ASPEN: Spectral-Temporal Fusion for Cross-Subject Brain Decoding

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Lee, Megan, Hwang, Seung Ha, Choi, Inhyeok, Darade, Shreyas, Zhang, Mengchun, Shapovalenko, Kateryna
Format: Preprint
Published: 2026
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914336243449856
author Lee, Megan
Hwang, Seung Ha
Choi, Inhyeok
Darade, Shreyas
Zhang, Mengchun
Shapovalenko, Kateryna
author_facet Lee, Megan
Hwang, Seung Ha
Choi, Inhyeok
Darade, Shreyas
Zhang, Mengchun
Shapovalenko, Kateryna
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2602_16147
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle ASPEN: Spectral-Temporal Fusion for Cross-Subject Brain Decoding
Lee, Megan
Hwang, Seung Ha
Choi, Inhyeok
Darade, Shreyas
Zhang, Mengchun
Shapovalenko, Kateryna
Machine Learning
Artificial Intelligence
Human-Computer Interaction
Signal Processing
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.
title ASPEN: Spectral-Temporal Fusion for Cross-Subject Brain Decoding
topic Machine Learning
Artificial Intelligence
Human-Computer Interaction
Signal Processing
url https://arxiv.org/abs/2602.16147