Novel Feature Extraction for Classification of Auditory-visual Stimuli from fNIRS Signals

This paper proposes a novel feature extraction method for functional near-infrared spectroscopy (fNIRS) signal analysis, aiming to enhance the robustness of Brain-Computer Interface (BCI) applications. Traditional techniques for fNIRS signal analysis often face challenges such as limited representation ability and sensitivity to noise. To overcome these limitations, we introduce a new approach that converts fNIRS signals into two-dimensional signals and utilizes a pre-trained convolutional neural network (CNN) model for image-based feature extraction. By effectively using the CNN-based feature representation ability, we can enhance the flexibility of fNIRS signal analysis, which has the potential to advance BCI technology.