Collaborative Research: NCS-FO: Intelligent Closed-Loop Neural Interface System for Studying Mechanisms of Somatosensory Feedback in Control of Functional and Stable Locomotion

移动腿部的感觉反馈对于功能性和动态稳定的运动至关重要。尽管很明显，与运动相关的感觉反馈会影响腿间协调和步态选择（步行、小跑、奔马等），但尚不清楚哪种感觉方式（例如，肌肉长度或力量相关信号）和反馈来源（例如，臀部或膝盖肌肉）介导这些运动变化。因此，该项目旨在了解提供有关臀部肌肉长度信息的感觉神经元如何调节肢体间协调和步态选择。这一目标将通过以智能、闭环和良好控制的方式选择性地、可逆地刺激这些感觉神经元来实现。该项目将开发新的神经植入工具和相关计算算法，用于在大型动物模型（猫）中体内操纵与运动相关的感觉信号。该项目的新发现和开发的方法将大大增强我们对感觉运动控制机制的理解，并有助于开发新的治疗干预措施。所提出的多学科研究方法还将显着扩展快速发展的光遗传学领域的实用性和能力，实现变革性研究并为神经科学提供前所未有的新实验工具。这项工作对社会最显着的长期好处将是改善大量受到从肢体丧失到感觉神经病等各种运动缺陷影响的人们的生活质量。这些人将受益于由机器学习算法控制的神经系统和工程系统之间的神经接口的发展。在整个项目中，将努力招募和培训来自代表性不足群体的研究生和本科生。还将组织外展活动，与教师、学生和代表性不足的群体分享资源、工具和知识。拟议的研究和教育活动的结果将通过科学博览会、出版物、研讨会、会议和互联网与学生、科学界和公众分享。该提案的总体目标是通过开发和利用体内智能闭环光电神经接口系统，表征猫模型中髋部肌肉纺锤体传入对肢体间协调和步态选择的体感控制机制。特别是，在该提案中，将开发高密度、高效和无线供电的植入式光电（WIOE）神经接口设备。每个 WIOE 异质地将 64 个透明微电极和 16 个微型发光二极管 (μLED) 组成的光电阵列、片上系统 (SoC) 和功率接收器 (Rx) 线圈集成在 mm3 尺寸的封装中，能够进行光遗传学刺激和神经活动的电记录。将实施无线遥测链路，以在外部数据采集/控制单元和 WIOE 植入体分布式阵列之间实现高效的经皮电力和宽带数据传输。多个 WIOE 设备将被植入猫选定的背根神经节 (DRG)。 DRG 神经元的神经活动、四肢选定肌肉的 EMG 活动以及全身运动运动学将被记录，并且在不受约束的猫运动过程中，将通过选定 DRG 中的光遗传学刺激来操纵纺锤体传入活动。利用信号中的时空结构并将 DRG 中的传入活动映射到肢体运动学的机器学习 (ML) 模型将用于实现光遗传神经调节的闭环控制。拟议的研究活动将由在生物MEMS、无线微电子、机器学习、人工智能和行为神经科学领域具有互补研究专业知识的合作者团队进行。所提出的智能闭环光电神经接口的成功开发将为一整套微创神经接口提供强大的构建模块，以研究运动的体感控制，以及监测或治疗体感病理状况。该奖项反映了 NSF 的法定使命，并通过使用基金会的智力价值和更广泛的影响审查标准进行评估，被认为值得支持。

Sensory feedback from moving legs is critical for functional and dynamically stable locomotion. Although it is clear that motion-related sensory feedback influences inter-leg coordination and selection of gaits (walking, trotting, galloping, etc.), it is not known which sensory modalities (e.g., muscle length- or force-related signals) and sources of feedback (e.g., hip or knee muscles) mediate these locomotor changes. Therefore, this project aims to understand how sensory neurons providing information about the length of hip muscles regulate interlimb coordination and gait selection. This goal will be accomplished by selectively and reversibly stimulating these sensory neurons in an intelligent, closed-loop, and well-controlled manner. This project will lead to the development of new neural implant tools and associated computational algorithms for an in-vivo manipulation of motion-related sensory signals in a large animal model, the cat. The new findings of this project and the developed methods will substantially enhance our understanding of the mechanisms of sensory locomotor control and contribute to developing novel therapeutic interventions. The proposed multidisciplinary research approaches will also significantly expand the utility and capabilities of the rapidly growing field of optogenetics, enabling transformative research and providing unprecedented new experimental tools for neuroscience. The most noticeable long-term benefits of this work to society will be an improvement in the quality of life for a sizable population of people affected by a wide range of movement deficits, from limb loss to sensory neuropathy. These individuals will benefit from the development of neural interfaces between the nervous and engineering systems controlled by machine learning algorithms. Throughout this project, efforts will be made to recruit and train graduate and undergraduate students from underrepresented groups. Outreach activities will also be organized to share resources, tools, and knowledge with teachers, students, and underrepresented groups. The results of the proposed research and educational activities will be shared with students, scientific communities, and the public through science fairs, publications, workshops, conferences, and the Internet.The overall goal of this proposal is to characterize the mechanisms of somatosensory control of interlimb coordination and gait selection by spindle afferents of hip muscles in the cat model by developing and utilizing in-vivo an intelligent and closed-loop optoelectronic neural interface system. In particular, in this proposal high-density, efficient, and wirelessly-powered implantable opto-electro (WIOE) neural interface devices will be developed. Each WIOE heterogeneously incorporates an optoelectronic array of 64 transparent microelectrodes and 16 microscale light-emitting-diodes (µLEDs), a system-on-a-chip (SoC), and a power receiver (Rx) coil in an mm3-size package, capable of optogenetic stimulation and electrical recording of neural activities. Wireless telemetry links will be implemented for efficient transcutaneous power and wideband data transmission between an external data-acquisition/control unit and the distributed array of WIOE implants. Multiple WIOE devices will be implanted in selected dorsal root ganglia (DRG) of the cat. Neural activities of DRG neurons, EMG activities of selected muscles of the four limbs, and full-body locomotor kinematics will be recorded, and spindle afferent activities will be manipulated via optogenetic stimulation in selected DRGs during unconstrained cat locomotion. Machine learning (ML) models leveraging the spatiotemporal structures in the signals and mapping afferent activities in DRGs to limb kinematics will be applied for achieving closed-loop control of the optogenetic neuromodulation. The proposed research activities will be conducted by a team of collaborators with complementary research expertise in the areas of bioMEMS, wireless microelectronics, machine learning, artificial intelligence, and behavioral neuroscience. The successful development of the proposed intelligent and closed-loop optoelectronic neural interface will yield a robust building block for a comprehensive set of minimally invasive neural interfaces to study somatosensory control of movement, as well as monitor or treat somatosensory pathological conditions.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.