Train-On-Request: An On-Device Continual Learning Workflow for Adaptive Real-World Brain Machine Interfaces

TL;DR

This work proposes TOR, a train-on-request workflow that enables user-specific model adaptation to novel conditions, addressing signal variability over time, and is the first demonstration of an online, energy-efficient, dynamic adaptation of a BMI model to the intrinsic variability of EEG signals in real-time settings.

Abstract

Brain-machine interfaces (BMIs) are expanding beyond clinical settings thanks to advances in hardware and algorithms. However, they still face challenges in user-friendliness and signal variability. Classification models need periodic adaptation for real-life use, making an optimal re-training strategy essential to maximize user acceptance and maintain high performance. We propose TOR, a train-on-request workflow that enables user-specific model adaptation to novel conditions, addressing signal variability over time. Using continual learning, TOR preserves knowledge across sessions and mitigates inter-session variability. With TOR, users can refine, on demand, the model through on-device learning (ODL) to enhance accuracy adapting to changing conditions. We evaluate the proposed methodology on a motor-movement dataset recorded with a non-stigmatizing wearable BMI headband, achieving up to 92% accuracy and a re-calibration time as low as 1.6 minutes, a 46% reduction compared to a naive transfer learning workflow. We additionally demonstrate that TOR is suitable for ODL in extreme edge settings by deploying the training procedure on a RISC-V ultra-low-power SoC (GAP9), resulting in 21.6 ms of latency and 1 mJ of energy consumption per training step. To the best of our knowledge, this work is the first demonstration of an online, energy-efficient, dynamic adaptation of a BMI model to the intrinsic variability of EEG signals in real-time settings.

Figures (6)

• Figure 1: The workflow: we pretrain our model on Session 1. For a new session, the user utilizes the model as long as the model performance is satisfactory, exemplified here by a passing test on subsession SS2.1. When the model performance is not satisfactory, e.g., on SS2.2, the model is finetuned on the subsequent subsession SS2.3. Once the model is adapted to the new data, the user can continue to use it for the remaining session.
• Figure 2: Test accuracy over multiple sessions for chain- baseline with different train/test data splits.
• Figure 3: Required training trials over multiple sessions for - and -based workflows with $T_{Acc} = 90\%$, $trls=10$. The horizontal dashed line represents the chain- baseline.
• Figure 4: Unsuccessful (a) and successful (b) workflows for $subss=10$ subsessions, $T_{Acc}=90\%$, $trls=10$ trials.
• Figure 5: Test accuracy over multiple sessions for - and -based workflows, and chain- baseline.