Neuropixels 2.0: A miniaturized high-density probe for stable, long-term brain recordings
Neuropixels 2.0: A miniaturized high-density probe for stable, long-term brain recordings
复制标题
Neuropixels 2.0：用于稳定、长期脑部记录的小型化高密度探针
DOI：
10.1101/2020.10.27.358291
复制
发表时间：
2020
期刊：
影响因子：
--
通讯作者：
Steinmetz N
中科院分区：
文献类型：
--
作者：
Steinmetz N
引言基于互补金属氧化物半导体硅制造技术的电极阵列，如Neuropixels探针，已经能够记录活体大脑中数千个单独的神经元。这些工具导致了关于感知和行动的大脑相关性的发现，主要是在急性，头部固定记录中使用。然而，为了研究神经元在不同时间尺度上的处理动力学，有必要在数周和数月内记录神经元，最好是在无约束的行为和小动物（如小鼠）中记录。探针和头台被小型化到原始尺寸的约三分之一（即，Neuropixels 1.0探头的尺寸），因此两个探头及其单个头台的重量仅约为1.1 g，而不会损失通道数（每个探头384个通道）。使用两个四柄探针可在一个种植体中提供10，240个记录位点。为了在大脑运动的情况下实现稳定的记录，我们优化了记录部位的布置。探头具有更密集的线性几何结构，允许使用新设计的算法进行事后计算运动校正。该算法在Kilosort 2.5软件包中实现，从尖峰数据中确定随时间的运动，并使用空间恢复进行校正，如在图像配准中。为了验证这些探针的长期记录，我们将它们长期植入六个实验室的21只大鼠和小鼠中。这21个植入物中有20个成功，并在数周和数月内产生了神经元，同时保持了良好的信号质量。使用新设计的植入物固定装置设计，探针可以可靠地恢复。为了测试运动校正算法的性能，我们使用已知的探针相对于大脑的强加运动进行记录。该算法提高了稳定神经元的产量，并在很大程度上消除了运动对记录的影响，该算法的一个版本允许跨天稳定记录神经元。我们通过对初级视觉皮层中长期记录的单个神经元进行“指纹识别”来评估这一点，这些神经元对一组图像的独特视觉反应。神经元跟踪是>90%成功的2周和>80%成功的2 months.CONCLUSIONThis工作演示了一套电生理工具，包括一个小型化的高密度探头，可恢复的慢性植入夹具，和算法的自动事后运动校正。这些工具可以使小动物（如小鼠）的记录位点数量增加一个数量级，并能够在长时间范围内稳定记录。Neuropixels 2.0探针可以进行前所未有的记录（A）Neuropixels 1.0和2.0设备设计的比较。Neuropixels 2.0器械是小型化的，有四个柄。每个头台可容纳两个探头。(B)在>300天内记录的跨皮层（Ctx）、海马（HC）和丘脑（Th）的尖峰活动模式。(C)来自长期植入小鼠体内的两个Neuropixels 2.0探针的尖峰光栅示例，显示在两个探针的10，240个可用位点中的6144个上记录的尖峰。从768个通道进行8次连续记录（不同颜色）。
INTRODUCTIONElectrode arrays based on complementary metal-oxide semiconductor silicon fabrication technology, such as Neuropixels probes, have enabled recordings of thousands of individual neurons in the living brain. These tools have led to discoveries about the brain-wide correlates of perception and action, primarily when used in acute, head-fixed recordings. To study the dynamics of neuronal processing across time scales, however, it is necessary to record from neurons over weeks and months, ideally during unrestrained behavior and in small animals, such as mice.RATIONALETo this end, we designed a miniaturized probe, called Neuropixels 2.0, with 5120 recording sites distributed over four shanks. The probe and headstage were miniaturized to about one-third of the original size (i.e., the size of the Neuropixels 1.0 probe), so that two probes and their single headstage weigh only ~1.1 g, without loss of channel count (384 channels per probe). Using two four-shank probes provides 10,240 recording sites in one implant. To achieve stable recordings despite brain movement, we optimized the recording site arrangement. The probe has a denser, linearized geometry that allows for post hoc computational motion correction using a newly designed algorithm. This algorithm, implemented in the Kilosort 2.5 software package, determines the motion over time from the spiking data and corrects it with spatial resampling, as in image registration.RESULTSTo validate these probes for long-term recordings, we implanted them chronically in 21 rats and mice in six laboratories. Twenty of these 21 implants succeeded and yielded neurons over weeks and months while retaining good signal quality. The probes were reliably recoverable using newly engineered implant fixture designs.To test the performance of the motion correction algorithm, we performed recordings with known imposed motion of the probe relative to the brain. The algorithm improved the yield of stable neurons and largely eliminated the impact of motion on the recording.A version of this algorithm allowed the recording of neurons stably across days. We assessed this by “fingerprinting” individual chronically recorded neurons in the primary visual cortex using their distinctive visual responses to a battery of images. Neuron tracking was >90% successful for up to 2 weeks and >80% successful for up to 2 months.CONCLUSIONThis work demonstrates a suite of electrophysiological tools comprising a miniaturized high-density probe, recoverable chronic implant fixtures, and algorithms for automatic post hoc motion correction. These tools enable an order-of-magnitude increase in the number of sites that can be recorded in small animals, such as mice, and the ability to record from them stably over long time scales.Neuropixels 2.0 probes allow unprecedented recordings(A) Comparison of the Neuropixels 1.0 and 2.0 device designs. The Neuropixels 2.0 device is miniaturized and has four shanks. Two probes can be hosted per headstage. (B) Pattern of spiking activity across the cortex (Ctx), hippocampus (HC), and thalamus (Th) recorded over >300 days. (C) Example spiking rasters from two Neuropixels 2.0 probes chronically implanted in a mouse, showing spikes recorded on 6144 of the 10,240 sites available across the two probes. Eight sequential recordings (different colors) were performed from 768 channels each.
登录
查看更多内容
影响因子：
25
作者：
通讯作者：
--
跨多个大脑的线性探针电极的精确定位
DOI：
--
发表时间：
2020
期刊：
bioRxiv
影响因子：
--
作者：
Liu D Liu;Susu Chen;M. N. Economo;Nuo Li;K. Svoboda
通讯作者：
K. Svoboda
影响因子：
14.8
作者：
Rik J J van Daal-Rik-J-J-van-Daal-2007891459;C. Aydın;Frédéric Michon;A. Aarts;M. Kraft;F. Kloosterman;S. Haesler
通讯作者：
Rik J J van Daal-Rik-J-J-van-Daal-2007891459;C. Aydın;Frédéric Michon;A. Aarts;M. Kraft;F. Kloosterman;S. Haesler
猫纹状皮层相邻神经元对自然刺激反应的异质性。
影响因子：
2.5
作者：
S. Yen;J. Baker;C. Gray
通讯作者：
C. Gray
影响因子：
2.5
作者：
Jackson, Andrew;Fetz, Eberhard E.
通讯作者：
Fetz, Eberhard E.