Bloorview Research Institute, Holland Bloorview Kids Rehabilitation Hospital, Canada.
Institute of Biomaterials and Biomedical Engineering, University of Toronto, 27 King's College Circle, Toronto, Ontario, Canada M5S 1A1, Canada.
Int J Neural Syst. 2020 Jun;30(6):2050026. doi: 10.1142/S0129065720500264.
Brain-computer interfaces (BCIs) can provide a means of communication to individuals with severe motor disorders, such as those presenting as locked-in. Many BCI paradigms rely on motor neural pathways, which are often impaired in these individuals. However, recent findings suggest that visuospatial function may remain intact. This study aimed to determine whether visuospatial imagery, a previously unexplored task, could be used to signify intent in an online electroencephalography (EEG)-based BCI. Eighteen typically developed participants imagined checkerboard arrow stimuli in four quadrants of the visual field in 5-s trials, while signals were collected using 16 dry electrodes over the visual cortex. In online blocks, participants received graded visual feedback based on their performance. An initial BCI pipeline (visuospatial imagery classifier I) attained a mean accuracy of [Formula: see text]% classifying rest against visuospatial imagery in online trials. This BCI pipeline was further improved using restriction to alpha band features (visuospatial imagery classifier II), resulting in a mean pseudo-online accuracy of [Formula: see text]%. Accuracies exceeded the threshold for practical BCIs in 12 participants. This study supports the use of visuospatial imagery as a real-time, binary EEG-BCI control paradigm.
脑机接口 (BCI) 可为患有严重运动障碍的个体提供一种沟通方式，例如闭锁综合征患者。许多 BCI 范式依赖于运动神经通路，但这些个体的运动神经通路通常受损。然而，最近的研究结果表明，视觉空间功能可能保持完整。本研究旨在确定以前未探索过的视觉空间意象任务是否可用于在线脑电图 (EEG) 基 BCI 中表示意图。18 名正常发育的参与者在 5 秒的试验中想象棋盘格箭头刺激在视野的四个象限，同时使用视觉皮层上的 16 个干电极收集信号。在在线块中，参与者根据表现获得分级视觉反馈。初始 BCI 管道 (视觉空间意象分类器 I) 在在线试验中以 [公式：见文本]%的准确率区分休息状态和视觉空间意象。通过限制到 alpha 波段特征（视觉空间意象分类器 II）进一步改进了该 BCI 管道，从而得到平均伪在线准确率为 [公式：见文本]%。在 12 名参与者中，准确率超过实用 BCI 的阈值。本研究支持将视觉空间意象用作实时、二进制 EEG-BCI 控制范式。
