Ultra High-density Optomechanic Neural Interfaces

项目摘要 对神经回路的多尺度、机械性理解，包括局部和全脑 互联互通仍然难以捉摸。根本的挑战之一是缺乏监测工具， 高时空分辨率，局部神经元的活动同时在大脑的不同区域 清醒、行为自由的动物的大脑。这就要求设计超高密度的神经探头 从数千个神经元以高时空分辨率进行记录。虽然已经有了巨大的 传统无源和有源电子神经探头的设计进展，这些技术包括 达到扩展限制。我们需要打破传统的录波和继电方案 使用被动或主动电子神经探头的神经信号，以实现突破性改进 我们可以从大脑记录的同步频道的数量。在这里，我们提出了一种颠覆性的方法 基于光学和微电子机械系统(MEMS)的根本进步，以提供 创新的光机械探测器，可能同时有数千个以上的探测器 有源记录电极与传统无源探头的占地面积相同。所有记录的神经细胞 我们的设计中的信号在光学域中进行编码，以利用超高带宽的光 在单个光波导上将记录的聚合神经信号传输到大脑外部。在这 方案中，每个记录通道被编码到沿其传播的单个波长的光上 波导型。这种波长域多路复用(WDM)方法能够真正同时记录多个 与在有源电子神经探头中使用的时间域多路复用(TDM)方案不同， 其依赖于多个频道的顺序记录。因此，我们的设计支持大规模扩展 同时记录的频道数，同时提高SNR、保留带宽并最小化 有源电子神经探头的不利影响，如在大脑内产生热量。的核心单元格 我们的神经探头是一种机电传感器，可以检测神经电信号并将其转换为小信号 薄膜的机械运动，而薄膜又调制光子微谐振器。因此，电气设备 神经信号被转换成机械信号，然后是光学信号。超高品质因数光学元件 微谐振器增强了检测到的信号。一种耦合到多个微谐振器的单个公共波导 将光信号传输到大脑外部的后端。这种新颖的设计使大规模扩展成为可能 在不增加神经探头尺寸的情况下记录通道的数量。此外，转换为 电信号到光信号的转换导致了信噪比(SNR)的提高，还使得 传输的信号不受不必要的电子干扰。在成功演示多路传输后 电光机械神经记录在本项目中，成果可以在今后的工作中推广，以i)发展 具有1000多个通道的更高密度的神经探针和ii)展示了其在体内的应用。

Project Summary A multi-scale, mechanistic understanding of neural circuits that includes both local- and whole-brain interconnections still remains elusive. One of the fundamental challenges is the lack of tools for monitoring, with high spatiotemporal resolution, the activity of local neuron ensembles simultaneously in different regions of the brain in awake, freely-behaving animals. This calls for the design of ultrahigh density neural probes capable of recording from thousands of neurons with high spatiotemporal resolution. While there has been tremendous progress on the design of conventional passive and active electronic neural probes, these technologies are reaching scaling limits. We need to break away from the conventional scheme of recording and relaying electrical neural signals using passive or active electronic neural probes to enable breakthrough improvements in the number of simultaneous channels that we can record from the brain. Here, we propose a disruptive approach based on fundamental advancements in optics and microelectromechanical systems (MEMS) to deliver an innovative opto-mechanical probe that can potentially have more than a couple of thousand simultaneously active recording electrodes in the same footprint of a conventional passive probe. All of the recorded neural signals in our design are encoded in the optics domain to leverage the ultrahigh bandwidth of light for communicating the recorded aggregate neural signals to outside the brain on a single optical waveguide. In this scheme, each recording channel is encoded onto a single wavelength of light that travels along the same waveguide. This wavelength domain multiplexing (WDM) method enables a true simultaneous recording of many channels, unlike the time domain multiplexing (TDM) scheme that is used in active electronic neural probes, which relies on sequential recording of multiple channels. Therefore, our design enables massive scaling of the number of simultaneously recorded channels, while enhancing SNR, preserving the bandwidth, and minimizing adverse effects of active electronic neural probes such as heat generation inside the brain. The core unit cell of our neural probe is an electromechanical sensor that detects electrical neural signals and converts them to small mechanical motions of a membrane, which in turn modulates a photonic microresonator. Therefore, the electrical neural signal is transformed to a mechanical and then an optical signal. The ultra-high quality factor optical microresonator enhances the detected signals. A single common waveguide coupled to multiple microresonators carries the optical signals to the backend outside the brain. This novel design enables massive scaling of the number of recording channels without increasing the size of the neural probe. Moreover, the conversion of electrical signals to optical signals results in enhanced signal-to-noise ratio (SNR) and also makes the transmitted signals immune to unwanted electrical interference. After successful demonstration of multiplexed electro-opto-mechanic neural recording in this project, the results can be extended in future efforts to i) develop even much higher density neural probes with more than 1000 channels and ii) demonstrate its in vivo application.