Skavhaug Ida-Maria, Bobell Rebecca, Vernon Ben, Joshi Sanjay S
Department of Mechanical & Aerospace Engineering, University of California Davis, I Shields Avenue, Davis, CA 95616, USA.
Annu Int Conf IEEE Eng Med Biol Soc. 2012;2012:1727-31. doi: 10.1109/EMBC.2012.6346282.
We are developing a new class of Brain-Computer Interface that we call a Brain-Muscle-Computer Interface, in which surface electromyography (sEMG) recordings from a single muscle site are used to control the movement of a cursor. Previous work in our laboratory has established that subjects can learn to navigate a cursor to targets by manipulating the sEMG from a head muscle (the Auricularis Superior). Subjects achieved two-dimensional control of the cursor by simultaneously regulating the power in two frequency bands that were chosen to suit the individuals. The purposes of the current pilot study were to investigate (i) subjects' abilities to manipulate power in separate frequency bands in other muscles of the body and (ii) whether subjects can adapt to preselected frequency bands. We report pilot study data suggesting that subjects can learn to perform cursor-to-target tasks on a mobile phone by contracting the Extensor Pollicis Longus (a muscle located on the wrist) using frequency bands that are the same for every individual. After the completion of a short training protocol of less than 30 minutes, three subjects achieved 83%, 60% and 60% accuracies (with mean time-to-targets of 3.4 s, 1.4 s and 2.7 s respectively). All three subjects improved their performance, and two subjects decreased their time-to-targets following training. These results suggest that subjects may be able to use the Extensor Pollicis Longus to control the BMCI and adapt to preselected frequency bands. Further testing will more conclusively investigate these preliminary findings.
我们正在开发一种新型的脑机接口,我们称之为脑-肌肉-计算机接口,其中来自单个肌肉部位的表面肌电图(sEMG)记录被用于控制光标移动。我们实验室之前的工作已经证实,受试者可以通过操纵来自头部肌肉(耳上肌)的sEMG来学习将光标导航至目标。受试者通过同时调节两个适合个体的频段的功率来实现光标的二维控制。当前初步研究的目的是调查:(i)受试者在身体其他肌肉中操纵不同频段功率的能力;(ii)受试者是否能够适应预先选择的频段。我们报告的初步研究数据表明,受试者可以通过使用对每个人都相同的频段收缩拇长伸肌(位于手腕上的一块肌肉),学会在手机上执行光标到目标的任务。在完成不到30分钟的简短训练方案后,三名受试者的准确率分别达到了83%、60%和60%(平均到达目标时间分别为3.4秒、1.4秒和2.7秒)。所有三名受试者的表现都有所提高,并且有两名受试者在训练后缩短了到达目标的时间。这些结果表明,受试者或许能够使用拇长伸肌来控制脑-肌肉-计算机接口并适应预先选择的频段。进一步的测试将更确凿地研究这些初步发现。