Saved in:
Bibliographic Details
Main Authors: Li, Xiaolong, Guo, Guiliang, Wen, Guangqi, Cao, Peng, Yang, Jinzhu, Wu, Honglin, Liu, Xiaoli, Wang, Fei, Zaiane, Osmar R.
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.19295
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866911530821353472
author Li, Xiaolong
Guo, Guiliang
Wen, Guangqi
Cao, Peng
Yang, Jinzhu
Wu, Honglin
Liu, Xiaoli
Wang, Fei
Zaiane, Osmar R.
author_facet Li, Xiaolong
Guo, Guiliang
Wen, Guangqi
Cao, Peng
Yang, Jinzhu
Wu, Honglin
Liu, Xiaoli
Wang, Fei
Zaiane, Osmar R.
contents Mental disorder populations exhibit pronounced heterogeneity -- that is, the significant differences between samples -- poses a significant challenge to the definition of positive pairs in contrastive learning. To address this, we propose a subtype-guided contrastive learning framework that models patient heterogeneity as latent subtypes and incorporates them as structural priors to guide discriminative representation learning. Specifically, we construct multi-view representations by combining patients' clinical text with graph structure adaptively learned from BOLD signals, to uncover latent subtypes via unsupervised spectral clustering. A dual-level attention mechanism is proposed to construct prototypes for capturing stable subtype-specific connectivity patterns. We further propose a subtype-guided contrastive learning strategy that pulls samples toward their subtype prototype graph, reinforcing intra-subtype consistency for providing effective supervisory signals to improve model performance. We evaluate our method on Major Depressive Disorder (MDD), Bipolar Disorder (BD), and Autism Spectrum Disorders (ASD). Experimental results confirm the effectiveness of subtype prototype graphs in guiding contrastive learning and demonstrate that the proposed approach outperforms state-of-the-art approaches. Our code is available at https://anonymous.4open.science/r/BrainSCL-06D7.
format Preprint
id arxiv_https___arxiv_org_abs_2603_19295
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle BrainSCL: Subtype-Guided Contrastive Learning for Brain Disorder Diagnosis
Li, Xiaolong
Guo, Guiliang
Wen, Guangqi
Cao, Peng
Yang, Jinzhu
Wu, Honglin
Liu, Xiaoli
Wang, Fei
Zaiane, Osmar R.
Machine Learning
Mental disorder populations exhibit pronounced heterogeneity -- that is, the significant differences between samples -- poses a significant challenge to the definition of positive pairs in contrastive learning. To address this, we propose a subtype-guided contrastive learning framework that models patient heterogeneity as latent subtypes and incorporates them as structural priors to guide discriminative representation learning. Specifically, we construct multi-view representations by combining patients' clinical text with graph structure adaptively learned from BOLD signals, to uncover latent subtypes via unsupervised spectral clustering. A dual-level attention mechanism is proposed to construct prototypes for capturing stable subtype-specific connectivity patterns. We further propose a subtype-guided contrastive learning strategy that pulls samples toward their subtype prototype graph, reinforcing intra-subtype consistency for providing effective supervisory signals to improve model performance. We evaluate our method on Major Depressive Disorder (MDD), Bipolar Disorder (BD), and Autism Spectrum Disorders (ASD). Experimental results confirm the effectiveness of subtype prototype graphs in guiding contrastive learning and demonstrate that the proposed approach outperforms state-of-the-art approaches. Our code is available at https://anonymous.4open.science/r/BrainSCL-06D7.
title BrainSCL: Subtype-Guided Contrastive Learning for Brain Disorder Diagnosis
topic Machine Learning
url https://arxiv.org/abs/2603.19295