Graph-variate flow matching generates high-resolution brain connectivity data
High-Resolution Dynamic Functional Connectivity Generation with Graph-Variate Flow Matching
Machine Learning
Summary
Measuring brain activity connections over very short time periods can be noisy and unclear. The authors developed a new way to create detailed brain connectivity data by combining quick interactions with stable longer-term patterns. Their method uses math that keeps the data realistic and smooth across time. It also uses a kind of AI model called a Transformer to generate these brain connection maps more accurately and efficiently than previous methods. This could help improve studies of how brain networks change quickly during tasks.
What this means in practice
- •For brain-computer interface developers: Generate realistic, high-resolution brain connectivity sequences to improve decoding models for motor imagery tasks without retraining for different time resolutions.
- •For medical imaging software teams: Create synthetic but realistic brain network data that capture temporal dynamics for testing and validating EEG analysis tools.
Authors
Om Roy, Yashar Moshfeghi, Keith Malcolm Smith
Abstract
High-resolution dynamic functional connectivity (DFC) can reveal rapidly evolving brain-network interactions, but short temporal windows yield noisy, often low-rank covariance estimates. Graph-Variate Dynamic (GVD) connectivity addresses this by modulating fast instantaneous interactions with stable trial-level support. This suppresses spurious fluctuations and emphasizes persistent, informative connections. We show that the Hadamard construction lifts low-rank instantaneous connectivity from the positive-semidefinite to the positive-definite cone, keeping high-resolution trajectories on the SPD manifold without ridge regularisation or post-hoc projection. We introduce GVD-CFM, a class-conditional generative model for high-resolution dynamic connectivity. Each trial is represented as SPD GVD matrices on a product Riemannian manifold, then mapped through a global log-Euclidean diffeomorphism and an invertible temporal DCT basis. A Transformer-based conditional flow models all spectral modes jointly and generates the full trajectory non-autoregressively in Euclidean coordinates while preserving exact correspondence with valid SPD sequences. Retaining the full DCT basis also enables decoding on denser temporal grids without retraining. Across multiple EEG motor-imagery datasets, GVD-CFM delivers the strongest overall results for held-out distributional fidelity, temporal-dynamics preservation, and synthetic-to-real classification. It also remains computationally efficient relative to strong raw-signal and direct GVD-space generative baselines. GVD-CFM therefore provides a practical framework for realistic, temporally coherent, high-resolution brain-network generation with preserved manifold structure and resolution-flexible decoding from a single trained model.