Chronic co-implantation of ultraflexible neural electrodes and a cranial window

Abstract. Significance: Electrophysiological recording and optical imaging are two prevalent neurotechnologies with complementary strengths, the combined application of which can significantly improve our capacity in deciphering neural circuits. Flexible electrode arrays can support longitudinal optical imaging in the same brain region, but their mechanical flexibility makes surgical preparation challenging. Here, we provide a step-by-step protocol by which an ultraflexible nanoelectronic thread is co-implanted with a cranial window in a single surgery to enable chronic, dual-modal measurements. Aim: The method uses 1-μm-thick polymer neural electrodes which conform to the site of implantation. The mechanical flexibility of the probe allows bending without breaking and enables long-lasting electrophysiological recordings of single-unit activities and concurrent, high-resolution optical imaging through the cranial window. Approach: The protocol describes methods and procedures to co-implant an ultraflexible electrode array and a glass cranial window in the mouse neocortex. The implantation strategy includes temporary attachment of flexible electrodes to a retractable tungsten-microwire insertion shuttle, craniotomy, stereotaxic insertion of the electrode array, skull fixation of the cranial window and electrode, and installation of a head plate. Results: The resultant implant allows simultaneous interrogation of brain activity both electrophysiologically and optically for several months. Importantly, a variety of optical imaging modalities, including wide-field fluorescent imaging, two-photon microscopy, and functional optical imaging, can be readily applied to the specific brain region where ultraflexible electrodes record from. Conclusions: The protocol describes a method for co-implantation of ultraflexible neural electrodes and a cranial window for chronic, multimodal measurements of brain activity in mice. Device preparation and surgical implantation are described in detail to guide the adaptation of these methods for other flexible neural implants and cranial windows.

1 Introduction adaptation, learning, development, and degeneration occur on broader timescales ranging from seconds to years and even decades. [2][3][4] Spatially, neural activity involves not only cellular and subcellular structural and functional changes, but also orchestrated activities distributed across multiple brain areas. In addition to neuronal activity, brain functions and dysfunctions also stem from the interaction between other cell types and surrounding vasculature. 5,6 Resolving these multifaceted activities requires technologies with cross-brain coverage, long-lasting functioning periods, and high spatial and temporal resolutions that match the network in question. 7 Electrophysiology, one of the "gold standard" tools in detecting neuronal dynamics, affords high temporal resolution detection and direct neural activation, but is limited by its low spatial resolution incapable of resolving fine cellular and subcellular structures. In contrast, optical imaging and modulation techniques offer high spatial resolution 8,9 and cell type specificity, 10 and can measure non-neuronal, non-electrical activities. 11,12 Nevertheless, they are limited by their generally low temporal resolution and insufficient tissue penetration depth. Integrating the electrical and optical measurements in the same brain leverages their complementary strengths, which has become an emerging approach for synchronously observing and controlling brain activities of a specific region. [13][14][15][16][17][18] Among various designs of neural electrodes, flexible electrodes 15,19,20 are uniquely wellsuited for concurrent, closely opposed optical and electrophysiological data collection. The mechanical flexibility allows these electrodes to be implanted beneath a cranial window and bent out of the way at a small radius without breaking or obstructing the optical view. 15,16,20 Furthermore, and importantly, their mechanical flexibility alleviates mechanical micromotions 21 and foreign-body responses [22][23][24][25][26][27] at the tissue-electrode interface, which enables long-lasting recordings and stability to track individual neurons. 15,[28][29][30] However, the same mechanical flexibility makes surgical implantation a challenge. Here we use a 1-μm-thick, ultraflexible nanoelectronic thread (NET) 15 to provide step-by-step guidance for surgical co-implantation of flexible electrodes and a cranial window in a single procedure. We include not only surgical procedures, but also preparation of ultraflexible electrodes for implantation and attachment of retractable shuttle devices (Fig. 1). To facilitate the adaptation of this method to other flexible electrodes, we use mostly off-the-shelf components, including cover glass as the cranial window, tungsten microwires as the shuttle devices, and polyetheretherketone (PEEK) microtubes as the guiding structures for shuttle devices. The same procedures can also be adapted to implanting other customized cranial windows of different sizes and materials.

Protocol
This protocol starts with pre-surgery device preparations for which a single-shank 32-channel NET electrode array is temporarily attached to an insertion shuttle made of a tungsten microwire. It proceeds with descriptions of surgical techniques and implantation steps required to successfully implant both the NET array and a cranial window in a mouse (Fig. 1). It ends with representative electrophysiological recordings and optical imaging from the co-implantation of NET and a cranial window. This protocol assumes the starting materials of a NET array before releasing from the fabrication substrate of glass, a tungsten microwire (diameter: 50 μm, length: 10 to 15 mm) as the insertion shuttle, the bio-dissolvable adhesive polyethylene glycol (PEG) as the temporary adhesive, 31 and a PEEK microtube as the guiding structures for the insertion shuttle (Fig. 2). The prereleased NET array is bonded with a customized printed circuit board (PCB) by a standard ball grid array packaging method, in which solder balls are manually placed on the bonding pad of NETs and brought into contact with the PCB that has matching copper pads. The assembly is then heated in a reflow oven to melt the solder balls and make a permanent bonding. After bonding, epoxy is applied at the edge of the NET device where it meets the PCB as mechanical reinforcement. The tip of the tungsten shuttle is pre-sharpened using KOH electrochemical etching as previously reported. 32 This protocol also uses a #1 cover glass (diameter: 3 mm) and a 3D-printed piece to securely attach the NET device to a stereotaxic micromanipulator. All animal-involved protocols described in this protocol have been approved by the Institutional Animal Care and Use Committee at Rice University.   to the PEEK tube. Reposition the shuttle using a pair of tweezers to the proper position and reapply PEG to re-fix its location. 5. Sterilize the assembled device by ethylene oxide low-temperature sterilization. The total weight of the implant is <1 g.

Retracted PI thread
Break-off section Capillary tube Break-off section 2.2 Surgical Procedures for Co-Implantation of NET and a Cranial Window

1.
Anesthetize the animal using isoflurane (3% for induction and 1% to 2% for maintenance). Place the animal on a homeothermic blanket at 37°C to maintain body temperature stable. Monitor the depth of anesthesia by periodic toe-pinch tests (every 15 min) and the breath rate (55 to 65 breaths∕ min).

2.
Administer Ethiqa-XR (3.25 mg∕kg) for analgesia and dexamethasone (2 mg∕kg, subcutaneous) for anti-inflammation after induction. Apply ophthalmic ointment on the eyes to prevent drying of the cornea. Administer warmed lactated ringers subcutaneously 1 ml∕100 g body weight per anesthesia hour to minimize postoperative dehydration.

3.
Shave fur on the surgical site with a clipper.

4.
Fix the mouse on the stereotaxic apparatus. Position the intra aural positioning studs to symmetrically fix the position of the animal's head in the apparatus. Use only blunted style ear studs for survival procedures. To insert the incisor adapter, use small forceps to pull down the animal's lower jaw and slowly move the incisor adapter into the mouth until the incisors fit in the opening; gently pull back slightly and fix the adapter in place.
Gently position the nose clamp by resting on top of the nose using very low pressure. Adjust the incisor screw to make the head level so that bregma and lambda are equal in height. Level the head horizontally in the caudal-to-rostral direction.
Incise into the skin of the scalp and resect to expose the skull between bregma and lambda skull sutures [ Fig. 5(a)]. Remove the periosteum and score the skull with the tip of a scalpel blade to roughen the skull. 7.
Measure the desired coordinates with a pair of calipers and use a scalpel blade to mark the position. Thin the bone along the marked craniotomy perimeter using a surgical drill. Drill a circular craniotomy with a diameter of 3 mm over the target brain areas [ Fig. 5(b)]. Note: Regularly rinse the skull with sterile saline during drilling to prevent heating. The thickness of the bone along the drilled perimeter can be checked by gently pressing on the bone. Drill until the central piece of the bone is detached from the rest of the skull. 8.
Using a pair of forceps, carefully lift the central piece of the skull. If dura is attached to some part of the skull, apply sterile saline, slowly pry the edge, and follow around. 9.
Immediately after removing the central piece of the skull, place a piece of saline-soaked gel foam in the craniotomy to keep the exposed brain moist and promote coagulation. 10.
Carefully remove any remaining thin layer of bone on the edge of the craniotomy. This prevents rapid bone regrowth under the window that impedes optical clarity for chronic studies. Remove small residues of bone dust and blood clots.

11.
Drill a burr hole at the contralateral hemisphere of the brain and insert a metal wire (made of platinum, silver, or medical-grade stainless steel) as the grounding reference.

Implantation of NET
1. Fix the assembled NET device onto a stereotaxic arm using a 3D-printed case and a pair of screws [ Fig. 6(a)]. 2. Remove any gel foam from the craniotomy. Aim the device to the insertion site. To avoid surface vasculature, move in z-direction toward the brain and fine adjust the implantation location [ Fig. 6(b-I)]. Note: Avoid prolonged periods of the device in close proximity to the brain before insertion. Water vapor and condensation from the brain surface may weaken the adhesion between the NET and the insertion shuttle, leading to premature detachment prior to or during insertion. 3. Insert the device quickly. For a thin layer (1 to 2 μm) of PEG as we demonstrate here, the insertion time should be within 5 s to avoid pre-mature separation between the shuttle wire and PEG. To achieve slower insertion speed, thicker layer of PEG can be used at the expense of larger implantation footprint. After insertion, dissolve PEG outside the brain by gently dripping body-temperature saline on the device where PEG is adhered to the shuttle needle. Wait until PEG outside the brain is completely dissolved (<10 min). The device will float in saline and separate from the shuttle needle, indicating it is ready for retraction [ Fig. 6 5. To achieve precise implantation depth after shuttle needle is retracted, slowly retract the stereotaxic arm, straighten the flexible section outside the brain, and pull out the implanted section to reach the designated depth. Depth markers are fabricated on NET as guidance [ Fig. 6(b-III)]. 6. Rotate the stereotaxic arm to reduce the mounting angle of the carrier chip on the skull.
Depending on the rotation axis, this may involve iterative steps of rotation and moving the micromanipulator on the stereotaxic arm. Avoid straightening the flexible section of NET outside the brain to minimize uncontrolled stress on it [ Fig. 6(b-III)]. The carrier chip should be more than 2 mm away from the edge of the craniotomy and at an angle of 45 deg or shallower to accommodate the optical objectives for two-photon imaging [ Fig. 6(c)]. The flexible section of NET outside the brain should have some slack (at least 500 μm) to accommodate possible tissue motion as initial implantation damage subsides with time.
Fix the stereotaxic arm when the carrier chip reaches the final location and angle.

Implantation of chronic cranial window
1. After implantation of NET, add sterile saline to the brain. Place a glass window on the exposed area. If there is space between the window and the remaining skull, fill it with Kwik-sil. 2. Dry the skull. Apply a layer of low viscosity super glue on the remaining skull. Apply a layer of high viscosity super glue around the edge of the window, ensuring contact with the skull. 3. Apply a layer of Metabond dental cement over the high-viscosity super glue and the Kwik-sil, and apply additional cement to fix the NET carrier chip to the skull [ Fig. 7(a)]. 4. Mount a lightweight titanium headbar on the skull for head-restrained experiment by applying another layer of Metabond cement [ Fig. 7(b)]. 5. Cap the glass window using a piece of silicone to protect the window from mechanical scratches and tissue from photobleaching by ambient light.

Postoperative care
Animals will be monitored and documented twice in the first 24 h post-surgery, then daily for four days on body weight, incisional appearance, posture and attitude (grimacing, squinting, walking on toes or hunched. Afterward, the mouse status (body weight; body condition, score as defined by Rice University's "Monitoring Metric Scoring for Humane Endpoints to Score Animal Discomfort After Surgery" form; head cap condition) is documented weekly post-surgery.

Representative Results
After this protocol, a 32-channel NET was co-implanted with a glass cranial window over the mouse motor cortex, resulting in closed opposed electrophysiological and optical data collection  Here "all unit" represents all sorted units from Mountain Sort after rejecting noise and artifact clusters. 33 "single unit" represents the units that meet the single unit criteria: the number of spikes with an inter-spike interval smaller than 2 ms is <2% of the total events. The optical window remained clear without significant dura or bone regrowth for the same 6 month period [ Fig. 8(e)]. In another animal with a co-implanted 32channel NET and a glass cranial window over the somatosensory cortex, we showcased a variety of imaging modalities, including wide-field fluorescence imaging 34 [ Fig. 8(f)] and two-photon imaging of vasculature 9 [ Fig. 8(g)], and laser speckle contrast imaging of cerebral blood flow 35 [ Fig. 8(h)]. Mice were awake, head-fixed on a custom-made free-moving treadmill for these measurements. These results show the unique capabilities of the method described in this protocol in enabling longitudinal measurements of the multifaceted brain activity.

Discussion
This protocol provides step-by-step instruction on the implantation preparation of ultraflexible NET and its co-implantation with a cranial window, with the goal of disseminating these reproducible methods to the neuroscience community. While the protocol described here uses a single-shank NET as an example, the same procedures apply to multi-shank flexible electrodes as previously discussed. 32 The methods use off-the-shelf, cost-effective components such as tungsten microwires and PEG, which is beneficial for their facile adaptation to other designs of flexible electrodes. One caveat associated with this method is the relatively low level of control in the dissolution time of PEG, resulting from the simple design of the shuttle device and adhesion mechanism. A more sophisticated design of the shuttle device, e.g., containing microfabricated reservoirs of PEG, 36 can alleviate this problem. An alternative shuttle-device attachment mechanism, such as the needle and thread sewing mechanism, 15 provides a promising approach to precisely control when and where the insertion shuttle detaches from the flexible electrodes. Table 1 lists materials and relevant information.