Hemodynamic Response Detection Using Integrated EEG-fNIRS-VPA for BCI

For BCI systems, it is important to have an accurate and less complex architecture control a device with enhanced accuracy. In this paper, novel methodology for more detection of the hemodynamic response has been developed using multimodal brain-computer interface (BCI). An integrated classifier achieving better classification accuracy two modalities. EEG-fNIRS-based vector-phase analysis (VPA) conducted. open-source dataset collected at Technische Universität Berlin, including simultaneous electroencephalography (EEG) functional near-infrared spectroscopy (fNIRS) signals 26 healthy participants during n-back tests, used research. Instrumental physiological noise removal done preprocessing techniques followed by individually detecting activity in both With resting state threshold circle, VPA detect fNIRS signals, whereas phase plots EEG constructed Hilbert Transform each trial. Multiple circles are drawn vector plane, where circle after task completion trial signal. Finally, processes into one plot get combined activity. Results study illustrate that EEG-fNIRS yields considerably higher average accuracy, 91.35%, as compared other classifiers such support machine (SVM), convolutional neural networks (CNN), deep (DNN) (with dual-threshold circles) accuracies 82%, 89%, 87% 86% respectively. Outcomes research demonstrate improved performance can be feasibly achieved hybrid data.