Fusion Learning from Dynamic Functional Connectivity: Combining the Amplitude and Phase of fMRI Signals to Identify Brain Disorders
Researchers have developed a new framework called MSFL that combines amplitude and phase information from fMRI signals to improve the detection of brain disorders. This multi-scale fusion learning approach leverages both sliding window correlation (SWC) for amplitude correlations and phase synchronization (PS) for phase coherence. When tested on datasets for autism spectrum disorder and major depressive disorder, MSFL demonstrated superior performance compared to existing models, with analysis indicating that both SWC and PS features contribute to accurate classification. AI
IMPACT This research introduces a novel fusion learning framework for analyzing fMRI data, potentially enhancing diagnostic capabilities for neurological and psychiatric conditions.