Collaborative Research: CDI Type I: Optimal and predictive control of neural prostheses using intracortical Brain Machine Interfaces

本研究的目的是为脑机接口（bmi）实现的神经假体开发一个最佳的预测反馈控制框架。广义上来说，bmi是连接大脑和机器（计算机）的系统，用于感知神经元活动和认知过程，以恢复残疾受试者受损的运动任务。本研究的重点是：(1)开发基于bmi的神经假体的计算建模框架，该框架可以纳入皮质内神经元测量的大数据集；(2)神经运动假体自然和“替代”最佳反馈感知路径的识别；(3)开发计算效率高的预测控制算法，利用神经假体中的模型和反馈通路；(4)模拟神经义肢多指抓握实验评价。目前的神经假体装置大多是“开环”的，由于无法整合多变量反馈、感觉和交互信号，因此性能有限。这项研究解决了这一限制，努力推进下一代反馈神经假体设备。将最优控制理论与神经工程和计算神经科学的原理相结合，为分析复杂的神经元互联系统中固有的众多反馈模式提供了一个框架。为了产生更广泛的影响，该研究寻求使基于bmi的神经假体和辅助装置能够在患有周围神经病变、脊髓损伤、神经肌肉疾病和截肢的人类受试者中稳定地扩展使用。该项目将在反馈控制教育中引入计算神经科学范式，同时培训生物医学工程师直接使用计算反馈控制理论的工具。伯利恒的阿萨帕克小学和巴尔的摩的中学暑期机器人夏令营将通过大脑功能的演示来激励和激发年轻人思考解决技术“重大挑战”。

The objective of this research is to develop an optimal predictive feedback control framework for neuroprostheses that are enabled by brain machine interfaces (BMIs). BMIs, broadly defined, are systems that interface the brain and a machine (computer) to sense neuronal activity and cognitive processes to restore impaired motor tasks in disabled subjects. The research focuses on: (1) development of a computational modeling framework for BMI-based neuroprostheses that can incorporate large data sets of intracortical neuronal measurements; (2) identification of natural and "surrogate" optimal feedback sensing paths in neuromotor prostheses; (3) development of computationally efficient predictive control algorithms that exploit models and feedback pathways in neuroprostheses; and (4) experimental evaluation through simulated neuroprosthetic multi-finger grasping.Current neuroprosthetic devices are largely "open-loop" and have limited performance due to their inability to incorporate multivariable feedback, sensory and interactive signals. This research addresses this limitation, striving to advance next-generation feedback-enabled neuroprosthetic devices. The proposed conceptual integration of optimal control theory with principles of neuroengineering and computational neuroscience provides a framework for analyzing the numerous feedback modalities intrinsically embedded in a complex interconnected system of neurons.For broader impacts, the research seeks to enable the transition of BMI-based neuroprostheses and assistive devices to stable extended use in human subjects suffering from peripheral neuropathies, spinal cord injuries, neuromuscular disorders, and amputations. The project will introduce a computational neuroscience paradigm in feedback control education while training biomedical engineers to directly employ tools from computational feedback control theory. A science outreach effort with Asa Packer Elementary School in Bethlehem and the middle school Summer Robotics Camp in Baltimore will motivate and excite young minds to think about solving technological "grand challenges" through demonstrations of brain functionality.