Temporal simplicial neural network improves EEG brain state decoding
T-SNN: Temporal Simplicial Neural Network for EEG Decoding
Machine Learning
Summary
Decoding brain states from EEG signals requires understanding both how brain activity changes over time and how different brain regions interact. Existing methods often miss complex group interactions. The authors present a new neural network called T-SNN that models these higher-order interactions as evolving shapes that change over time. This approach outperforms other methods in recognizing emotions from EEG data and works even better when combined with eye movement information.
What this means in practice
- •For brain-computer interface developers: Enhance emotion recognition systems by modeling complex brain region interactions over time using the T-SNN framework.
- •For multimodal neurotechnology teams: Combine eye movement data with EEG signals using T-SNN to improve decoding accuracy for brain-state monitoring devices.
Authors
Nikita Malik, Shubhajit Roy, Mohit Kataria, Isuru Herath, Suraj Yadav, Inés García-Redondo, Dhananjay Bhaskar
Abstract
Decoding brain states requires models that capture both the evolution of neural activity and interactions among groups of brain regions. Existing EEG methods often treat recordings as multivariate time series or represent functional connectivity with pairwise graphs, leaving dynamic higher-order interactions largely unmodeled. We introduce the Temporal Simplicial Neural Network (T-SNN), which represents EEG recordings as sequences of evolving simplicial complexes. By combining simplicial convolutions with recurrent updates, T-SNN jointly learns higher-order interactions and their temporal evolution. On the seven-class SEED-VII emotion recognition task, T-SNN outperforms convolutional, recurrent, graph-based, and Transformer methods in both trial-wise and cross-subject evaluations. Incorporating eye-movement features further improves performance, demonstrating the framework's potential for multimodal brain-state decoding.