Brain-inspired Spiking Neural Network Frameworks for Multimodal Spatiotemporal Brain Data Integration: A Case Study on EEG-fMRI Data
This study proposes a biologically inspired Spiking Neural Network framework utilizing early, intermediate, and a novel explainable X-Meta late fusion strategy to effectively integrate heterogeneous EEG and fMRI data, demonstrating superior accuracy and biological interpretability in decoding eyes-open versus eyes-closed brain states compared to conventional machine learning approaches.