MHNet: Multi-view High-order Network for Diagnosing Neurodevelopmental Disorders Using Resting-state fMRI

TL;DR

MHNet addresses the need for more informative representations in NDD diagnosis from rs-fMRI by integrating multi-view high-order features learned in both Euclidean and non-Euclidean spaces. The model combines ESFE and Non-ESFE branches, with FFC fusion to produce robust subject-level predictions, and it leverages a novel G-IBHN-G hierarchical brain network along with HGNN and HCNN modules. Across ABIDE-I, ABIDE-II, and ADHD-200 datasets, MHNet using Brainnetome atlas outperforms state-of-the-art methods, with ablations confirming the value of multi-view and high-order representations. The work highlights interpretability via brain region importance, offers atlas-related insights, and suggests pathways to incorporate non-imaging data in future brain disorder diagnosis frameworks.

Abstract

Background: Deep learning models have shown promise in diagnosing neurodevelopmental disorders (NDD) like ASD and ADHD. However, many models either use graph neural networks (GNN) to construct single-level brain functional networks (BFNs) or employ spatial convolution filtering for local information extraction from rs-fMRI data, often neglecting high-order features crucial for NDD classification. Methods: We introduce a Multi-view High-order Network (MHNet) to capture hierarchical and high-order features from multi-view BFNs derived from rs-fMRI data for NDD prediction. MHNet has two branches: the Euclidean Space Features Extraction (ESFE) module and the Non-Euclidean Space Features Extraction (Non-ESFE) module, followed by a Feature Fusion-based Classification (FFC) module for NDD identification. ESFE includes a Functional Connectivity Generation (FCG) module and a High-order Convolutional Neural Network (HCNN) module to extract local and high-order features from BFNs in Euclidean space. Non-ESFE comprises a Generic Internet-like Brain Hierarchical Network Generation (G-IBHN-G) module and a High-order Graph Neural Network (HGNN) module to capture topological and high-order features in non-Euclidean space. Results: Experiments on three public datasets show that MHNet outperforms state-of-the-art methods using both AAL1 and Brainnetome Atlas templates. Extensive ablation studies confirm the superiority of MHNet and the effectiveness of using multi-view fMRI information and high-order features. Our study also offers atlas options for constructing more sophisticated hierarchical networks and explains the association between key brain regions and NDD. Conclusion: MHNet leverages multi-view feature learning from both Euclidean and non-Euclidean spaces, incorporating high-order information from BFNs to enhance NDD classification performance.

Figures (8)

• Figure 1: The framework of the proposed MHNet. Multi-view data of rs-fMRI are used for the diagnosis of NDD.
• Figure 2: Training loss curves on three datasets. The plots show the convergence of the model during training: (a) ABIDE-I dataset with 240 epochs, (b) ABIDE-II dataset with 200 epochs, and (c) ADHD-200 dataset with 300 epochs.
• Figure 3: Illustration of the relationship curve between the percentage of retained edges and the cutoff threshold.
• Figure 4: Illustration of classification performance achievable with different cutoff thresholds.
• Figure 5: Comparision of the classification performance between using AAL1 and Brainnetome. * indicates a statistically significant differences.