ExplainabilityFairnessSafetycs.OH
FBCNet: A Multi-view Convolutional Neural Network for Brain-Computer Interface
Ravikiran Mane, Effie Chew, Karen Chua, Kai Keng Ang, Neethu Robinson, A. P. Vinod, Seong-Whan Lee, Cuntai Guan
March 17, 2021
arXiv.org
267 citations
41 influential
Key Takeaways
-
Improved accuracy with limited data: FBCNet achieves state-of-the-art performance (76.20% accuracy on standard benchmarks) while requiring less training data than previous methods, addressing a practical barrier to deploying brain-computer interfaces in real-world settings where collecting extensive EEG recordings is difficult and costly.
-
Transparency in medical AI: The researchers used explainable AI techniques to identify and report differences in brain signal patterns between healthy individuals and stroke patients, demonstrating how responsible AI development can provide interpretable insights rather than "black box" predictions—crucial for clinical acceptance and trust.
-
Open science approach: By making FBCNet source code publicly available, the authors enable reproducibility and allow other researchers to build upon their work, supporting responsible AI practices through transparency and community collaboration in medical technology development.
-
Practical design for real patients: Testing on datasets from chronic stroke patients (not just healthy volunteers) shows the algorithm works on the populations who would actually benefit from brain-computer interfaces, reflecting responsible AI's emphasis on real-world applicability and inclusive testing beyond idealized conditions.
Abstract
Lack of adequate training samples and noisy high-dimensional features are key challenges faced by Motor Imagery (MI) decoding algorithms for electroencephalogram (EEG) based Brain-Computer Interface (BCI). To address these challenges, inspired from neuro-physiological signatures of MI, this paper proposes a novel Filter-Bank Convolutional Network (FBCNet) for MI classification. FBCNet employs a multi-view data representation followed by spatial filtering to extract spectro-spatially discriminative features. This multistage approach enables efficient training of the network even when limited training data is available. More significantly, in FBCNet, we propose a novel Variance layer that effectively aggregates the EEG time-domain information. With this design, we compare FBCNet with state-of-the-art (SOTA) BCI algorithm on four MI datasets: The BCI competition IV dataset 2a (BCIC-IV-2a), the OpenBMI dataset, and two large datasets from chronic stroke patients. The results show that, by achieving 76.20% 4-class classification accuracy, FBCNet sets a new SOTA for BCIC-IV-2a dataset. On the other three datasets, FBCNet yields up to 8% higher binary classification accuracies. Additionally, using explainable AI techniques we present one of the first reports about the differences in discriminative EEG features between healthy subjects and stroke patients. Also, the FBCNet source code is available at https://github.com/ravikiran-mane/FBCNet.