Department of Information and Communications Engineering, Tokyo Institute of Technology, Yokohama, Japan.
Institute of Innovative Research, Tokyo Institute of Technology, Yokohama, Japan.
J Healthc Eng. 2020 Feb 8;2020:1418437. doi: 10.1155/2020/1418437. eCollection 2020.
Error-related brain activation has been investigated for advanced brain-machine interfaces (BMI). However, how a delayed response of cursor control in BMI systems should be handled is not clear. Therefore, the purpose of this study was to investigate how participants responded to delayed cursor control. Six subjects participated in the experiment and performed a wrist-bending task. For three distinct delay intervals (an interval where participants could not perceive the delay, an interval where participants could not be sure whether there was a delay or not, and an interval where participants could perceive the delay), we assessed two types of binary classifications ("Yes + No" vs. "I don't know" and "Yes" vs. "No") based on participants' responses and applied delay times (thus, four types of classification, overall). For most participants, the "Yes vs. No" classification had higher accuracy than "Yes + No" vs. "I don't know" classification. For the "Yes + No" vs. "I don't know" classification, most participants displayed higher accuracy based on response classification than delay classification. Our results demonstrate that a class only for "I don't know" largely contributed to these differences. Many independent components (ICs) that exhibited high accuracy in "Yes + No" vs. "I don't know" response classification were associated with activation of areas from the frontal to parietal lobes, while many ICs that showed high accuracy in the "Yes vs. No" classification were associated with activation of an area ranging from the parietal to the occipital lobes and were more broadly localized in cortical regions than was seen for the "Yes + No" vs. "I don't know" classification. Our results suggest that small and large delays in real-time cursor control differ not only in the magnitude of the delay but should be handled as distinct information in different ways and might involve differential processing in the brain.
错误相关脑区激活已在高级脑机接口（BMI）中进行了研究。然而，对于 BMI 系统中延迟的光标控制响应应如何处理尚不清楚。因此，本研究旨在探讨参与者如何响应延迟的光标控制。六名受试者参与了实验并执行了手腕弯曲任务。对于三个不同的延迟间隔（参与者无法感知延迟的间隔、参与者不确定是否存在延迟的间隔和参与者可以感知延迟的间隔），我们根据参与者的反应评估了两种类型的二分类（“是+否”与“不知道”和“是”与“否”），并应用了延迟时间（因此，总体上有四种分类）。对于大多数参与者，“是与否”分类的准确性高于“是+否”与“不知道”分类。对于“是+否”与“不知道”分类，大多数参与者基于反应分类的准确性高于延迟分类。我们的结果表明，一个只包含“不知道”的类别在很大程度上导致了这些差异。在“是+否”与“不知道”的反应分类中表现出高准确性的许多独立成分（IC）与额叶到顶叶区域的激活有关，而在“是与否”分类中表现出高准确性的许多 IC 与顶叶到枕叶区域的激活有关，与“是+否”与“不知道”分类相比，更广泛地定位于皮质区域。我们的结果表明，实时光标控制中的小延迟和大延迟不仅在延迟幅度上有所不同，而且应该以不同的方式作为不同的信息进行处理，并且可能涉及大脑中的不同处理。
