Graph-level contrastive learning with self-aware and cross-sample topology augmentation for brain disorder diagnosis using rs-fMRI.

Researchers developed a new computer method to help diagnose brain conditions like autism and depression using brain scans. The method works better than current approaches and can identify unusual brain connection patterns. This could help doctors make more accurate diagnoses using brain imaging, though the research is still in early stages and needs more testing.

This study presents GCSC-TA, a novel machine learning approach for diagnosing brain disorders using resting-state functional MRI data. The method addresses the challenge of limited labeled training data in clinical settings by using contrastive learning with brain network augmentation strategies. The approach generates complementary views of brain networks for each subject and uses a specialized loss function to preserve brain topology structure. Testing on datasets for Major Depressive Disorder and Autism Spectrum Disorder demonstrated superior classification performance compared to existing methods.

The approach also identified abnormal brain connectivity patterns, potentially advancing clinical interpretability of fMRI for diagnosis.

This computational approach shows promise for improving autism diagnosis using brain imaging, but requires extensive clinical validation before implementation. The ability to work with limited labeled data could make advanced neuroimaging more accessible in clinical practice, though safety and reliability standards must be established.

Sample sizes not reported. Study appears to be primarily computational/methodological without detailed clinical validation. Limited information about study populations or clinical characteristics. Unclear generalizability to real-world clinical settings.

Resting-state functional MRI (rs-fMRI) is widely used for diagnosing and analyzing brain disorders. However, existing fMRI studies have shown that learning-based approaches depend heavily on labeled training data, which is difficult to obtain due to the substantial time and effort required for annotation in clinical settings. To address these challenges, we propose GCSC-TA (Graph-level Contrastive Learning with Self-aware and Cross-sample Topology Augmentation) for brain disorder diagnosis and analysis using rs-fMRI. The proposed GCSC-TA generates two complementary augmented brain networks for each subject by introducing self-aware and cross-sample topology augmentations.

This dual-view strategy enhances the identification of individual-specific features and also amplifies inter-subject functional heterogeneity. Moreover, we designed a min-max contrastive loss function to accommodate augmented brain networks, overcoming the limitations of traditional projection-based methods while performing graph-level contrastive learning on the original integrity of the brain topology structure. Extensive experiments on a private Major Depressive Disorder (MDD) dataset and the publicly available Autism Spectrum Disorder (ABIDE) dataset demonstrate the superior classification performance of GCSC-TA over several state-of-the-arts. Furthermore, GCSC-TA also identifies abnormal brain connectivity patterns associated with MDD and ASD, thereby advancing the interpretability and clinical utility of rs-fMRI for clinical diagnosis.