Reanimating paralyzed hands using an implantable, brain-controlled functional electrical stimulation neuroprosthesis

项目摘要 这项研究的长期目标是使用完全可植入的脑控装置使瘫痪的手恢复活力。 功能性电刺激神经假体可供脊髓损伤患者随时使用。整体而言 这项提案的目标是提出一项 非人灵长类可植入脑控手神经假体使功能恢复瘫痪 通过电刺激肌肉，而不会牺牲性能。以前的大脑控制 功能性电刺激神经假体需要数百根电线连接到计算机塔上 它们以对可移植性不合理的速率消耗功率以获得所呈现的解码性能，呈现 神经假体的使用仅限于实验室(Bouton等人。2016年，Ajiboye等人。2017年)。中环 假设300-1,000赫兹尖峰频段功率(SBP)功能将允许安全地植入功率级别 同时保持30 kSPS阈值跨越的译码性能。建议的理由是 研究表明，与传统记录模式相比，带宽减少了15倍，并且单个单元的专用性 大幅降低了提取特征所需的功率，而不会对单机性能造成任何影响。在 首先，在嵌入式平台上开发了一种低功耗的多自由度译码方法。 欧文等人。证明SBP可以高性能地预测开环手指位置(Irwin等人)。2016)。 然而，猴子执行了单自由度两个目标的捕获任务。目前还不清楚是否 SBP在解码复杂动作时将保持高性能。因此，SBP将用于 在低功耗嵌入式设备上解码更复杂的中心向外多指任务，如 Bullard，Nason等人。2018年(提交)。假设SBP解码器的性能将好于阈值 在闭环多指任务中交叉解码器，甚至在嵌入式设备上也是如此。第二个的目的是 目的是研究利用嵌入的神经对手部肌肉进行闭环功能电刺激 非人灵长类的信号处理器和网络神经假体。到目前为止，网络 凯斯西储大学开发的神经假体一直无法提供直观的多重 对颈椎水平脊髓损伤患者进行手指控制。假设需要一个大脑接口来 使联网神经假体变得直观，但目前还没有完全可植入的解决方案。该设备来自 第一个目标将用于提供一种可植入的手神经假体，为人类临床试验做准备。这个 这项工作的贡献有望是一种可植入的、直观的、由大脑控制的功能性电信号 刺激手神经假体使脊髓损伤患者恢复部分独立性。这一贡献 将具有重大意义，因为它将提供一种手神经假体，患者可以将其带回家完全- 时间使用。研究人员认为，这项拟议的研究具有创新性，因为这是第一个系统 能够以比标准低一个数量级的功率采集特定于单个单元的信号。

Project Summary The long-term goal of this study is to reanimate paralyzed hands using a fully implantable brain-controlled functional electrical stimulation neuroprosthesis for spinal cord injured patients to use at any time. The overall objective of this proposal, which is the next step toward attainment of the long-term goal, is to present an implantable brain-controlled hand neuroprosthesis in non-human primates that returns function to paralyzed musculature through electrical stimulation and does not sacrifice performance. Previous brain-controlled functional electrical stimulation neuroprostheses required hundreds of wires connected to towers of computers that consume power at rates unreasonable for portability to obtain the presented decode performance, rendering usage of the neuroprostheses restricted to the laboratory (Bouton et al. 2016, Ajiboye et al. 2017). The central hypothesis is that the 300-1,000 Hz spiking band power (SBP) feature will allow safely implantable power levels while maintaining the decode performance of 30 kSps threshold crossings. The rationale of the proposed research is that the 15x bandwidth reduction over conventional recording paradigms and single unit specificity of SBP dramatically cut the power needed to extract features without any loss in single-unit performance. In the first aim, a low-power multiple degree of freedom decoding method will be developed on an embedded platform. Irwin et al. demonstrated that SBP can predict open-loop finger position with high performance (Irwin et al. 2016). However, the monkey performed a single degree of freedom two target acquisition task. It remains unknown if SBP will maintain high performance when decoding complex movements. Consequently, SBP will be used to decode the more complicated center-out multiple finger task on the low-power embedded device presented in Bullard, Nason et al. 2018 (in submission). It is hypothesized that SBP decoders will perform better than threshold crossing decoders in closed-loop multiple finger tasks, even on the embedded device. The purpose of the second aim is to investigate closed-loop functional electrical stimulation of hand muscles using the embedded neural signal processor and the Networked Neuroprosthesis in a non-human primate. To date, the Networked Neuroprosthesis developed at Case Western Reserve University has been unable to provide intuitive multiple finger control to cervical level spinal cord injury patients. It is hypothesized that a brain interface is required to make the Networked Neuroprosthesis intuitive, but there exists no fully implantable solution yet. The device from the first aim will be used to present an implantable hand neuroprosthesis ready for human clinical trials. The contribution of this work is expected to be an implantable, intuitive, brain-controlled functional electrical stimulation hand neuroprosthesis to return some independence to spinal cord injured patients. This contribution will be significant because it will provide a hand neuroprosthesis that patients can take home with them for full- time use. The proposed research is innovative, in the opinion of the researchers, because it is the first system capable of acquiring signals specific to single units using an order of magnitude less power than the standard.