Massively Multiplexed Gold Microprobe Arrays for Whole-Mouse-Brain Recording

摘要 了解大脑功能需要同时记录数千或数万个 数千个神经元参与了神经回路中的动态活动。基于CMOS基电极 技术构成了与生命系统进行电交互的唯一手段，超越了这种规模，并在 毫秒级的时间分辨率，但受限于记录深度和硅片的侵入性 这可能会禁止他们在人体实验中使用。尽管如此，人们越来越意识到，利用 商业上可用的大规模cmos技术可能会解决脑成像中的规模挑战。 我们在此提出了一种新的设备概念，它包含了两种独特和创新的组合 广泛使用现有技术--金属微线和cmos电子技术--以创造协同效应。通过 在薄而可靠的CMOS片上开发小直径、深穿透的金微探针 电子路由器，我们的目标是在一个只有几根I/O线的情况下实现1000多个亚细胞神经探针 商业上可行的方式，可扩展到数万，甚至更高的大规模活体大脑 映射。 我们将首先开发一种可行的电化学沉积工艺来获得高深宽比的金 直径最小为10微米，长度最大为1毫米的微型探头。一种96 mm的无源金微探针阵列 将首先开发密度高达100个探头/mm2的通道。同时，我们将大规模开发一种 SOI衬底上的多路CMOSASIC设计将被减薄到小于... 支持Kapton基板上的3微米器件厚度。我们将形成大规模的多元化穿透 通过在精简的ASIC上合成金微探针，可实现多达1000个电极的阵列。植入物 在啮齿动物皮质将被用来评估记录的可靠性和组织反应。整体式制造 黄金微探针的工艺将支持在厘米处放大到10,000-100,000个或更高的微探针 比例。 该项目充分利用了材料科学家、电路设计师、设备 达特茅斯学院和犹他大学(犹他州)的工程师和电生理学家，以实现大- 规模化、亚细胞、深度穿透的金微探针阵列，作为从全鼠到全鼠的尺度基础- 大脑记录。

Abstract Understanding brain function requires the ability to record simultaneously from thousands or tens-of- thousands of neurons contributing to the dynamic activity in a neural circuit. CMOS based electrode technology constitutes the only means to electrically interact with living systems beyond this scale and at sub- millisecond time resolution, but suffers from the limited recording depth and the invasiveness of silicon wafer which might prohibit their use in human experimentation. Still, there is a growing awareness that leveraging commercially available large-scale CMOS technologies might address the scaling challenge in brain mapping. We herein propose a new device concept which encompasses a unique and innovative combination of two widely-used existing technologies - metal microwires and CMOS electronics - to create a synergistic result. By developing small-diameter, deep-penetrating, gold microprobes monolithically on a thin yet reliable CMOS electronic ‘router’, we aim to achieve over one thousand subcellular neuro probes with only a few I/O wires in a commercially viable way, with scalability up to tens of thousands and even higher for large-scale in vivo brain mapping. We will first develop a viable electrochemical deposition process to achieve high-aspect-ratio gold microprobes with down to 10µm diameter, and up to 1mm length. A passive gold microprobe array with 96 channels will be developed first at a density up to 100 probes/mm2. In parallel, we will develop a massively multiplexed CMOS ASIC design on SOI substrates to be thinned down to the buried oxide layer with less-than- 3µm device thickness on a supporting Kapton substrate. We will form the massively multiplexed penetrating arrays with up to 1000 electrodes by synthesizing gold microprobes on the thinned-down ASIC. Implantations in rodent cortex will be used to assess recording reliability and tissue response. The monolithic fabrication process of the gold microprobes will support up-scaling to 10,000 - 100,000 microprobes or higher at the cm scale. This project leverages a vibrant collaboration between material scientists, circuit designers, device engineers and electrophysiologists at Dartmouth College and the University of Utah (Utah), to realize large- scale, subcellular, deep-penetrating gold microprobe arrays that serve as a basis to scale to whole-mouse- brain recording.