Making Brain-Computer Interfaces More Secure

In three linesStudy on adversarial robustness of EEG-based brain-computer interfaces. Researchers propose a lightweight CNN architecture and test it against adversarial attacks on two EEG datasets, comparing with EEGNet, DeepConvNet, and SleepEEGNet. The proposed model demonstrates improved robustness to perturbations.

Summary generated by Claude — human-verified