Timely and accurate detection of seizures from Electroencephalogram (EEG) signals is critical for the effective management of epilepsy. Although deep learning and artificial neural networks (ANNs) have substantially advanced automated seizure detection, their reliance on dense and continuous computations results in high energy consumption, thereby hindering practical clinical deployment. To address this challenge, we propose Spatio Temporal Attention with Spiking Neural Networks (STASNN), a novel architecture that integrates the strong feature extraction capability of attention mechanisms with the event driven energy efficiency of biologically inspired neurons. The STASNN framework comprises two complementary modules, namely Spatial Attention with SNN (SASNN) and Temporal Attention with SNN (TASNN), which are designed to capture inter channel spatial dependencies and long-range temporal dynamics, respectively. By encoding information through sparse binary spikes, the proposed method significantly reduces theoretical energy consumption while maintaining high detection accuracy. Experimental results on the CHB-MIT and Siena datasets demonstrate that STASNN outperforms existing state-of-the-art methods. The code is publicly available at https://github.com/peutim564-gif/STASNN.
Liu et al. (Thu,) studied this question.