Overcoming subject variability in P300 decoding based on quantification by deep learning
- 1School of Electrical and Computer Engineering, University of Georgia
Abstract
Binary P300 decoding is a cornerstone of P300-based brain-computer interfaces, enabling applications such as assistive communication, neurorehabilitation, and cognitive monitoring. However, substantial subject variability in EEG patterns poses a major challenge to the development of generalized P300 decoders. To address this problem, this study proposes SQ_EMoE, a subject-variability-aware enhanced Mixture-of-Experts framework that uses the confidence produced by a deep subject-identification network as a practical proxy for trial-level subject distinctiveness. High-confidence trials are routed to subject-specific experts, whereas uncertain trials are redirected to a pooled general expert, allowing the framework to dynamically balance subject-specific specialization and cross-subject generalization. Latent subject representations are further visualized to qualitatively examine subject separability and its relationship with routing behavior. Experiments on eight random subjects from each of two public P300 datasets yielded mean ROC–AUC values of approximately 0.86 on BNCI Horizon 2020 and 0.89 on P300-LINI, outperforming pooled EEGNet and architecture-matched pooled P300MCNN baselines. Ablation experiments further demonstrated the contribution of confidence-controlled fallback to the general expert. To evaluate scalability, the framework was additionally tested on all 22 subjects of P300-LINI. With the original subject-identification network, SQ_EMoE achieved a mean ROC–AUC of 0.8518; after strengthening the subject representation and identification network, performance increased to 0.8760, exceeding the pooled P300MCNN baseline (0.8482) and slightly surpassing the oracle routing or within-subject decoding (0.8744). Leave-one-subject-out experiments further showed that the under-realized SQ_EMoE conservatively redirects most completely unseen-subject trials to the general expert when reliable subject-specific routing is unavailable. These results demonstrate that effective modeling of subject variability, together with adaptive integration of shared and subject-specific representations, provides a promising alternative to conventional one-size-fits-all EEG decoding.
Related
- Leakage-Audited Benchmarking Reveals Limited Evidence for Cross-Subject Auditory-Evoked EEG Vowel Perception Decoding
- Cross-subject decoding of imagined speech from scalp EEG: a leakage-controlled evaluation on three public corpora
- Accounting for Variability Across Years, Depths, and Non‐Target Reads Improves Quantification of Root Community Composition in Plant Mixtures
- Window leakage in clinical EEG machine learning: matched subject-independent evaluation across ADHD and Parkinson cohorts
- Electroencephalography-based Alzheimer's disease detection using multi-scale information retention transform: a subject-level comparison of machine learning classifiers
BibTeX
@article{liu2026overcoming,
title = {{Overcoming subject variability in P300 decoding based on quantification by deep learning}},
author = {Maohua Liu and Fred R. Beyette},
journal = {Frontiers in Neuroscience},
year = {2026},
month = oct,
doi = {10.3389/fnins.2026.1940267},
url = {https://www.frontiersin.org/articles/10.3389/fnins.2026.1940267/full},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Overcoming subject variability in P300 decoding based on quantification by deep learning
AU - Maohua Liu
AU - Fred R. Beyette
PY - 2026
DA - 2026/10/22
JO - Frontiers in Neuroscience
DO - 10.3389/fnins.2026.1940267
UR - https://www.frontiersin.org/articles/10.3389/fnins.2026.1940267/full
AB - Binary P300 decoding is a cornerstone of P300-based brain-computer interfaces, enabling applications such as assistive communication, neurorehabilitation, and cognitive monitoring. However, substantial subject variability in EEG patterns poses a major challenge to the development of generalized P300 decoders. To address this problem, this study proposes SQ_EMoE, a subject-variability-aware enhanced Mixture-of-Experts framework that uses the confidence produced by a deep subject-identification network as a practical proxy for trial-level subject distinctiveness. High-confidence trials are routed to subject-specific experts, whereas uncertain trials are redirected to a pooled general expert, allowing the framework to dynamically balance subject-specific specialization and cross-subject generalization. Latent subject representations are further visualized to qualitatively examine subject separability and its relationship with routing behavior. Experiments on eight random subjects from each of two public P300 datasets yielded mean ROC–AUC values of approximately 0.86 on BNCI Horizon 2020 and 0.89 on P300-LINI, outperforming pooled EEGNet and architecture-matched pooled P300MCNN baselines. Ablation experiments further demonstrated the contribution of confidence-controlled fallback to the general expert. To evaluate scalability, the framework was additionally tested on all 22 subjects of P300-LINI. With the original subject-identification network, SQ_EMoE achieved a mean ROC–AUC of 0.8518; after strengthening the subject representation and identification network, performance increased to 0.8760, exceeding the pooled P300MCNN baseline (0.8482) and slightly surpassing the oracle routing or within-subject decoding (0.8744). Leave-one-subject-out experiments further showed that the under-realized SQ_EMoE conservatively redirects most completely unseen-subject trials to the general expert when reliable subject-specific routing is unavailable. These results demonstrate that effective modeling of subject variability, together with adaptive integration of shared and subject-specific representations, provides a promising alternative to conventional one-size-fits-all EEG decoding.
SN - 1662-453X
ER -