Geometry- and Relation-Aware Diffusion for EEG Super-Resolution

2026-02-02

AI Summary

This work addresses the limitation of existing EEG spatial super-resolution methods, which often neglect the underlying physiological spatial structure, thereby constraining generation performance. To overcome this, the authors propose TopoDiff, a novel framework that, for the first time, integrates the global geometric context of EEG topographic maps and dynamic electrode relationship graphs into a latent diffusion generative model. This enables joint modeling of topology-aware embeddings and time-varying inter-channel dependencies. Evaluated across multiple benchmark datasets—including SEED, PhysioNet, and TUSZ—the method significantly enhances the fidelity and physiological plausibility of generated EEG signals and consistently improves performance on downstream tasks such as emotion recognition, motor imagery classification, and seizure detection.

Technology Category

• Computer Vision: Diffusion Models for Vision
• Machine Learning: Deep Generative Models & Autoencoders
• Knowledge Representation and Reasoning: Geometric, Spatial, and Temporal Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications

Abstract

Recent electroencephalography (EEG) spatial super-resolution (SR) methods, while showing improved quality by either directly predicting missing signals from visible channels or adapting latent diffusion-based generative modeling to temporal data, often lack awareness of physiological spatial structure, thereby constraining spatial generation performance. To address this issue, we introduce TopoDiff, a geometry- and relation-aware diffusion model for EEG spatial super-resolution. Inspired by how human experts interpret spatial EEG patterns, TopoDiff incorporates topology-aware image embeddings derived from EEG topographic representations to provide global geometric context for spatial generation, together with a dynamic channel-relation graph that encodes inter-electrode relationships and evolves with temporal dynamics. This design yields a spatially grounded EEG spatial super-resolution framework with consistent performance improvements. Across multiple EEG datasets spanning diverse applications, including SEED/SEED-IV for emotion recognition, PhysioNet motor imagery (MI/MM), and TUSZ for seizure detection, our method achieves substantial gains in generation fidelity and leads to notable improvements in downstream EEG task performance.

Problem

• EEG super-resolution
• spatial structure
• physiological topology
• channel relations
• generation fidelity

Innovation

• EEG super-resolution
• diffusion model
• topology-aware embedding
• dynamic channel-relation graph
• spatial geometry