arXiv eess.SP· Kunkun Zhang, Kehao Wu, Zhitao Chen, Haigen Hu, Qianwei Zhou·· 9 小时前AI 评分40
MOPC：免源数据的 EEG 运动想象在线测试时自适应框架
Memory-Augmented Online Test-Time Adaptation for EEG Motor Imagery via Perturbation Consistency and Confidence Calibration
AI 导读
研究提出免源数据框架 MOPC，让 EEG 脑机接口在推理阶段在线更新解码参数，实现单次前向自适应。该方法用随机掩码扰动约束预测一致性，并以置信度校准重加权损失抑制不可靠伪标签。在 BCI IV IIb 和 OpenBMI 数据集上分别超越此前方法 2.22% 和 3.67%，单试次自适应延迟约 40–60 ms。
正文 · 原文
Abstract:Recently, electroencephalography (EEG)-based brain--computer interfaces (BCIs) have gained traction as portable neural measurement systems. However, the practical deployment of these instruments is often hindered by nonstationarity and cross-subject distribution shifts, which lead to significant intersubject inconsistency. Existing adaptation methods remain impractical for real-time measurement scenarios, as they typically require offline access to source or aggregated target data. To address these challenges, this article proposes memory-augmented online test-time adaptation (OTTA) with perturbation consistency and confidence calibration (MOPC), a source-free framework designed to enhance the reliability of BCI instrumentation. MOPC enables the measurement system to update its decoding parameters on the fly during the inference stage, facilitating single-pass adaptation without accessing sensitive source data. We introduce random-mask perturbations to enforce prediction consistency against inherent EEG signal noise and acquisition artifacts. Furthermore, a confidence-calibrated reweighting loss is designed to integrate feature-space similarity, effectively mitigating unreliable pseudolabels and class imbalance during continuous signal acquisition. To ensure system stability and prevent the catastrophic forgetting of established neural patterns, a lightweight memory module is incorporated to regularize parameter updates within a tight latency budget. Evaluated on the BCI IV IIb and OpenBMI datasets, MOPC achieves state-of-the-art performance, outperforming prior methods by 2.22~\% and 3.67~\%, respectively, while keeping the per-trial adaptation latency to approximately 40--60 ms. This study provides a robust, recalibration-free solution for real-time EEG decoding, marking a significant advancement in the precision and usability of practical BCI measurement instruments.
| Comments: | Accepted for publication in IEEE Transactions on Instrumentation and Measurement (TIM), 2026 |
| Subjects: | Machine Learning (cs.LG); Signal Processing (eess.SP) |
| Cite as: | arXiv:2610.30422 [cs.LG] |
| (or arXiv:2610.30422v1 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2610.30422 arXiv-issued DOI via DataCite (pending registration) |
|
| Related DOI: | https://doi.org/10.1109/TIM.2026.3736160
DOI(s) linking to related resources |
Submission history
From: Kunkun Zhang [view email]
[v1]
Wed, 21 Oct 2026 20:32:45 UTC (1,458 KB)
来源：arXiv eess.SP · arxiv.org