ASD Classification via Multi-View Renormalized Graph Convolutional Networks.

Saved in:

Bibliographic Details
Title:	ASD Classification via Multi-View Renormalized Graph Convolutional Networks.
Authors:	Wang, Chuang¹ wangchuang4078@126.com, Sun, Sixiang¹ ssx1026@126.com, Liu, Junli¹ liujunli2023@126.com, Li, Yuanyuan¹ forkp@djtu.edu.cn, Zhou, Wenjie² tiankong305@163.com, Li, Dongyan¹ lidy@djtu.edu.cn
Source:	IAENG International Journal of Computer Science. May2026, Vol. 53 Issue 5, p1750-1758. 9p.
Subjects:	Autism spectrum disorders, Graph neural networks, Computer-assisted image analysis (Medicine), Machine learning, Large-scale brain networks, Biomarkers, Functional magnetic resonance imaging
Abstract:	The integration of Graph Convolution Networks (GCNs) with functional networks constructed from resting-state functional Magnetic Resonance Imaging (rs-fMRI) demonstrates promising potential for early diagnosis of Autism Spectrum Disorder (ASD). Current methodologies typically reduce dimensionality through community detection algorithms prior to graph convolution operations. However, these approaches fail to capture the hierarchical structural patterns inherent in local connectivity networks. Therefore, this paper proposes a Multi-view Renormalization Graph Convolution Network (MVR-GCN) framework that integrates multi-view renormalization to delineate structural information within brain networks. The Box-based Graph Convolution (BoxGraphConv) module employs a hierarchical graph convolutional architecture coupled with a multi-view feature learning strategy, enabling effective deep feature extraction. This framework significantly enhances the model's capability to interpret complex network topologies and improves predictive performance. Experimental results demonstrate that MVR-GCN outperforms existing methods on the Autism Brain Imaging Data Exchange (ABIDE) dataset. Specifically, it achieves notable improvements in classification accuracy and Area Under the Curve (AUC), with increases of approximately 2.47% and 1.81%, respectively. Moreover, the biomarkers identified by MVR-GCN align closely with established medical knowledge, offering new insights for the clinical diagnosis of ASD. [ABSTRACT FROM AUTHOR]
	Copyright of IAENG International Journal of Computer Science is the property of International Association of Engineers (IAENG) and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database:	Engineering Source

Description
Abstract:	The integration of Graph Convolution Networks (GCNs) with functional networks constructed from resting-state functional Magnetic Resonance Imaging (rs-fMRI) demonstrates promising potential for early diagnosis of Autism Spectrum Disorder (ASD). Current methodologies typically reduce dimensionality through community detection algorithms prior to graph convolution operations. However, these approaches fail to capture the hierarchical structural patterns inherent in local connectivity networks. Therefore, this paper proposes a Multi-view Renormalization Graph Convolution Network (MVR-GCN) framework that integrates multi-view renormalization to delineate structural information within brain networks. The Box-based Graph Convolution (BoxGraphConv) module employs a hierarchical graph convolutional architecture coupled with a multi-view feature learning strategy, enabling effective deep feature extraction. This framework significantly enhances the model's capability to interpret complex network topologies and improves predictive performance. Experimental results demonstrate that MVR-GCN outperforms existing methods on the Autism Brain Imaging Data Exchange (ABIDE) dataset. Specifically, it achieves notable improvements in classification accuracy and Area Under the Curve (AUC), with increases of approximately 2.47% and 1.81%, respectively. Moreover, the biomarkers identified by MVR-GCN align closely with established medical knowledge, offering new insights for the clinical diagnosis of ASD. [ABSTRACT FROM AUTHOR]
ISSN:	1819656X