A Cyclic Adaptation-Generalization Framework with Uncertainty-Guided Self-Paced Learning for Long-Term Brain-Machine Interfaces
Jiyu Wei, Di Hong, Zhanjie Zhang, Dazhong Rong, Qinming He, Yueming Wang
AI summary
Problem
Invasive brain-machine interfaces suffer from progressive neural drift that degrades decoding accuracy over time. Existing methods treat domain adaptation and generalization as isolated processes, failing to capture fine-grained subdomain shifts and limiting long-term robustness.
Approach
The authors propose UnSPC, which uses uncertainty estimates to progressively mine reliable pseudo-labeled samples from unlabeled target sessions. These samples drive a cyclic optimization process that alternates between domain adaptation and domain generalization to align with evolving neural distributions while preserving robust representations.
Key results
- First framework to cyclically integrate domain adaptation and generalization with pseudo-labeling for neural decoding
- Uncertainty-guided self-paced pseudo-labeling effectively mines reliable samples and prevents error propagation
- Cyclic optimization mitigates both global and subdomain neural drift through complementary adaptation and generalization
- Significantly outperforms state-of-the-art baselines across multiple non-human primate datasets without target labels
Why it matters
Enables stable, recalibration-free long-term control for invasive brain-computer interfaces, advancing neurorehabilitation and human-robot interaction applications.
Abstract
Brain-Machine Interfaces (BMIs), which link the brain to external devices, hold great potential in rehabilita- tion, human performance augmentation, and human-centered robotics. However, invasive BMIs face a critical challenge for long-term deployment due to neural drift, which degrades decoding performance over time and necessitates frequent recalibration. Existing methods designed to mitigate neural drift typically rely on either domain adaptation (DA) or domain generalization (DG) alone and often fail to capture fine- grained distribution shifts across neural subdomains, resulting in limited performance. To overcome these limitations, we propose Uncertainty-guided Self-paced Cycling (UnSPC), a robust framework that synergizes DA and DG for target do- main refining under an Uncertainty-guided Self-paced Pseudo- labeling (UnSPL) mechanism. To handle subdomain neural drift across domains, UNSPL is proposed to iteratively mine reliable pseudo-labeled samples with a noise-robust ranking strategy for further fine-tuning. Leveraging these high-quality samples, we introduce a novel Cycling Adaptation and Generalization (CycAG) strategy, which integrates DA and DG within an iter- ative cycle to progressively mitigate both global and subdomain drift. This cyclic process enables effective alignment to evolving target distributions while preserving robust and transferable representations, thereby mitigating performance degradation under long-term neural drifts. Extensive experiments on multi- ple neural decoding datasets demonstrate the effectiveness and robustness of UnSPC. To our knowledge, our proposed UnSPC is the first to cyclically integrate DA and DG with pseudo- labeling, paving the way toward stable long-term BMI controls.