![[2602.16147] ASPEN: Spectral-Temporal Fusion for Cross-Subject Brain Decoding](https://arxiv.org/static/browse/0.3.4/images/arxiv-logo-fb.png){kind=logo}
[2602.16147] ASPEN: Spectral-Temporal Fusion for Cross-Subject Brain Decoding
Summary
The paper presents ASPEN, a novel architecture that enhances cross-subject brain decoding by integrating spectral and temporal features, demonstrating improved accuracy in EEG-based brain-computer interfaces.
Why It Matters
Cross-subject generalization in EEG-based brain-computer interfaces is crucial for developing effective BCIs that can work across different individuals. ASPEN's approach to combining spectral and temporal features could significantly improve the reliability and accuracy of these systems, making them more applicable in real-world scenarios.
Key Takeaways
- ASPEN utilizes a hybrid architecture for EEG signal processing.
- Spectral features provide more stable representations than temporal signals for cross-subject decoding.
- The model dynamically balances spectral and temporal inputs based on the EEG paradigm.
- ASPEN achieved superior accuracy on multiple benchmark datasets.
- The research highlights the importance of cross-modal agreement in feature propagation.
Computer Science > Machine Learning arXiv:2602.16147 (cs) [Submitted on 18 Feb 2026] Title:ASPEN: Spectral-Temporal Fusion for Cross-Subject Brain Decoding Authors:Megan Lee, Seung Ha Hwang, Inhyeok Choi, Shreyas Darade, Mengchun Zhang, Kateryna Shapovalenko View a PDF of the paper titled ASPEN: Spectral-Temporal Fusion for Cross-Subject Brain Decoding, by Megan Lee and 5 other authors View PDF HTML (experimental) Abstract:Cross-subject generalization in EEG-based brain-computer interfaces (BCIs) remains challenging due to individual variability in neural signals. We investigate whether spectral representations offer more stable features for cross-subject transfer than temporal waveforms. Through correlation analyses across three EEG paradigms (SSVEP, P300, and Motor Imagery), we find that spectral features exhibit consistently higher cross-subject similarity than temporal signals. Motivated by this observation, we introduce ASPEN, a hybrid architecture that combines spectral and temporal feature streams via multiplicative fusion, requiring cross-modal agreement for features to propagate. Experiments across six benchmark datasets reveal that ASPEN is able to dynamically achieve the optimal spectral-temporal balance depending on the paradigm. ASPEN achieves the best unseen-subject accuracy on three of six datasets and competitive performance on others, demonstrating that multiplicative multimodal fusion enables effective cross-subject generalization. Subjects: Machine Learn...
