Brain-MGF: Multimodal Graph Fusion Network for EEG-fMRI Brain Connectivity Analysis Under Psilocybin

2025-11-23

AI Summary

This study investigates how psilocybin reshapes large-scale functional brain connectivity and characterizes its neurovascular coupling mechanisms across multimodal EEG–fMRI networks. To address the challenge of integrating heterogeneous neuroimaging modalities, we propose an adaptive Softmax-gated fusion framework that enables individualized, context-sensitive weighting of EEG and fMRI data. Leveraging partial correlation graph modeling and Pearson-spectrum-based node embedding, we employ graph convolutional networks to learn subject-level representations. Our method significantly enhances interpretability and discriminative power: it achieves 76.0% classification accuracy and 85.8% ROC-AUC in distinguishing psilocybin-administered versus placebo conditions during both resting-state and meditation tasks. UMAP visualization confirms improved class separability in the fused representation space. This work establishes the first interpretable, multimodal graph-network modeling paradigm for elucidating the neural mechanisms of psychedelic compounds.

Technology Category

• Machine Learning: Multimodal Learning
• Computer Vision: Multi-modal Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Psychedelics, such as psilocybin, reorganise large-scale brain connectivity, yet how these changes are reflected across electrophysiological (electroencephalogram, EEG) and haemodynamic (functional magnetic resonance imaging, fMRI) networks remains unclear. We present Brain-MGF, a multimodal graph fusion network for joint EEG-fMRI connectivity analysis. For each modality, we construct graphs with partial-correlation edges and Pearson-profile node features, and learn subject-level embeddings via graph convolution. An adaptive softmax gate then fuses modalities with sample-specific weights to capture context-dependent contributions. Using the world's largest single-site psilocybin dataset, PsiConnect, Brain-MGF distinguishes psilocybin from no-psilocybin conditions in meditation and rest. Fusion improves over unimodal and non-adaptive variants, achieving 74.0% accuracy and 76.5% F1 score on meditation, and 76.0% accuracy with 85.8% ROC-AUC on rest. UMAP visualisations reveal clearer class separation for fused embeddings. These results indicate that adaptive graph fusion effectively integrates complementary EEG-fMRI information, providing an interpretable framework for characterising psilocybin-induced alterations in large-scale neural organisation.

Problem

• Analyzing psilocybin-induced brain connectivity changes across EEG and fMRI modalities
• Developing multimodal graph fusion to integrate complementary electrophysiological and haemodynamic data
• Characterizing psychedelic reorganization of neural networks using adaptive fusion framework

Innovation

• Multimodal graph fusion network for EEG-fMRI connectivity
• Adaptive softmax gate fuses modalities with sample-specific weights
• Graph convolution learns subject-level embeddings from brain networks