EEG-TCNet: An Accurate Temporal Convolutional Network for Embedded Motor-Imagery Brain-Machine Interfaces

TL;DR

EEG-TCNet is a lightweight, accurate temporal convolutional network designed for embedded motor-imagery brain-machine interfaces, enabling efficient EEG classification on resource-limited devices with high generalization across datasets.

Contribution

The paper introduces EEG-TCNet, a novel TCN architecture that balances high accuracy with low computational complexity for embedded EEG-based MI classification.

Findings

Achieves 77.35% accuracy on BCI Competition IV-2a dataset

Improves to 83.84% accuracy with subject-specific hyperparameters

Outperforms state-of-the-art on MOABB benchmark

Abstract

In recent years, deep learning (DL) has contributed significantly to the improvement of motor-imagery brain-machine interfaces (MI-BMIs) based on electroencephalography(EEG). While achieving high classification accuracy, DL models have also grown in size, requiring a vast amount of memory and computational resources. This poses a major challenge to an embedded BMI solution that guarantees user privacy, reduced latency, and low power consumption by processing the data locally. In this paper, we propose EEG-TCNet, a novel temporal convolutional network (TCN) that achieves outstanding accuracy while requiring few trainable parameters. Its low memory footprint and low computational complexity for inference make it suitable for embedded classification on resource-limited devices at the edge. Experimental results on the BCI Competition IV-2a dataset show that EEG-TCNet achieves 77.35% classification accuracy in 4-class MI. By finding the optimal network hyperparameters per subject, we further improve the accuracy to 83.84%. Finally, we demonstrate the versatility of EEG-TCNet on the Mother of All BCI Benchmarks (MOABB), a large scale test benchmark containing 12 different EEG datasets with MI experiments. The results indicate that EEG-TCNet successfully generalizes beyond one single dataset, outperforming the current state-of-the-art (SoA) on MOABB by a meta-effect of 0.25.