Restoration of motor function by brain-computer interface with cortical implant

在这项研究中，我们的目标是开发脑机接口（BCI），它允许一个人通过检测他的意图而不使用肌肉或周围神经与他的环境进行交流。在实际肌肉收缩之前，检查由自主运动引起的皮质电图（ECoG）和脑磁图（MEG）来检测运动强度。MEG成功地以精确的时空分辨率显示了肌肉收缩前几毫秒的运动准备场、运动场和运动诱发场。在运动相关皮质中，β带的事件相关去同步（ERD）和高γ带的事件相关同步（ERS）被一致地诱发。ECoG显示更广泛的皮层激活表现为β ERD，反映了运动编程过程的顺序和并行流。光谱分析和相干分析显示ERD/ERS定位与电刺激映射之间的地形一致性。特别是，与α - β波段ERD相比，高频段ERD与运动类型的相关性更强。与来自初级运动区域的ECoG信号相比，来自前运动区域的ECoG信号在邻近电极之间更加一致。这一证据表明，运动控制程序最初是在运动前皮层产生的，具有广泛的功能单元协作，然后变得更加具体，以驱动初级运动皮层中的每个肌肉。应用带通滤波技术对流ECoG信号进行滤波，可以区分运动的意图。通过对条件的优化，可以达到80%的准确度，最高可达92%。这表明我们的方法达到了开发实用、高速bci的目的。在语言任务中观察到类似的ECoG激活，这意味着使用ECoG进行更直接的基于脑机接口的交流是可行的。

In this study we aimed to develop brain-computer interface (BCI) which allows a person to communicate with his environment by detecting his intention without the use of muscles or peripheral nerves. Electrocorticograms (ECoG) as well as magnetoencephalograms (MEG) evoked by voluntary movement were examined to detect the intension of movement before actual muscle contraction.1.MEG successfully revealed with precise spatiotemporal resolution the motor readiness field several millisecond prior to muscle contraction, the motor field and the motor evoked field. Event related desynchronization (ERD) of beta band and event related synchronization (ERS) of high gamma band were consistently evoked in the motor related cortices.2.ECoG revealed wider cortical activations displayed as the beta ERD reflecting the sequential and parallel flows of programming process for motion.3.The spectrographical and coherence analysis revealed topographical concordance between ERD/ERS localization and electrical stimulation mapping. Especially, ERS of high frequency band correlated more selective with the type of motion than ERD of alfa to beta bands. The ECoG signals from premotor areas were more coherent between nearby electrodes compared to those from primary motor areas. This evidence suggested the motor control program are generated initially in the premotor cortices with wide-spread collaboration of functional units and then became more specific to drive each muscle in the primary motor cortices.4.Applying a band-pass filtering to stream ECoG signals, the intention of motion could be differentiated. We could attain the accuracy at the level of 80% and up to 92% by optimizing the conditions. This suggests our method attained to develop practical and high speed BCI.5.Similar ECoG activation was observed in the language task which implied feasibility of more direct BCI based communication using ECoG.