A comparative and critical study of EEGNet for fNIRS-driven cognitive load classification
A new study published on arXiv evaluates the effectiveness of EEGNet for classifying cognitive load using fNIRS signals. The research systematically examined various parameters, including temporal segmentation, window lengths, feature extraction methods, and learning rates. Results indicated that while overlapping segmentation and fixed learning rates achieved high accuracy in random-split experiments, subject-independent evaluation showed a significant drop in performance, highlighting generalization challenges. AI