EDAPT: Towards Calibration-Free BCIs with Continual Online Adaptation

2025-08-14

AI Summary

Brain–computer interfaces (BCIs) suffer from limited practical deployment due to neural signal non-stationarity and substantial inter-subject variability, necessitating frequent offline calibration. This work proposes the first task- and model-agnostic continual online adaptation framework for BCIs, integrating multi-subject pretraining, supervised online fine-tuning, and unsupervised domain adaptation to achieve fully calibration-free operation. The framework incurs sub-200 ms model update latency on consumer-grade hardware and supports cross-task generalization. Evaluated on nine public BCI datasets, it demonstrates continuous decoding accuracy improvement with accumulating data—significantly outperforming static baselines—without reliance on subject-specific or trial-specific data splits. To our knowledge, this is the first BCI system achieving true “plug-and-play” usability: zero calibration, real-time adaptation, and robust performance across users and tasks.

Technology Category

• Cognitive Modeling & Cognitive Systems: Adaptive Behavior
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Humans and AI: Brain-Sensing and Analysis

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Economics, Online Markets and Human Computation: Economic ramifications for generative AI infrastructure and applications
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Brain-computer interfaces (BCIs) suffer from accuracy degradation as neural signals drift over time and vary across users, requiring frequent recalibration that limits practical deployment. We introduce EDAPT, a task- and model-agnostic framework that eliminates calibration through continual model adaptation. EDAPT first trains a baseline decoder using data from multiple users, then continually personalizes this model via supervised finetuning as the neural patterns evolve during use. We tested EDAPT across nine datasets covering three BCI tasks, and found that it consistently improved accuracy over conventional, static methods. These improvements primarily stem from combining population-level pretraining and online continual finetuning, with unsupervised domain adaptation providing further gains on some datasets. EDAPT runs efficiently, updating models within 200 milliseconds on consumer-grade hardware. Finally, decoding accuracy scales with total data budget rather than its allocation between subjects and trials. EDAPT provides a practical pathway toward calibration-free BCIs, reducing a major barrier to BCI deployment.

Problem

• Reduces BCI accuracy loss from neural signal drift
• Eliminates need for frequent recalibration in BCIs
• Improves decoding accuracy via continual online adaptation

Innovation

• Continual online adaptation for calibration-free BCIs
• Population-level pretraining with personalized finetuning
• Efficient model updates within 200 milliseconds