2021/03/04 by Byoung-Hee Kwon, Kwon, Byoung-Hee, Ji-Hoon Jeong +3
Neuroscience · Computer Science · Engineering · #EEG and Brain-Computer Interfaces #Gaze Tracking and Assistive Technology #Advanced Memory and Neural Computing
paper · pdf · doi:10.48550/arxiv.2103.02851
Brain-computer interfaces (BCIs) use brain signals such as\nelectroencephalography to reflect user intention and enable two-way\ncommunication between computers and users. BCI technology has recently received\nmuch attention in healthcare applications, such as neurorehabilitation and\ndiagnosis. BCI applications can also control external devices using only brain\nactivity, which can help people with physical or mental disabilities,\nespecially those suffering from neurological and neuromuscular diseases such as\nstroke and amyotrophic lateral sclerosis. Motor imagery (MI) has been widely\nused for BCI-based device control, but we adopted intuitive visual motion\nimagery to overcome the weakness of MI. In this study, we developed a\nthree-dimensional (3D) BCI training platform to induce users to imagine\nupper-limb movements used in real-life activities (picking up a cell phone,\npouring water, opening a door, and eating food). We collected intuitive visual\nmotion imagery data and proposed a deep learning network based on functional\nconnectivity as a mind-reading technique. As a result, the proposed network\nrecorded a high classification performance on average (71.05%). Furthermore, we\napplied the leave-one-subject-out approach to confirm the possibility of\nimprovements in subject-independent classification performance. This study will\ncontribute to the development of BCI-based healthcare applications for\nrehabilitation, such as robotic arms and wheelchairs, or assist daily life.\n