method works with both structural and functional brain scans, handles missing data, and outperforms previous approaches in tests. NormMoX helps not only predict autism but also understand the specific brain differences involved.
What this means in practice
- •For clinical imaging teams: Use a normative brain-based system to detect and explain autism-related brain abnormalities from MRI scans in patients.
- •For medical device developers: Incorporate a reliable fusion model that integrates structural and functional MRI to improve autism diagnostic tools with explainability features.$Commercial implications: Enables development of advanced diagnostic software that provides detailed brain-based insights, appealing to hospitals and clinics.
Authors
Muhammad Hasan Masrur, Fani Deligianni
Abstract
While structural and functional magnetic resonance imaging (sMRI, fMRI) offer profound potential for researching autism spectrum disorder (ASD) and understanding individual neurodivergence, current classifiers lack the explanatory power needed for scientific study, clinical use and personal-level insights. Most models learn from imaging features that also vary with age, sex, scanner site and head motion. They fuse the two modalities without a common reference and explain decisions against arbitrary reference inputs rather than a typical brain. We present NormMoX, a normative mixture of experts that makes the typical brain the shared reference for prediction, fusion and explanation. Normative models trained on large healthy cohorts give the value each brain measure is expected to take for a person's age and sex. NormMoX learns from each person's deviations from these expected values. For each modality, a neural network expert turns the deviations into a score of evidence for ASD. A fusion head, MoX, compares each score with those of typical controls before combining them. It therefore needs no imputation when a modality is missing and shows in which modality each person is atypical. On 1,035 participants of ABIDE I, NormMoX reached an AUROC of $0.728 \pm 0.004$ in repeated cross-validation, above linear, single-modality and deep fusion baselines. NormMoX thus shows not only whether a person is likely to have ASD, but where, and in which modality, their brain departs from typical. The implementation is available at https://github.com/Novelex/NormMox.