Brain signal analysis by Spiking neural networks

Anxiety disorders are among the most prevalent mental health conditions worldwide, with rising incidence and growing demand for effective, objective diagnostic tools. While electroencephalography (EEG) has emerged as a non-invasive and cost-efficient modality for detecting neurophysiological correlates of anxiety, most existing classification approaches rely on conventional machine learning algorithms, which often struggle to capture the temporal complexity of EEG signals. In this study, we investigate the feasibility of applying Spiking Neural Networks (SNNs) - a class of biologically inspired models capable of processing event-driven data - to the task of anxiety classification. Using the DASPS EEG dataset, we developed a custom SNN architecture based on Integrate-and-Fire neurons, implemented via the SpikingJelly framework. The model processes spike-encoded EEG inputs and outputs binary predictions of anxiety severity.