A Virtual Reality Simulator to Study VLSBA and Test Brain-Actuating Technologies

摘要： 在过去的25年里，我的研究重点是研究哺乳动物大脑中允许神经回路产生感觉、运动和认知行为的基本神经生理学原理。为了实现这一目标，我的实验室开发了一系列高度创新的实验方法，结合了电生理学、遗传学、药理学、行为学、计算和工程工具。利用这些，我们的实验室开创了一种革命性的范式，称为脑机接口(BMI)。利用BMI，我们已经证明了非人类灵长类动物和人类受试者可以有效地利用他们的脑源性电活动来直接控制复杂的人工设备的运动，如计算机工具和假肢。然而，BMI研究几乎没有触及脑驱动技术在基础神经科学和临床神经科学的未来可能具有的巨大生物医学潜力。为了开始探索这一未来，我提议开发第一个共享的大脑控制虚拟现实环境(BC-VRE)，旨在研究超大规模大脑活动(VLSBA)的动态特性，以及大脑驱动技术治疗神经疾病的全部潜力。这个虚拟现实模拟器的核心将通过将现代电生理、脑磁图(MEG)和脑成像设备连接到一个超级计算机集群来形成，该集群能够呈现一个虚拟现实环境，在这个环境中，所有组成元素，包括各种计算和机器人工具，甚至虚拟身体，都可以由交互主体的VLSBA直接控制。最初，VLSBA将使用一种新的方法在非人类灵长类动物中进行大通道(高达100,000个单个神经元)记录。稍后，BC-VRE还将接受来自人类受试者的本地或远程脑磁图/磁共振数据。BC-VRE最初将用于：(1)测量人工工具如何被大脑的S表征同化 (2)对S全身神经假体装置的设计进行测试。 严重瘫痪的病人。

Abstract: For the past 25 years, my research has focused on the investigation of the basic neurophysiological principles that allow neural circuits in the mammalian brain to generate sensory, motor, and cognitive behaviors. To pursue this goal, my laboratory has developed a series of highly innovative experimental approaches that combine electrophysiological, genetic, pharmacological, behavioral, computational, and engineering tools. Utilizing those, our laboratory has pioneered a revolutionary paradigm known as brain-machine interfaces (BMIs). Using BMIs, we have demonstrated that non-human primates and human subjects can effectively use their brain-derived electrical activity to directly control the movements of complex artificial devices, such as computer tools and prosthetic limbs. Yet, BMI research has barely touched the enormous biomedical potential that brain-actuating technologies will likely have in the future of both basic and clinical neuroscience. To start probing this future, I propose to develop the first shared brain-controlled virtual reality environment (BC-VRE) designed to investigate the dynamic properties of very-large scale brain activity (VLSBA) and the full potential of brain-actuating technologies for treating neurological disorders. The core of this virtual reality simulator will be formed by interfacing modern electrophysiological, magnetoenecephalographic (MEG), and brain imaging devices to a supercomputer cluster capable of rendering a virtual reality environment in which all constituent elements, including a variety of computational and robotic tools, and even virtual bodies, can be directly controlled by the VLSBA of interacting subjects. Initially, VLSBA will be obtained using a new method for mega-channel (up to 100,000 single neurons) recordings in non-human primates. Later on, the BC-VRE will also accept local or remote MEG/MRI data from human subjects. The BC-VRE will be initially used to: (1) measure how artificial tools are assimilated by the brain¿s representation of the subject¿s body, and (2) test the design of a whole-body neuroprosthetic device for severely paralyzed patients.