FPED: A Functional-Network Prior-Guided Mixture-of-Experts Framework for Interpretable Brain Decoding

2026-05-18

AI Summary

Existing brain decoding approaches flatten fMRI signals, thereby disrupting the topological structure of functional brain networks and compromising neuroscientific interpretability. This work proposes the FPED framework, which for the first time integrates functional brain network priors into a Mixture-of-Experts (MoE) architecture by modeling each functional network as a specialized expert. An adaptive routing mechanism dynamically combines these experts’ complementary contributions to visual semantic processing, while multi-network fMRI features are aligned within the CLIP latent space. With only 0.68 billion parameters, FPED achieves high-performance semantic reconstruction and reveals biologically plausible mappings between functional brain networks and modality-specific semantic representations, substantially enhancing both model interpretability and alignment with established neuroscience principles.

Abstract

Visual image reconstruction from functional Magnetic Resonance Imaging (fMRI) is a fundamental task in brain decoding, providing a crucial pathway for understanding human perceptual mechanisms and developing advanced brain-computer interfaces (BCIs). However, most current methods simply flatten fMRI signals from localized visual cortices into one-dimensional (1D) vectors, mapping them directly into latent spaces such as that of Contrastive Language-Image Pre-training (CLIP). This paradigm not only disrupts the inherent network topology of the brain-leading to limited neuroscientific interpretability-but also overlooks the synergistic contributions of other distributed functional networks in processing high-level visual semantics. To address these limitations, we propose FPED, a Functional-Network Prior-Guided Mixture of Experts (MoE) framework for interpretable brain decoding. FPED explicitly models different functional brain networks as specialized experts and employs adaptive routing to capture their complementary contributions to visual semantic understanding. Unlike conventional homogeneous decoding paradigms, our framework incorporates neurobiologically grounded priors to enable structured and interpretable network-level representation learning. Experimental results demonstrate that FPED achieves highly competitive semantic reconstruction performance with only 0.68B parameters. The learned routing dynamics reveal biologically meaningful correspondence between functional brain networks and modality-specific semantic processing, providing transparent neuroscientific interpretability. This suggests that brain network-aware expert modeling is a promising direction for bridging neural decoding and biologically inspired artificial intelligence.

Problem

• brain decoding
• fMRI
• functional brain networks
• visual image reconstruction
• neuroscientific interpretability

Innovation

• Functional-Network Prior
• Mixture-of-Experts
• Interpretable Brain Decoding
• fMRI Reconstruction
• Neurobiologically Grounded AI