Sources of adaptation during BMI control

PI: Moxon, Karen a .提案编号：1402984机构：Drexel大学标题：BMI控制期间的适应来源目前有超过270,000人患有衰弱性脊髓损伤，每年给国家造成97.3亿美元的医疗保健和生产力损失。大多数人至少腰部以下瘫痪，恢复对腿部的自主控制是他们最期望的结果之一。重要的是，在开发改进的刺激器方面取得了重大进展，这些刺激器可以激活腿部的神经或肌肉并恢复运动。与此同时，科学家们已经展示了记录大脑神经元活动的能力，解码病人移动四肢的意图信息，然后利用这些信息来控制机器人设备或外骨骼。该项目的目标是将这些技术结合起来，开发出最优的解码器，可以将大脑神经元的活动模式“翻译”为可用于控制电刺激和恢复患者自己腿的意志控制的信息。这项工作很重要，因为它代表了一种创新的方法来恢复病人的健康。脊髓损伤后的自主能力。这项研究的一个更广泛的影响是，这种方法可能应用于其他神经系统疾病或疾病。其他更广泛的影响包括，这项工作是德雷塞尔大学一个更大的神经工程项目的一部分，该项目既训练学生，又开发外展项目。该项目将通过提高对脊髓损伤的认识，同时为学生提供参与各级教育（K-12，本科，研究生）的工程设计和开发的机会，为该计划做出贡献。重要的是，这是一个跨学科的项目，涉及工程师、生物学家和神经科学家，他们组成一个团队，努力解决一个复杂的问题，并将特别强调妇女和其他在科学、技术和工程领域未被充分代表的群体的教育。迄今为止，大多数脑机接口（BMI）研究都是在健康的动物模型中进行的。虽然这些数据表明神经元广泛适应“学习”控制外部设备，但问题是损伤后会发生广泛的可塑性和重组，这对BMI的影响尚不清楚。该项目的长期目标是设计有效的解码器，作为闭环BMI系统的一部分，该系统可以控制脊髓损伤后的功能性电刺激，以恢复患者自己肢体的意志控制。这项工作的中心目标是评估解码算法与任务复杂性在脊髓损伤前后神经适应中的相对作用。这一目标将通过使用一种创新的BMI模式来实现，该模式允许大鼠在执行任务时持续与解码器互动：1)它们非常有动力执行任务；2)不需要训练；3)即使在完全脊柱横断后它们也可以执行任务。这项任务要求动物通过神经控制（即BMI）来保持平衡，以应对意想不到的姿势扰动。使用这种新颖的BMI范式，本提案的中心目标将有两个目的。目的1是确定初级运动皮层用于控制平衡的计算机制。通过记录和分析单个神经元群的活动、后肢屈肌和伸肌肌电图以及地面反作用力，将健康大鼠与脊髓横断大鼠在倾斜任务中使用的编码机制进行比较，从而实现这一目标。目的2是比较不同解码器和任务下神经元在BMI控制过程中的适应性。通过在闭环BMI范式中使用倾斜任务的变化，比较健康动物和脊髓横断大鼠记录的神经元的适应性，可以实现这一目标。

PI: Moxon, Karen A. Proposal Number: 1402984 Institution: Drexel University Title: Sources of adaptation during BMI controlThere are currently over 270,000 people who have sustained a debilitating spinal cord injury costing the nation $9.73 billion per year on healthcare and lost productivity. Most of these people are paralyzed from at least the waist down and restoring voluntary control of their legs is among their most desired outcomes. Importantly, significant advances have been made in developing improved stimulators that can activate nerves or muscles of the legs and restore movement. At the same time, scientists have demonstrated the ability to record the activity of neurons in the brain, decode information about the intention of a patient to move their limbs and then use that information to control a robotic device or exoskeleton. The goal of this project is to combine these technologies and develop optimal decoders that can "translate" the patterns of activity of neurons in the brain to information that could be used to control electrical stimulation and restore volitional control of the patients own legs. This work is important because it represents an innovative approach to restoring a patient?s autonomy after a debilitating spinal cord injury. A broader impact of this research is that this approach could likely be applied to other neurological disorders or diseases. Additional broader impacts include the fact that this work is part of a larger Neuroengineering Program at Drexel University that both trains students and develops outreach programs. This project will contribute to the program by raising awareness about spinal cord injury while providing opportunities for students to participate in engineering design and development at all levels of education (K-12, undergraduate, graduate). Importantly, this is an interdisciplinary program involving engineers, biologists and neuroscientists in a team effort to solve a complex problem and special emphasis will be placed on the education of women and other underrepresented groups in science, technology and engineering. To date, most brain-machine interface (BMI) studies have been done in healthy animal models. While these data show extensive adaptation of neurons to "learn" to control an external device, the problem is that extensive plasticity and reorganization occur after injury and the effects of this on BMI are unknown. The long-term goal of this project is to design effective decoders as part of a closed-loop BMI system that could control functional electrical stimulation after spinal cord injury to restore volitional control of a patient's own limbs. The central goal of this proposed work is to assess the relative role of the decoding algorithm compared to that of task complexity on neural adaptation both before and after complete spinal cord injury. This goal will be accomplished by using an innovative BMI paradigm that allows a rat to continuously interact with the decoder while performing a task 1) they are highly motivated to perform, 2) does not require training and 3) they can perform even after a complete spinal transection. The task requires the animal to maintain its balance in response to unexpected perturbations of posture using neural control (i.e. BMI). Using this novel BMI paradigm, the central goal of this proposal will be addressed with two Aims. Aim 1 is to identify computational mechanisms used by primary motor cortex for the control of balance. This Aim will be accomplished by comparing encoding mechanisms utilized by healthy rats to those of rats with spinal transection in the tilting task by recording and analyzing the activity of populations of single neurons, electromyography from hindlimb flexors and extensors and ground reaction forces. Aim 2 is to compare the adaptability of neurons during BMI control using different decoders and tasks. This Aim will be accomplished by comparing the adaptability of neurons recorded from healthy animals to those recorded from rats with spinal transection using variations of the tilting task in a closed-loop BMI paradigm.