3D printed guide tube system for acute Neuropixels probe recordings in non-human primates

Abstract

Objective.Neuropixels (NP) probes are a significant advance in electrophysiological recording technology that enable monitoring of hundreds of neurons in the brain simultaneously at different depths. Application of this technology has been predominately in rodents, however widespread use in non-human primates (NHPs) such as rhesus macaques has been limited. In this study we sought to overcome two overarching challenges that impede acute NP implantation in NHPs: (1) traditional microdrive systems that mount to cephalic chambers are commonly used to access cortical areas for microelectrode recordings but are not designed to accommodate NP probes, and (2) NHPs have thick dura mater and tissue growth within the cephalic chambers which poses a challenge for insertion of the extremely fragile NP probe.Approach.In this study we present a novel NP guide tube system that can be adapted to commercial microdrive systems and demonstrate an implant method using the NP guide tube system. This system was developed using a combination of CAD design, 3D printing, and small part machining. Software programs, 3D Slicer and SolidWorks were used to target cortical areas, approximate recording depths and locations, and for in-silico implant testing.Main results.We performedin vivotesting to validate our methodology, successfully implanting, explanting, and reimplanting NP probes. We collected stable neurophysiological recordings in the premotor cortex of a rhesus macaque at rest and during performance of a reaching task.Significance.In this study we demonstrate a robust Neuropixels implant system that allows multiple penetrations with the same NP probe and share design files that will facilitate the adoption of this powerful recording technology for NHP studies.

Keywords: 3D printing; CAD; Neuropixels; acute recordings; electrophysiology; rhesus macaques.

Figure 1.

Conventional microelectrode recordings and associated design constraints imposed by utilizing a microdrive approach for Neuropixels probes. (a) Diagram of Neuropixels probe and headstage. (b) Standard microelectrode implantation steps and constraints traditional microelectrode implantation impose on a Neuropixels probe. Microelectrodes use a single lowering phase dedicated to lowering the microelectrode through a pre-placed guide and puncture cannula while a Neuropixels probe requires two lowering phases. The chamber’s 19 mm inner diameter and 25 mm height constrains targetable cortical area since the Neuropixels probe base has a width of 10 mm. Dura mater and periosteum thicken over time after a craniotomy (skull thickness approximated to be 3 mm). Microelectrode implant steps are: 1. Attach the microdrive (modeled after a Narishige microdrive) to the recording chamber, 2. Puncture the dura mater with a guide cannula to provide passage for the microelectrode through the dura mater into the brain, and 3. Descend the microelectrode into the brain to record neuronal activity.

Figure 2.

Two-stage Neuropixels implantation system and cortical targeting overview. (a) CAD rendering showing components of the Alpha Omega microdrive system with custom Neuropixels guide tube system. (b) Gross adjustment of the assembly is performed by turning the gross lowering stage ball screw, attached to the ring platform, and fine adjustment is performed moving the fine stage clamp with the electrode positioning system. The fine adjustment independently controls the movement of the Neuropixels probe. (c) Design modifications adapting the Neuropixels to the two-stage microdrive system. The Neuropixels probe is placed into the probe holder and manually aligned to the 3 mm guide tube, which bridges the dura mater and periosteum growth. The probe holder fits into vertical guide channel, maintaining alignment of probe shank to guide tube during fine stage lowering. (d) Primate anatomy is imaged using 7 T MRI, registered to CT, and segmented in 3D slicer showing the: (i) coronal segmentation scans, (ii) whole 3D segmentation, and (iii) view through recording chamber. (e) (i) Microdrive model fixed to chamber in CAD software, including segmentation generated previously. (ii) Targeting location is determined using a grid rendered over the region of interest. The superior precentral dimple (spcd) and arcuate sulcus (as) are labeled. (iii) Both stages are lowered until the probe is inserted into the target. (iv) Gross and fine stage depth estimates are determined through measurements of the model (f). Approximate Neuropixels probe trajectory and recording locations can be visualized by importing relative model positions from CAD software back into 3D slicer, shown here in oblique sagittal, axial, and oblique coronal 7 T MRI views.

Figure 3.

Implantation sequence. (a) Dura mater is punctured using the Narishige system (step 1). The Flex MT microdrive is placed on the chamber after dura mater puncture and the guide tube is lowered using the gross stage into the puncture location, gently depressing, and stabilizing the surrounding tissue (steps 2 and 3). Data transfer cables and grounding wires are connected, and the NP probe is lowered into the cortex using the fine stage (steps 4 and 5). (b) 2D side view lowering schematic to further illustrate probe insertion. (c) Spiking patterns from a cell shifting across the channels during implantation. At time (t = 1 s) spiking activity is detected on a distal contact and as the probe is advanced (t = 1–5 s) the spiking activity shifts to contacts more proximal to the probe base.

Figure 4.

Left hemisphere Neuropixels implant locations NP 1.0 and NHP 1.0 implantations were located within the dorsal premotor cortex (PMd), defined by the region anterior to the central sulcus (CS) and immediately posteromedial to the arcuate sulcus (AS) and anterolateral to the superior precentral dimple (SPcD). NP 1.0 probes implant locations are demarcated by light grey circles with a black border. NHP 1.0 implantation locations are demarcated by black circles with a light grey boarder. The NP 1.0 probe used for recording area 2 was successfully reused for recording area 3; the same NHP NP 1.0 probe used for recording area 4 was reused for recording area 5.

Figure 5.

Neuropixels recording in primate PMd cortex during a reaching task. (a) Recording data from 384 recording sites closest to the tip of the probe (yellow) with the recording trajectory and depth approximated with model measurements. A representative example raster plot of SU and MU spike times identified from the Neuropixels probe recording channels during a single reach behavior trial. Trial start, food presentation, reach start and return end are labeled and indicated with vertical lines. (b) Selected spike traces show constant firing, modulated firing, increased firing, bursting, and small chewing artifacts during various aspects of the reach across the array. Signal amplitudes are scaled to the max absolute voltage detected on each channel for the displayed time. (c) Three sample spike waveforms (n = 200, mean in red) recorded from channel 246, 248, and 250 at recording start and at 1 h after recording start. (d) Peristimulus time histograms from three example units aligned to reach start and located at different depths along the probe. All data presented are from recording 5 in the table 1.

Figure 6.

Drift maps and traces for recordings. The top row has drift maps only for Recording 1. (a) Drift maps for Recording 1 display the drift initially assessed by Kilosort 2.5, (b) after Kilosort 2.5 initial sorting and drift correction, and (c) after manual curation with Phy GUI. The blue bar shows the corresponding time in the Recording 1 drift trace and drift map. Darker points have a higher spike amplitude and lighter points have smaller spike amplitudes. (d) Recording 1 drift trace shows a higher amount of drift than (e) Recording 5 drift trace. (f) Box plots display median and quartiles for drift during a reaching task for each recording session listed in table 1.