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

Psychedelics like psilocybin reorganize large-scale brain connectivity, but how these changes appear across EEG and fMRI networks has been unclear. A new multimodal graph fusion network, Brain-MGF, jointly analyzes EEG-fMRI connectivity by constructing graphs with partial-correlation edges and Pearson-profile node features, then learning subject-level embeddings via graph convolution. An adaptive softmax gate fuses modalities with sample-specific weights. Tested on the world's largest single-site psilocybin dataset, PsiConnect, the model distinguishes psilocybin from no-psilocybin conditions during meditation and rest. Fusion achieves 74.0% accuracy and 76.5% F1 score on meditation, and 76.0% accuracy with 85.8% ROC-AUC on rest, improving over unimodal and non-adaptive variants. UMAP visualizations show clearer class separation for fused embeddings, suggesting adaptive graph fusion effectively integrates complementary EEG-fMRI information for characterizing psilocybin-induced neural reorganization.

Study at a glance

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.