Effect of subject training on a movement-related cortical potential-based brain-computer interface. (March 2018)

• Record Type: Journal Article
• Authors: Jochumsen, Mads; Niazi, Imran Khan; Nedergaard, Rasmus Wiberg; Navid, Muhammad Samran; Dremstrup, Kim
• Is Part Of: Biomedical signal processing and control. Volume 41(2018)
• Page Start: 63
• Page End: 68
• Publication Date: 2018-03
• Subjects: Brain-computer interface -- Movement-related cortical potential -- Movement intention -- Training -- Movement kinetics; Signal processing -- Periodicals; Biomedical engineering -- Periodicals; Signal Processing, Computer-Assisted -- Periodicals; Image Processing, Computer-Assisted -- Periodicals; Biomedical Engineering -- Periodicals; 610.28
• DOI: 10.1016/j.bspc.2017.11.012
• Languages: English
• ISSNs: 1746-8094

Abstract

Graphical abstract: Highlights: Online detection of MRCPs was stable over eight weeks. Subject (healthy) training did not significantly affect BCI performance. The morphology of MRCPs was similar over eight weeks. There was no difference in BCI performance when using new and old data for calibration. All subjects could operate the BCI. Abstract: Brain-computer interfaces (BCIs) have been developed for several purposes in communication, control, and rehabilitation. To use the BCI efficiently, the system must be technically tuned, and the user must learn to operate it. In this study, we investigated if the user could be trained to improve the performance of online detection of movement-related cortical potentials (MRCPs) associated with fast and slow movements. Seven healthy subjects participated in nine experiments over eight weeks while the ability of the online system to detect the movements was accessed. The movements were detected using template matching. No training effect was observed on the performance or MRCP morphology over the eight weeks. The system correctly detected ∼80% of the movements with ∼1.5 false positive detections/min. The findings suggest that the detection of MRCPs is stable from the first session and that several training sessions are not needed to obtain control of the BCI; this may have implication for the applicability of BCIs for movement detection.