BrainNet-MoE: Brain-Inspired Mixture-of-Experts Learning for Neurological Disease Identification
🤖 AI Summary
Alzheimer’s disease (AD) and dementia with Lewy bodies (LBD) exhibit substantial clinical overlap and pathological heterogeneity, compounded by the scarcity of LBD cases, rendering early differential diagnosis highly challenging. To address this, we propose a brain-inspired hierarchical system-level Mixture-of-Experts (MoE) network—the first to integrate whole-brain functional subnetwork modeling, disease-specific expert grouping, dynamic gating, and cross-subnetwork Transformer-based communication. Methodologically, it employs region-wise decoupled modeling for interpretable whole-brain representation; mechanistically, it introduces disease-aware gating to guide expert specialization and enhance discriminative specificity; architecturally, it enables long-range inter-subnetwork interactions to capture cross-regional pathological synergies. Experiments demonstrate significant improvements in AD/LBD classification accuracy and enable quantitative attribution of diagnostic contributions from individual functional subnetworks. This work establishes a mechanism-driven paradigm for differential diagnosis of neurodegenerative disorders.
Application Category
📝 Abstract
The Lewy body dementia (LBD) is the second most common neurodegenerative dementia after Alzheimer's disease (AD). Early differentiation between AD and LBD is crucial because they require different treatment approaches, but this is challenging due to significant clinical overlap, heterogeneity, complex pathogenesis, and the rarity of LBD. While recent advances in artificial intelligence (AI) demonstrate powerful learning capabilities and offer new hope for accurate diagnosis, existing methods primarily focus on designing"neural-level networks". Our work represents a pioneering effort in modeling system-level artificial neural network called BrainNet-MoE for brain modeling and diagnosing. Inspired by the brain's hierarchical organization of bottom-up sensory integration and top-down control, we design a set of disease-specific expert groups to process brain sub-network under different condition, A disease gate mechanism guides the specializa-tion of expert groups, while a transformer layer enables communication be-tween all sub-networks, generating a comprehensive whole-brain represen-tation for downstream disease classification. Experimental results show superior classification accuracy with interpretable insights into how brain sub-networks contribute to different neurodegenerative conditions.
Problem
Research questions and friction points this paper is trying to address.
Differentiate Lewy body dementia from Alzheimer's disease early.
Model system-level neural networks for brain disease diagnosis.
Enhance classification accuracy with interpretable brain sub-network insights.
Innovation
Methods, ideas, or system contributions that make the work stand out.
BrainNet-MoE models system-level neural networks
Disease-specific expert groups process brain sub-networks
Transformer layer enables communication between sub-networks
🔎 Similar Papers
No similar papers found.
