Search papers, labs, and topics across Lattice.
This study introduces a dynamic multimodal Mixture-of-Experts (MoE) framework that integrates functional and structural MRI data to improve the early diagnosis of post-traumatic epilepsy (PTE). By employing time-aware functional-structural encoding and class-conditioned expert routing, the framework allows for the learning of complementary representations from both modalities, leading to enhanced diagnostic accuracy. Experimental results show that this approach outperforms static fusion baselines across three binary classification tasks while providing interpretable insights into region-of-interest interactions in the brain.
Dynamic expert routing in multimodal MRI analysis reveals class-dependent brain interaction patterns that significantly enhance PTE diagnosis.
Post-traumatic epilepsy (PTE) is a severe complication of traumatic brain injury (TBI), yet early identification remains challenging due to the complex structural and functional alterations it induces in the brain. To address this, we propose a dynamic multimodal Mixture-of-Experts (MoE) framework that integrates functional and structural MRI through time-aware functional-structural encoding and class-conditioned expert routing. Within this framework, modality-specific and cross-modal experts learn complementary representations, while a Modality-Class MoE (MCoE) module dynamically dispatches expert weights according to each classification objective. Experimental results across three binary classification tasks demonstrate that the framework consistently outperforms static fusion baselines, and high-interpretability analyses further reveal meaningful region-of-interest (ROI) interactions. This dynamic multimodal expert framework effectively captures class-dependent brain interaction patterns and provides an interpretable approach for PTE diagnosis and risk stratification.
