I\textsuperscript{2}RiMA: Spectral Riemannian Representation with Temporal Attention for Mental Stress Detection based on EEG Signals

Short summary

I²RiMA is a neural network approach for detecting mental stress from EEG signals using spectral Riemannian geometry and attention mechanisms. It achieves 82.78% accuracy while remaining efficient at 1.6M parameters. The method addresses cross-subject variability by learning frequency-specific stress patterns.

• •Novel Riemannian manifold method for EEG-based stress detection
• •Outperforms five baselines with 82.78% balanced accuracy
• •Lightweight architecture: 1.60M parameters, 31.95M FLOPs

Generated with AI, which can make mistakes.