Metal microdrive and head cap system for silicon probe recovery in freely moving rodent

High-yield electrophysiological extracellular recording in freely moving rodents provides a unique window into the temporal dynamics of neural circuits. Recording from unrestrained animals is critical to investigate brain activity during natural behaviors. The use and implantation of high-channel-count silicon probes represent the largest cost and experimental complexity associated with such recordings making a recoverable and reusable system desirable. To address this, we have designed and tested a novel 3D printed head-gear system for freely moving mice and rats. The system consists of a recoverable microdrive printed in stainless steel and a plastic head cap system, allowing researchers to reuse the silicon probes with ease, decreasing the effective cost, and the experimental effort and complexity. The cap designs are modular and provide structural protection and electrical shielding to the implanted hardware and electronics. We provide detailed procedural instructions allowing researchers to adapt and flexibly modify the head-gear system.


Introduction
Action potentials are the common currency of communication between neurons and they can be detected as voltage fluctuation in the extracellular space (Adrian and Moruzzi, 1939). However, recording from representative ensembles of neurons simultaneously requires electrodes with multiple recording sites. Multi-wire twisted electrodes (tetrodes) and silicon probes offer the possibility to record tens to hundreds of neurons simultaneously from multiple cortical and subcortical structures simultaneously in freely moving animals (McNaughton et al., 1983;Wise and Najafi, 1991;Csicsvari et al., 2003;Buzsáki, 2004;Blanche et al., 2005;Montgomery et al., 2008;Jun et al., 2017). For cost benefits, microwire arrays are a popular choice for neuroscientists, despite the amount of manual labor involved (Edell et al., 1992). Silicon probes, while more expensive, do not require assembly, the tissue-volume displacement is minimal (Buzsáki, 2004;Kipke et al., 2008), recording properties are consistent (site impedance and locations) and geometric configurations (number of shanks, distance, and pattern of recording sites) can be customized to suit the architecture of the particular brain structure under study (Wise and Najafi, 1991;Scholvin, 2016). The affordability of high-channel-count electrophysiology amplifier chips (e.g., RHD-2132 andRHD-2164, Intan Technologies, Los Angeles, CA;Berényi et al., 2014) and integrated designs (Jun et al., 2017) have accelerated the spread of large-scale recordings. Integration of mLEDs into silicon-based electrodes and can offer unique spatiotemporal control of neuronal activity (Wu et al., 2015;Kim et al., 2020).
The most expensive component of the head gear is the recording probe (from $1000 for 32-channel passive recording probes to > $3000 for mLED probes or larger channel count probes). Therefore, reusing silicon probes in freely moving animals is an important current goal (Juavinett et al., 2019). In addition to reducing costs, repeated usage of the same probe/headgear would allow for better consistency in recordings across animals, enhances data reproducibility, and would reduce electrode/headgear preparation for surgery. Achieving this goal requires an integrated design of a reusable microdrive and head gear to increase recording stability and protect/ shield sensitive drive and electronic components (Chung et al., 2017;Senzai et al., 2019).
Currently available headgear systems are typically large, reducing the ability to record from multiple brain structures in mice. In contrast to recent progress in recording electrodes, the development of implantation techniques such as the Flexdrive, Shuttledrive, DMCdrive and the Hyperdrive (Voigts et al., 2013;Voigts et al., 2020;Kim et al., 2020;Lu et al., 2018) has lagged behind. Electrodes are either fixed in brain tissue or attached to a microdrive to allow the advancement of the electrode after implantation (Chung et al., 2017;Fee and Leonardo, 2001;Korshunov, 2006;Vandecasteele et al., 2012;Wilson and McNaughton, 1993;Yamamoto and Wilson, 2008). Microdrives and accompanying head gear protection and shielding inevitably add extra weight (weight = 0.12 g -1 g, drives designed for mice) and volume (skull surface area = 7.68-252 mm 2 , drives designed for mice) to the implant ( Table 1). The weight, volume and footprint of the microdrive can limit comfortable movement of small rodents and can prevent flexible multiregional recordings in mice (Headley et al., 2015). Yet, chronic recordings from freely behaving subjects are essential in many experiments, where the relationship between neuronal activity and movement, perception, learning and memory, decision making, and other forms of cognition are studied to disambiguate overt behavior and hidden variables (Juavinett et al., 2019;Jun et al., 2017;Steinmetz, 2020). An ideal microdrive should have high precision movement, mechanical stability, minimal size, low weight, and the ability for flexible customization. Commercially available microdrives are expensive and hard to customize. Disposable 3-D printed customized drives and head gear have reduced costs (Headley et al., 2015;Chung et al., 2017;Allen et al., 2020). Most importantly, recovery and reimplantation of recording probes are limited with currently available headgears.
Below, we report the design and testing of an integrated 3D printed headgear system (including microdrives and protective head cap) that is adaptable for multiple recording devices for both mice and rats. Our design reduces surgery time substantially and the small footprint of the metal Microdrive allows targeting multiple brain structures. The fast and reliable recovery of the probe and reuse of the same system in multiple animals decreases costs and experimenter effort.

Results
Recoverable metal microdrive 3D printing has taken science and industries by the storm, offering in-house design customization, fast iterative development, and cheap production using professional printers based on filament extrusion (e.g., MakerBot Industries, New York, NY) and liquid resin (e.g., Form three by Formlabs, Sommerville, MA). Yet, plastic prints have limitations mostly due to the low strength of the materials. Repeated use of plastic threads results in rapid deterioration, which can be prevented by metal-tometal connection (Figure 1-figure supplement 1). Recently, metal printing has become affordable offering increased strength, with options for printing in aluminum, stainless steel and even titanium with similar printing resolution to plastics. Here, we have taken advantage of this and constructed a 3D printed microdrive from stainless steel (stainless steel 316L, 20 mm resolution), which offers superior strength and form factor compared to plastic prints (Young's modulus of stainless steel:~180 GPa vs plastic:~2 GPa). The metal printing allows reuse of the drives with minimal wear, driving the effective cost down. The microdrive is composed of three metal parts: an arm, a body, and a base ( Figure 1A and B) and has a footprint of 15.5 mm 2 (width: 3.1 mm, length: 5 mm, height: 15.3 mm). The detachable base allows for easy recovery of probes. The arm/shuttle is mounted on a screw to the drive body, allowing it to move linearly along the vertical axis simply by turning the screw (270 mm/ turn). The constructed microdrive has a total travel distance of 6 mm (maximum distance between arm and bottom of the drive body), allowing one to record from multiple brain regions across days and weeks. Due to its small form factor, multiple probes can be implanted in the same animal ( Figure 1C). The drives come with a stereotaxic implantation tool, printed in plastic (clear v4 resin from FormLabs), for user-friendly and reliable implantations and probe recovery (Figure 1-figure supplement 2), consisting of a stereotactic manipulator attachment and a microdrive holder ( Figure 1D and E, Figure 1-video 1 and 2).
The fully assembled steel microdrive weighs 0.87 g (base: 0.23 g, shuttle/arm with nut: 0.16 g, drive body with screw and metal bar: 0.49 g). This weight and dimensions are similar to other commercially available or custom-made electrode microdrives ( Table 1). The design files for the microdrive can be submitted to commercial 3D printing companies (e.g., Proto Labs, Maple Plain, MN, https://www.protolabs.com/) allowing for high-quality printing and fast production. The printing costs of the three components can be as low as $170 (base: $50, body: $60, arm: $55), a highly competitive price compared to commercial microdrives.
Inclusion criteria: microdrive systems that used silicon probes in freely moving mice

Mouse cap
To reuse silicon probes in multiple experiments, both the microdrive and the head cap have to be sturdy, easy to disassemble and reassemble. The mouse cap is composed of three parts: a base, a left-side wall, and a right-side wall ( Figure 2A). The cap-base is attached to the skull of the animal during anesthesia using a ring of Metabond cement, serving as a base for the rest of the cap. There is no need for skull support screws, making the head cap minimally invasive. The cap has a large internal window shaped as an elongated octagon, following the outer ridge of the skull, giving wide access for various surgical needs ( Figure 2B). The sidewalls provide structural support, electrical shielding (by acting as a Faraday cage), and physical protection of the silicon probes, hardware, and electronics. The internal volume allows for great flexibility and can fit two Omnetics preamplifierconnectors, as well as optic fibers. The sidewalls attach to the base using a rail and with three support screws (Figure 2-video 1).
The entire cap weighs 2.2 g (base: 0.19 g, walls with male header pins and copper mesh: 0.98 g each, and 000-120 screws: 0.05 g; Figure 2C). A chronically implanted mouse can carry this cap with one (or more) implanted silicon probe and with a custom connector for electrical stimulation ( Figure 2D). High-quality electrophysiological signals can be collected from freely moving mice for weeks ( Figure 2E and F). The system can be customized as needed, using our CAD files (see Methods section). We recommend printing the cap system on the Formlabs Form 2/3 resin printer or a comparable 3D printer (requires 25-50 mm resolution).

Rat cap
The typical Long-Evans rat is approximately ten times heavier than the mouse (~400 g), and requires a sturdier cap system, capable of withstanding forceful impacts and provide increased protection of the electronics and hardware. The rat cap is composed of four parts: a base, a left-side wall, a rightside wall, and a top cover ( Figure 3A). The octagon-shaped base aligns with the outer rim of the rat's dorsal skull surface and is attached with Metabond cement, with no need for skull support  Figure 1. Reusable metal microdrive. (A) The metal microdrive consists of three main parts: a drive body, a movable arm/shuttle, and a removable base. All components are 3D printed in stainless steel. Additional necessary components are a 00-90, 1/2 'brass screw, a 00-90 brass hex nut, a 000-120, 1/8' stainless steel screw fixing the drive to the base, and a male header pin. (B) The assembled drive with dimensions. (C) Schematic showing three microdrives, with silicon probes attached, implanted in a rat to target hippocampus, medial and lateral entorhinal cortices. 3D printed resin head cap is shown in purple. (D) 3D printed stereotaxic attachment and drive holder together with assembly pieces: male header pin, four 00-90 brass hex nuts, three 00-90, 1/4' and a 3/16' stainless steel screw. (E) Stereotaxic attachment with the metal drive assembled, and a probe attached, ready for implantation (red circle highlights the temporary soldering joint for the Omnetics connector). The backend of the silicon probe is attached to the arm using cyanoacrylate glue. The online version of this article includes the following video and figure supplement(s) for figure 1:    . The two side walls are attached to the base with a single rotation-axis located in the front of the base, attached with a long screw ( Figure 3B, top part). The walls are held in place on the base, using a rail and two screws in the back. The sidewalls have two sets of male header pins for soldering standard Omnetics probe connectors (see Surgical Instructions). The lid can be locked with a thumb screw and has holes for air ventilation ( Figure 3B bottom part, Figure 3-video 1). High-quality electrophysiological signals can be collected from freely moving rats for weeks either using silicon probes ( Figure 3C and D) or Neuropixels probes ( The entire design weighs 11.03 g (base: 1.04 g, right wall with male header pins and copper tape: 3.48 g, left wall with male header pins and copper tape: 3.68 g, top with thumb screw: 2.35 g and 00-80 screws: 0.48 g; Figure 3B bottom, right).   Surgical advantages using the head cap systems The modular system decreases the duration of the surgery and allows for faster post-operative recovery of the animal, due to four important modifications. 1. The head cap is prepared before surgery and can be reused easily. 2. The cap does not need support screws, reducing the invasiveness of the surgery and accelerating the animal's recovery. 3. The 3D printed cap-base is secured with a single step, by attaching it to the dorsal surface of the skull with Metabond cement. This ensures alignment precision relative to the brain surface, easier probe recovery, and reusability. 4. The electric shielding and structural support is implemented in the reusable head cap, decreasing extra manual steps for the construction of the protective cap from copper mesh, male header pins and grip cement during surgery (Vandecasteele et al., 2012).
These steps offer a time savings from 40 to 90 min (Figure 3-videos 2 and 3), compared to a manually constructed cap during surgery (Vandecasteele et al., 2012).
Further, the modular cap system substantially increases flexibility during an implantation procedure. Because the sides can easily be disassembled and reassembled, complex surgical procedures can be split into multiple sessions when needed. In the first session the skull is prepared, and the base of the cap is attached to the skull. After recovery, the craniotomy and implantation are performed in a second surgery. This double-step procedure results in a speedy recovery of the animal and reduces the likelihood of human error during extended procedures. Additionally, subsequent troubleshooting can be performed through the course of long chronic experiments with minimal disruption to the animal and the implanted components.

Probe recovery
To recover the probe at the end of a chronic experiment, the drive holder is aligned with the drive using the stereotactic frame. Once the position is aligned in the x-y plane, the drive holder is moved downwards ( Figure 4A, step 1). Next, the top of the drive is secured with the screw located on the side of the drive holder ( Figure 4A, step 2). The 000-120 screw is removed from the base ( Figure 4B, step 1) and the drive is moved upwards carefully ( Figure 4B step two and C). We recommend to carefully monitor the shanks of the probe under a microscope during the entire recovery procedure and, if any unexpected movement of the probe is observed, return to the previous step to make sure that everything is secured properly (Figure 4-videos 1 and 2).
The removed silicon probe (NeuroNexus, Cambridge Neurotech, Diagnostic Biochips products; Neuropixels) is cleaned by initially rinsing it in distilled water, then contact lens solution (containing protease) and distilled water again; each washing step should last for at least 12 hr. Soak the silicon shanks only (avoid soaking the backend area). If extra tissue or debris is detected between the shanks, it can be carefully removed by a fine needle (26 gauge or smaller) under a microscope. After recovering Neuropixels 1.0 probes, the probe shank should be soaked in 1% tergazyme (Alconox) for 24-48 hr, then rinse in distilled water and isopropyl alcohol (Luo et al., 2020).

Quantification of single unit quality measures
Microdrives allow experimenters to record from novel sets of neurons in successive sessions, surveying thousands of neurons from the same structure in a single animal (Girardeau et al., 2017). With the recoverable metal microdrive, we recorded across several days from the same animal while adjusting the implantation depth (500 mm to 780 mm) across days ( Figure 5A-F). Across days of recordings, while the probe was moved to record from different depths, the unit yield increased ( Figure 5H), the waveform amplitude increased ( Figure 5E) while the relative noise level decreased ( Figure 5F), suggesting either that the distance between the electrode sites and neuron bodies decreased or that large size neurons were recorded.

Effect of head gear type on behavior
Finally, we compared the behavioral effect of the 3D printed head cap system with manually built headmounts. To this end, we compared the running speed of rats and mice between subject implanted with the 3D printed and manually built headgears. Rats and mice were water deprived and had to collect water reward on a linear track or figure-eight circular maze. We observed no significant difference between the median running speed of the two rat groups (Kolmogorov-Smirnov test (KS-test); p=0.35) or the 95 percentiles (KS-test; p=0.95). We also performed the same test on mice and found a significant difference between the median running speed of the two groups (KStest; p=0.045G) but no significant difference between the 95 percentile speeds (KS-test; p=0.24; Figure 6H).

Discussion
We have developed a recoverable microdrive printed in stainless steel and a head cap system for chronic electrophysiological recordings in freely behaving rats and mice. The cap system allows for considerably faster and more standardized surgeries to be performed and faster post-surgical recovery of the animals. Importantly, recovery of the probe and head cap becomes an easy and routine procedure, allowing the same silicon probes to be used in multiple animals, offering substantial savings.
Our head caps are minimally invasive and do not require supportive bone screws, shortening surgery time and postoperative recovery. Except for the base, the entire headgear is reusable, making experiments performed on multiple animals less variable. For multiple surgeries (e.g., virus injection for optogenetic or pharmacogenetic experiments), implantation of the base during the first surgery provides fixed coordinates for a subsequent surgery. The head cap system is flexible, due to the large internal volume, and allows for multiple probe implants, optical fiber implants, and other optional components. In contrast, manually constructed cap systems are time-consuming to build, require extensive experience, and its construction may vary from animal to animal and across investigators even in the same laboratory. The main disadvantage of existing headgears is the limited success for probe recovery. Even after successful recovery of the recording probe, a new protective cap  must be built from scratch in subsequent surgeries. In contrast, our modular cap system is prepared before surgery, decreasing the time the animal spends under anesthesia, reducing potential complications during and after surgery. Using this strategy, we were able to explant and implant the same silicon probe in >10 mice (Senzai et al., 2019). The metal microdrive weighs 0.87 g with a footprint area of 15.5 mm 2 , allowing the implantation of multiple probes in rats and even in mice. Because the entire headgear can be removed from the base with a screwdriver, recovery of the silicon probes is simple and highly successful. The drives are printed in stainless steel, with a stiffness prints approximately 100 times higher than that of plastic (Young's modulus of stainless steel:~180 GPa vs. plastic:~2 GPa). Steel drives provide higher stability, potentially better recording quality, and prevent potential wobbling while turning the screw to adjust the probe's position in the brain. Commercially available drives are typically built from plastic, are non-recoverable, and more expensive. Hand-made drives introduce variability across drives and experiments. In contrast, 3D steel printing provides high consistency across drives, reducing interexperimental variability.
Our system allows electrophysiologists to record the neuronal activity from multiple brain regions simultaneously in freely moving rodents. The large internal volume of the head cap and the small footprint of the metal microdrive enable researchers to perform more than one silicon probe implantation in freely moving mice and rats. Despite the highly successful recovery of silicon probes, the probe itself can deteriorate over time, limiting the number of reimplants (decreased signal-to-noise over time, reduced number of high-quality single unit clusters). Appropriate cleaning procedures can extend the lifetime of silicon probes.
To facilitate wide use of the 3D printed designs, we share all necessary details of parts, fabrication process, and vendor source for easy replication by other laboratories (see Materials and methods). We offer several video tutorials, which describe the construction of the microdrive, the cap systems, the probe implantation, and the probe recovery. The CAD system allows different laboratories to customize both the drive and headgear according to their specific goals and needs.

Microdrive assembly instructions
The base of the microdrive anchors the body of the microdrive via a tapped hole in the back (000-120 tap) and four rectangular holes inside the base (0.5 Â 0.5 mm 2 ). Thin walls around the base prevent cement flowing between the base and the body during surgery ( Figure 1A). Glue a nut inside the arm (referred to as 'arm nut'; 00-90 brass nut) before attaching it to the body. The body has an opening in the top part of the back where a nut can fit inside ('top nut'; 00-90 brass nut). Insert the 'top nut' from the back, then insert the arm from the front and introduce a screw (00-90, 1/2', brass screw) through the 'top nut' and the 'arm nut'. Tighten the screw completely and release it a quarter-turn (or less). Fix the 'top nut' and the screw together using solder so the arm can be moved linearly relative to the body by turning this screw. Attach the body-arm complex to the base using a screw in the back (000-120, 1/8', stainless steel screw). Finally, insert a male header pin into the body and secure it using dental acrylic cement (Unifast Trad). This can be used as a soldering joint during surgery. Finally, attach the backend of the silicon probe to the arm using cyanoacrylate glue and solder the Omnetics connector (Omnetics Connector Corporation) of the probe to the male header pin of the drive holder. The fully assembled microdrive weighs 0.87 g (base: 0.225 g, arm with nut: 0.159 g, body with screw and metal bar: 0.486 g).
Assembly_instructions_microdrive_metal_v7.pdf contains instructions with photographic documentation. Figure 1-video 1 shows the assembly of the metal microdrive. Figure 1-video 2 shows the attachment of a Neuropixels probe to metal microdrive.

Implantation/recovery tool assembly instructions
Insert and glue one nut (00-90, brass nut) and a male header pin into the stereotactic attachment using cyanoacrylate glue. Insert a 00-90, 1/4' stainless steel screw into the nut. Tightening this screw will secure this piece to the electrode holder of the stereotactic arm (Model 1770, Kopf Instruments). The male header pin should be used as a temporary soldering joint for the Omnetics connector of the silicon probe. Insert and glue two nuts (00-90, brass nut) into the bottom of the drive holder and one nut (00-90, brass nut) into the body of the drive holder. Insert a 00-90, 3/16' stainless steel screw through this latter nut. This screw should be used to secure the top part of the body of the drive to the drive holder. Attach the stereotaxic attachment to the drive holder using 00-90 screws (00-90, 1/4', T2 screw). Assembly_instructions_implantation_tool_metal_v7.pdf contains instructions with pictures.

Mouse cap assembly instructions
The base has a rectangular hole for a male header pin (0.8 Â 0.8 mm 2 ) for fixing the left and right walls temporarily during surgery ( Figure 2B). This can help to open the cap using a fine pair of tweezers. The tip of the tweezer is squeezed between the rectangle and the walls. Pushing the tweezer against this rectangle readily opens the walls. The right wall has one tapped hole in the front and one in the lower part of the back (000-120 thread, 1.9 mm length). In addition, it has a hole in the upper part of the back (1 mm in diameter, 1.4 mm length). The left wall has one hole in the front and one in the lower side of the back (1 mm in diameter, 1.4 mm length) and a tapped hole in the upper side of the back (000-120 thread, 1.9 mm length). In addition, there are two rectangular holes in each wall (0.8 Â 0.8 mm 2 ) in which male header pins are glued with cyanoacrylate glue to serve as soldering points for the Omnetics connector and for the shielding copper mesh. To reduce weight, walls are perforated and covered with light copper mesh by gluing it with dental acrylic (Unifast Trad). The walls are closed using two screws in the back and one screw in the front (000-120, 1/8' stainless steel pan head torx screws).
Assembly_instructions_mouse_hat_10_39 mm_v11.pdf file contains instructions with pictures. Figure 2-video 1 shows the assembly of the mouse cap.

Rat cap assembly instructions
The base has a hole for a brass screw-to-expand insert (00-80 thread size, 1/8' installed length) and serves to hold together the left and right walls. It has a rectangular protrusion in the back (3 Â 1.5 Â 1.67 mm 3 ) to help opening of the cap using a fine tweezer. The right and left walls have a front hole (diameter 1.8 mm) in which a screw can be passed (00-80, 1' 18-8 stainless steel socket head screw) for fixing the walls to the metal insert of the base. In addition, there is a rail on each wall at the bottom part that grabs onto the base piece (1.2 mm height and 1 mm deep). During surgery, the walls are kept open with the screw loosely tightened ( Figure 3B, top part). After all the connectors are attached to the male header pins, the walls are closed, and the front screw is tightened. The right wall has a hole in the upper side of the back (1.8 mm, 2 mm length) and a tapped hole in the lower side of the back (00-80 thread, 2 mm length). The left wall has a hole in the lower side of the back (diameter 1.8 mm, 2 mm length) and a tapped hole in the upper side of the back (00-80 thread, 2 mm length). The walls are closed in the back using two screws (18-8 stainless steel socket head screw, diameter 0-80, 5/32" length). The left wall also has an insert in the upper part of the back side for a nut (18-8 stainless steel thin hex nut, M2.5 Â 0.45 mm thread). This latter nut serves as a locking mechanism for the top cover. There are four rectangular through-holes in each wall (0.8 Â 0.8 mm) in which male header pins are glued with epoxy (Araldite Standard Epoxy) and serve as soldering points. The locations of the holes can be modified according to user specifications to adapt different connector placements. To protect the implanted electrodes, the rat cap is covered by either self-adherent wrap (3M Coban) or the plastic top cover. The edge is extruded on the outer surface on top of the walls to provide extra surface for better adhesion. The plastic cover is attached to the walls using the front slide-in slot and the back screw (stainless steel flared-collar knurled-head thumb screw, M2 Â 0.40 mm thread size, 4 mm long). To protect the neuronal signal from environmental electromagnetic interference noise, conductive copper coil electrical tape is glued to the walls by cyanoacrylate glue (copper tape: 1" wide, McMaster product number: 76555A724) and connected to the ground.
Assembly_instructions_rat_cap_v8.pdf file contains instructions with photographs. Figure 3-video 1 shows the assembly of the rat cap.

3D designing and printing parts
All parts were designed in Autodesk Fusion 360 (https://www.autodesk.com/products/fusion-360). We tested and printed cap designs on a Form two printer from Formlabs with 50 mm resolution using their standard resins. The metal microdrive prints were produced by Proto Labs (https://www. protolabs.com/services/3d-printing/direct-metal-laser-sintering). All designs are available from our GitHub repository https://github.com/buzsakilab/3d_print_designs (copy archived at swh:1:rev: a073716d89c32f13eb76a5ac5e7fa6f7fa11e18a; Vö rö slakos, 2021). Proto Labs (Proto Labs, Inc, MN, USA) currently charges 250$ for a single print but substantial savings are available for larger print orders. Coordination of orders across laboratories therefore can reduce the price.

Subjects
Rats (adult male Long-Evans, 250-450 g, 3-6 months old, n = 11) and mice (adult male C57BL/ 6JxFVB mice, 32-40 g, n = 6) were kept in a vivarium on a 12 hr light/dark cycle and were housed two per cage before surgery and individually after it. All experiments were approved by the Institutional Animal Care and Use Committee at New York University Medical Center. Animals were handled daily and accommodated to the experimenter before the surgery and behavioral recording.

Surgery instructions
The following instructions cover surgeries in both rats and mice, with differences highlighted. Prior to surgery, prepare the 3D printed cap, the microdrive(s), the implantation tool and attach a silicon probe to the microdrive (as described above).
We recommend measuring the impedance of the silicon probe before implantation using the RHD USB interface board from Intan (Intan Technologies LLC, CA, USA). Lower the probe into 0.9% saline and ground the saline to the recording preamplifier ground. Connect the probe to an Intan preamplifier headstage (RHD 32-or 64-channel recording headstages) and to the main Intan board to perform the impedance test measurement at 1 kHz frequency. This measurement provides a quick and rough estimate about the quality of the recording sites. If higher precision is required, users can perform impedance measurement with a nanoZ device following the manufacturers recommendation (nanoZ Impedance Tester, Plexon Inc, TX, USA).
2. Place the animal back in a home cage. The cage should be placed on a heating pad during the first night. 3. Inject Buprenex subcutaneously after 20 min (0.05-0.1 mg/kg).

General notes
. Apply mineral oil to the eyes of the animal at regular intervals.
. To keep the animal properly hydrated during the postoperative days, provide an aqua-gel and a small container with water. Provide regular rodent pills.

Additional implantation information
Rats and mice were implanted with silicon probes to record local field potential and spikes from the CA1 pyramidal layer in rats and from the prelimbic cortex from mice. Silicon probes (NeuroNexus, Ann-Arbor, MI and Cambridge Neurotech, Cambridge, UK) were implanted in the dorsal hippocampus (rats: antero-posterior (AP) À3.5 mm from Bregma and 2.5 mm from the midline along the medial-lateral axis (ML); mice: antero-posterior (AP) +1.75 mm from Bregma and 0.75 mm from the midline, 10 degree relative to the sagittal axis). The probes were mounted on the recoverable metal microdrive and previous design iterations made in plastic (unpublished work; stl files and instructions are available at our GitHub repository https://github.com/buzsakilab/3d_print_designs/tree/master/ Microdrives/Plastic_recoverable), allowing for precise vertical movement after implantation and implanted by attaching the base of the micro-drives to the skull with dental cement (Supplementary file 1). The small footprint of the metal microdrive enables researchers to perform more than one silicon probe implantation in freely moving mice. For this purpose, larger mice (>35 g) were selected. After the post-surgical recovery, we moved the probes gradually in 50 mm to 150 mm steps until the tips reached the pyramidal layer of the CA1 region of the hippocampus. The pyramidal layer of the CA1 region was identified by physiological markers: increased unit activity and the presence of ripple oscillations (Mizuseki et al., 2011). In mice, the probe was implanted 500 mm below the surface of the brain and recordings were performed each day. The probe was moved 70 mm after each recording day. Data was collected daily. The implanted animals were single housed, and they do not carry the headstage while in the vivarium. During recordings, the headstage is attached and a counterbalanced pulley system ensures that the animal is not carrying the extra weight of the headstage and cable.

Behavioral data
Rats were trained to run on a 2.3 m linear track, or a 120 cm diameter circular track with a diagonal path allowing the animals to run in a figure-eight pattern (n = 3 rats implanted with the 3D printed head gear, n = 13 sessions, and n = 4 control subjects with manually built headgear, n = 22 sessions). In both behavioral paradigms, rats were water deprived and had to collect water reward (~0.02 ml).
Mice were either trained to run on a 1.1 m linear track (n = 3 mice implanted with the 3D printed head gear, n = 9 sessions), or to run on a 80 cm diameter circular maze with a diagonal path allowing the animals to run in a figure-eight pattern (same layout as the rats but a smaller maze, n = 5 control subjects with manually built headgear, n = 54 sessions). In all behavioral paradigms, mice were water deprived and had to collect water reward (~0.02 ml).
The linear tracks had 'reward areas' on each end where water reward was delivered via a custombuilt infrared-beam triggered system. Animals only received water reward for trials in which they travelled from one reward site to the other. On the circular maze the animals ran along the central arm after which they ran along the outer circle in a alternation fashion. Water reward was delivered in the reward area on correct trials.
The behavior of the animals was recorded using the Optitrack camera system (NaturalPoint, Inc, OR, USA).