Material preparation and characterization

Aqueous GO solution was diluted in deionized water to obtain a 0.15 mg ml⁻¹ solution and vacuum filtered through a nitrocellulose membrane with pores of 0.025 µm, forming a thin film of GO. The thin film was then transferred to the target substrate using wet transfer in deionized water and further thermal annealing at 100 °C for 2 min. The GO film–substrate stack was hydrothermally reduced at 134 °C in a standard autoclave for 3 h to form EGNITE. The base substrate for all characterization studies of EGNITE was a square (1 × 1 cm²) of Si/SiO₂ (400 μm/1 μm).

XPS

XPS measurements were performed with a Phoibos 150 analyser (SPECS) in ultra-high-vacuum conditions (base pressure, 5 × 10⁻¹⁰ mbar) with a monochromatic Al Kα X-ray source (1,486.74 eV). Overview spectra were acquired with a pass energy of 50 eV and step size of 1 eV and high-resolution spectra were acquired with pass energy of 20 eV and step size of 0.05 eV. The overall resolution in those last conditions is 0.58 eV, as determined by measuring the full width at half maximum of the Ag 3d5/2 peak of sputtered silver. The XPS analysis shows a strong decrease after the hydrothermal treatment of the C–O peak (associated with epoxide groups), but a small contribution of C–OH, C=O and C(O)OH due to hydroxyls, carbonyls and carboxyls that remain after reduction. The deconvolution of the O1s peak confirms such behaviour. The main contribution to the C1s signal after the hydrothermal reduction, however, comes from sp² hybridized C–C orbitals^34,57.

X-ray diffraction

X-ray diffraction measurements (θ–2θ scan) were performed in a Materials Research Diffractometer (Malvern PANalytical). This diffractometer has a horizontal ω–2θ goniometer (320 mm radius) in a four-circle geometry and worked with a ceramic X-ray tube with Cu Kα anode (λ = 1.540598 Å). The detector used is a Pixcel which is a fast X-ray detector based on Medipix2 technology.

Raman spectroscopy

Raman spectroscopy measurements were performed using a Witec spectrograph equipped with a 488 nm laser excitation line. For the measurements, Raman spectra were acquired using a 50× objective and a 600 grooves per nm grating; laser power was kept below 1.5 mW to avoid sample heating.

TEM

A focused ion beam lamella was prepared with a Helios NanoLab DualBeam (LMA-INA) for the cross-section study of the EGNITE sample. Structural analyses were performed by means of TEM using a Tecnai F20 microscope operated at 200 kV, including HRTEM and high-angle annular dark-field STEM techniques. The STEM-EELS experiment was performed in a Tecnai F20 microscope working at 200 KeV, with 5 mm aperture, 30 mm camera length, a convergence angle of 12.7 mrad and a collection angle of 87.6 mrad. As we used 0.5 eV per pixel and 250 eV as the starting energy in the core-loss acquisition, we did not acquire the Si K-edge expected at 1,839 eV, the Pt M-edge at 2,122 eV and the Au M-edge at 2,206 eV. The relative C–O atomic composition has been obtained by focusing our attention in the reduced GO layer and assuming that the edges analysed (C and O in our case) sum to 100%. This assumption is valid in our case as evidenced in the Supplementary Information maps. The energy differential cross section was computed using the Hartree–Slater model and the background using a power-low model.

Electrical conductivity

Electrical conductivity measurements were performed using a Keithley 2400 sourcemeter in two-point configuration. The samples measured consisted of EGNITE films of 1 × 1 cm² on top of a SiO₂ substrate.

Data analysis

X-ray diffraction, Raman and XPS data were analysed using Python 3.7 packages (Numpy, Pandas, Scipy, Xrdtools, Lmfit, Rampy, Peakutils, Matplotlib). The distance between planes was calculated from the X-ray diffraction measurements according to Snell’s law. Once the data were moved into the spatial domain, the maximum of the peaks was fitted. The corresponding distance gave a mean value of the distance between planes. Deviations from those mean values were calculated from the full width at half maximum of the Lorentzian fittings of the peaks on the spatial domain. XPS and Raman spectroscopy measurements were analysed by fitting a convolution of peaks on expected locations for the corresponding features. The conductivity values of the GO and EGNITE were obtained by fitting the I–V curves measured in the electrical conductivity measurements to Ohm’s law. Data are n = 1 for each measurement.

Flexible array fabrication

The fabrication of the devices is shown in Supplementary Fig. 4. Devices were fabricated on 4 inch Si/SiO₂ (400 μm/1 μm) wafers. First, a 10-µm-thick layer of PI (PI-2611, HD MicroSystems) was spin coated on the wafer and baked in an atmosphere rich in nitrogen at 350 °C for 30 min. Metallic traces were patterned using optical lithography of the image reversal photoresist (AZ5214, Microchemicals). Electron-beam evaporation was used to deposit 20 nm of titanium and 200 of gold and lift-off was performed. We used an EGNITE film of around 1 μm thickness as a trade-off between electrochemical performance and array flexibility. After transferring the GO film, aluminium was e-beam evaporated and areas on top of the future microelectrodes were defined by using a negative photoresist (nLOF 2070, Microchemicals) and lift off. Next, the GO film was etched everywhere apart from the future microelectrodes using an oxygen reactive ion etching (RIE) for 5 min at 500 W and the protecting aluminium columns were etched with a diluted solution of phosphoric and nitric acids. Then, a 3-µm-thick layer of PI-2611 was deposited onto the wafer and baked as previously described. PI-2611 openings on the microelectrode were then defined using a positive thick photoresist (AZ9260, Microchemicals) that acted as a mask for a subsequent oxygen RIE. Later, the devices were patterned on the PI layer, again using AZ9260 photoresist and RIE. The photoresist layer was then removed in acetone and the wafer cleaned in isopropyl alcohol and dried out. Finally, the devices were peeled off from the wafer and were ready to be placed in sterilization pouches to be hydrothermally treated at 134 °C in a standard autoclave for 3 h.

Microelectrode electrochemical characterization

Electrochemical characterization of the microelectrodes was performed with a Metrohm Autolab PGSTAT128N potentiostat in 1× PBS (Sigma-Aldrich, P4417) containing 10 mM phosphate buffer, 137 mM NaCl and 2.7 mM KCl at pH 7.4 and using a three-electrode configuration. An Ag/AgCl electrode (FlexRef, WPI) was used as reference and a platinum wire (Alfa Aesar, 45093) was used as counter-electrode.

Prior to performance evaluation, electrodes were pulsed with 10,000 charge-balanced pulses (1 ms, 15 µA). Exposure of electrodes to continuous pulsing protocols proceeded by 100 cyclic voltammetry cycles (−0.9 to +0.8 V) at 50 mV s⁻¹, 20 repetitions of 5,000 pulses (1 ms) and redetermination of the open circuit potential.

Data analysis

Electrochemical characterization data were analysed using Python 3.7 packages (Numpy, Pandas, Scipy, Pyeis, Lmfit, Matplotlib). Impedance spectroscopy data were fitted to an equivalent electric model consisting of a resistance (R) in series with a constant phase element (CPE). From there, the CPE value was approximated to a capacitance and divided by the microelectrode geometric area to obtain an equivalent value for the interfacial capacitance of EGNITE. Microelectrode charge storage capacitance (CSC) was calculated from cyclic voltammetry measurements by integrating the cathodic and anodic regimes of the measured current and normalizing by the scan rate. The cathodic and anodic charge storage capacitance (cCSC and aCSC) at 100 mV scan rate of EGNITE are 45.9 ± 2.4 and 34.6 ± 2.8 mC cm⁻², respectively (n = 3). As reported for other materials⁵⁸, the obtained CSCs depend on the scan rate (Supplementary Fig. 5). To assess the presence of oxygen reduction reactions, we measured the CV waveform under nitrogen-purged electrolyte⁵⁹ and did not observe substantial differences in waveform (Supplementary Fig. 6). However, our results do not fully address the impact of oxygen reduction reactions in the charge injection capacity of EGNITE and additional work needs to be done to properly investigate this. Microelectrode charge injection capacity (CIC) was established by determining the current pulse amplitude that elicited a voltage difference (after removing the ohmic drop) that matched the electrode electrochemical water window (−0.9 V for cathodic and +0.8 V for anodic versus Ag/AgCl) (Supplementary Fig. 17)⁶⁰.

Statistical analysis

Data are mean ± s.d., n = 18 for EIS and n = 3 for chronopotentiometries. Data of the map of cathodic capacitive voltage excursion are the mean of the cathodic capacitive voltage excursions for one event for each pulse shape of n = 3 electrodes.

Mechanical stability evaluation

Ultrasound sonication

EGNITE electrode arrays were placed inside a beaker filled with water in an ultrasound water bath (Elmasonic P 180H). Sonication was applied at 37 kHz for 15 min at 200 W, and followed by an additional 15 min of sonication at 37 kHz with the power elevated to 300 W. Images of electrodes were acquired before and after the sonication steps.

Bending test

The bending set-up (Fig. 2k) consisted of three cylindrical rods; the middle one (diameter, 700 µm) was lowered down, producing bending angles of 131°. Three flexible microelectrode arrays were used for the bending test. Each array contained 18 microelectrodes of 50 µm diameter. Two arrays were measured after 10 and 20 cycles while one device was measured only for 10 cycles as it was damaged during handling after measuring. The bending test cycle consisted of a 10-s-long load application plus 10 s with no load. Devices were electrochemically characterized (EIS and CV) before and after 10 and 20 bending cycles.

Epicortical neural recording

Epicortical implantation

All experimental procedures were performed in accordance with the recommendations of the European Community Council and French legislation for care and use of laboratory animals. The protocols were approved by the Grenoble ethical committee (ComEth) and authorized by the French ministry (number 04815.02). Sprague–Dawley rats (male, 4 months old, weighing ∼600 g) were anaesthetized intramuscularly with ketamine (50 mg per kg (body weight)) and xylazine (10 mg per kg (body weight)), and then fixed to a stereotaxic holder. Removing the temporal skull exposed the auditory cortex. Dura mater was preserved to avoid damaging the cortical tissue. A hole was drilled at the vertex to insert the reference electrode, and a second hole, 7 mm toward the front from the first one, was drilled to insert the ground electrode. The electrodes were 0.5-mm-thick pins used for integrated circuit sockets. They were placed to make electrical contact with the dura mater and fixed to the skull with dental cement. We then mounted the surface microelectrode ribbon on the auditory cortex as shown in Fig. 3b. The vein patterns identify the auditory cortex, in area 41 of Krieg’s rat brain map. Cortical signals were simultaneously amplified with a gain of 1,000 and digitized at a sampling rate of 33 kHz. A speaker 20 cm in front of a rat’s ear, contralateral to the exposed cortex, delivered acoustic stimuli. The stimuli delivered were monitored by a 0.25 inch microphone (Brüel & Kjaer, 4939) placed near the ear and presented in sound pressure level (dB SPL re 20 μPa). We examine the vertex-positive (negative-up) middle-latency responses evoked by alternating clicks at 80 dB SPL, and tone burst stimuli at 70 dB SPL with frequencies ranging from 5 to 40 kHz, a rise and fall time of 5 ms and a duration of 200 ms.

Data analysis

Electrophysiological data were analysed using Python 3.7 packages (Numpy, Pandas, Scipy, Neo, Elephant, Sklearn Matplotlib) and the custom library PhyREC (https://github.com/aguimera/PhyREC). r.m.s. values were calculated with a sliding window of 20 ms at frequencies above 200 Hz. Spectrograms were calculated for a range between 70 Hz and 1.1 kHz. PSD was calculated over 60 s of continuous recordings. For a given electrode array, two PSDs were calculated: in vivo (IV) and post-mortem (PM). The SNR is expressed in dB (20 × ln(r.m.s.(IV)/r.m.s.(PM))) and interpolated for 20 points logarithmically spaced between 10 Hz and 1 kHz.

Statistical analysis

Epicortical neural data presented in Fig. 3 are taken from individual measurements on a single animal. In Fig. 3c, data from 64 electrodes are presented. In Fig. 3d, data from two selected electrodes are presented. In Fig. 3f, the PSD and SNR are calculated from 64 EGNITE electrodes and are shown as mean ± s.d. In Supplementary Fig. 12c,d median data are presented for 192 EGNITE electrodes from n = 3 experiments and 60 platinum electrodes from n = 1 experiment.

Intracortical neural recording

Intracortical implantation

Animals were anaesthetized with a mixture of ketamine/xylazine (75:1, 0.35 ml/28 g i.p.) and this state was maintained with an inhalation mask providing 1.5% isoflurane. Several microscrews were placed into the skull to stabilize the implant, and the one on top of the cerebellum was used as a general ground. The probe was implanted in the prefrontal cortex (coordinates: AP, 1.5 mm; ML, ±0.5 mm; DV, −1.7 mm from bregma). The implantation was performed by coating the probe with maltose (see protocol below) to provide temporary probe stiffness and facilitate probe insertion. The probe was sealed with dental cement. TDT-ZifClip connectors were used to connect the probe to the electrophysiological system via a miniaturized cable. After the surgery, the mouse underwent a recovery period of 1 week receiving analgesia (buprenorphine) and anti-inflammatory (meloxicam) treatments. Neural activity was recorded with the multichannel Open Ephys system at a sampling rate of 30 kHz with an Intan RHD2132 amplifier. The auditory task experiments were conducted in a soundproofed box, with two speakers inside using protocols based on previously described work⁶¹. The sound stimulus consisted of a 15-ms-long white noise click, repeated 100 times (cycles), each separated by 5 s (interstimulus interval). During the task, the animal was able to move freely.

Maltose stiffener protocol

An aqueous solution of maltose is heated up to the glass transition point (T_g), between 130 and 160 °C, using a hot plate or a microwave. Once the maltose is viscous, the backside of the probe is brought into contact only with the maltose. As the maltose cools down, it rigidifies and stiffens the probe.

Data analysis

Neural signals from each electrode were filtered offline to extract SUA and LFPs. SUA was estimated by filtering the signal between 450 and 6,000 Hz and the spikes from individual neurons were sorted using principal-component analysis with Offline Sorter v.4 (Plexon). To obtain LFPs, signals were downsampled to 1 kHz, detrended and notch-filtered to remove noise line artefacts (50 Hz and its harmonics) with custom-written scripts in Python. AEP SNR was calculated as the ratio of the peak N1 amplitude and the s.d. of a 20 ms period prior to the stimulus.

Statistical analysis

Data shown in Fig. 3h,i are mean ± s.d., n = 30 as the number of averaged trials. Data recorded from the same electrode are shown at days 30, 60 and 90. Data from a single animal are presented.

Chronic epicortical biocompatibility

Surgical implantation of devices

A total of 27 adult, male, Sprague–Dawley rats were used for this study (Charles River). Animals were housed at an ambient temperature of 21 ± 2 °C and a humidity of 40–50%, on a 12 h light/12 h dark cycle. Rats were housed in groups and given free access to diet and water throughout the experimental period. Experimental procedures were carried out in accordance with the Animal Welfare Act (1998), under the approval of the UK Home Office and the local animal welfare ethical review body (AWERB). Animals were anaesthetized with isoflurane (2–3%) for the duration of surgery, and the depth of anaesthesia was monitored by the toe pinch reflex test. Animals were placed in a stereotaxic frame (Kopf, 900LS), located above a thermal blanket to maintain body temperature. A craniotomy hole (∼5 mm ×4 mm) was made 1 mm away from the midline using a dental drill with a 0.9 mm burr drill bit, the dura was removed and the epicortical device placed on the cortical surface of the brain. The craniotomy hole was sealed with Kwik-sil, followed by dental cement to secure, and the skin sutured closed. Subcutaneous injections of saline (1 ml per kg (body weight)) and buprenorphine (0.03 mg per kg (body weight)) were given to replace lost fluids and reduce postoperative pain, and anaesthesia was withdrawn.

Tissue collection and processing

Animals were terminated at 2, 6 or 12 weeks postimplantation by an appropriate method for the type of analysis to be performed.

Histology and immunohistochemistry

At 2, 6 or 12 weeks postimplantation rats were terminated via cardiac perfusion with heparinized (10 U ml⁻¹, Sigma-Aldrich) PBS, followed by 4% paraformaldehyde (PFA, Sigma-Aldrich) in PBS. Brains were postfixed in 4% PFA for 24 h, then transferred to 30% sucrose in PBS for at least 48 h before freezing in isopentane. The brains were then stored at −80 °C until cryosectioned at 25 µm. The tissue was then stained for ionized calcium binding adaptor molecule 1 (Iba-1) to determine the level of microglial activation. Briefly, tissue sections were blocked with 5% goat serum in PBS with 0.1% Triton-X for 1 h before overnight incubation at 4 °C with the primary antibody anti-Iba-1 (1:1,000, 019-19741; Wako). Sections were then stained with secondary antibody, anti-rabbit Alexa Fluor 594 (1:400, A-11012; Thermo Fisher) for 1 h at room temperature. Slides were mounted with coverslips using Prolong Gold anti-fade mounting media with 4,6-diamidino-2-phenylindole (Thermo Fisher). The probe covered an area of 3 × 3.7 mm² on the cortical surface of the brain; tissue sections selected for staining covered 3.2 mm in length of this region. Slides were imaged using a 3DHistech Pannoramic-250 microscope slide scanner at 20× and images were analysed using CaseViewer v.2.4 (3DHistech). To assess for microglia activation, a 3.2 mm area was covered, with one image analysed every 100 µm. Images were taken at 8.5× magnification which detailed a section of the epicortical probe site, 3 mm from the midline of the brain, encompassing the area directly under the probe site.

Image processing

The microscopy data were image-processed using an algorithm for microglia phenotype characterization (Supplementary Fig. 13). Microglial activation was analysed using a custom CellProfiler* (Broad Institute, v.3.1.9 from https://cellprofiler.org/) pipeline. First, the EnhanceOrSuppressFeatures module was used to enhance filamentous structures like neurites by applying the tubeness enhancement method. From the enhanced images, cells were segmented using the IdentifyPrimaryObjects module. Preliminary measurements of the cells suggested that the appropriate object diameter range was 3–40 pixels. Objects outside this diameter range or touching the edge of the image were discarded. The cells were segmented using a two-class Otsu adaptive thresholding strategy with an adaptive window size of 50 pixels. The objects identified by the IdentifyPrimaryObjects module were input to the MeasureObjectSizeShape module to calculate the necessary properties for cell classification. In the ClassifyObjects module, the category on which to base classifications was specified to be AreaShape, and Extent was selected as the corresponding measurement. The cells were classified as ‘activated’ or ‘non-activated’ based on their Extent property, which is the ratio of the area occupied by the cell to the area occupied by its bounding box. This classification approach was rationalized by the fact that activated microglia have large cell bodies and no processes, and thus occupy a far larger proportion of their bounding boxes than their non-activated counterparts. Finally, the CalculateMath and ExportToSpreadsheet modules were used to calculate and output the desired statistics.

Statistical analysis

Data sets are n = 3 for each device type (PI-only implant (PI); PI with exposed microfabricated gold (gold); and PI with microfabricated gold and EGNITE (EGNITE) at all time points) with the exception of 6 week gold which is n = 2 for ELISA data. Contralateral hemispheres were combined at each time point to give n = 9 at 2 and 12 weeks postimplantation and n = 8 at 6 weeks postimplantation. Analysis of the data was done using GraphPad Prism v.8 software. Statistical analysis was completed using a two-way analysis of variance (ANOVA) with Tukey’s multiple-comparisons test where appropriate; P < 0.05 was deemed to be significant.

ELISA

Following the implantation period, animals were terminated by cervical dislocation. Brain tissue was extracted from both the right and left hemisphere of the brain, snap frozen in liquid nitrogen and stored at −80 °C until further use. Tissue was lysed using NP-40 lysis buffer (150 mM NaCl, 50 mM Tris-Cl, 1% Nonidet P40 substitute, Fluka, pH adjusted to 7.4) containing protease and phosphatase inhibitor (Halt Protease and Phosphatase Inhibitor Cocktail, Thermo Fisher), followed by mechanical disruption of the tissue (TissueLyser LT, Qiagen). Samples were then centrifuged for 10 min at 5,000 r.p.m., and the supernatant stored at 4 °C until further use. The LEGENDplex Rat Inflammation Panel (catalogue number 740401, BioLegend), a bead-based multiplex ELISA kit, was run to quantify the following cytokines; IL-1α, IL-1β, IL-6, IL-10, IL-12p70, IL-17A, IL-18, IL-33, CXCL1 (KC), CCL2 (MCP-1), granulocyte–macrophage colony-stimulating factor, interferon-γ and tumour necrosis factor. The kit was run according to the manufacturer’s instructions, with protein loaded at a fixed volume of 15 µl. Following incubation with supernatant the beads were run on a BD FACSVerse flow cytometer, and the data analysed using LEGENDplex data analysis software.

Neural stimulation

Intrafascicular implantation

All animal experiments were approved by the Ethical Committee of the Universitat Autònoma de Barcelona in accordance with the European Communities Council Directive 2010/63/EU. Animals were housed at 22 ± 2 °C under a 12 h light/12 h dark cycle with food and water freely available. The sciatic nerve of anaesthetized female Sprague–Dawley rats (250–300 g, ∼18 weeks old) was surgically exposed and the TIME electrodes were implanted transversally across the sciatic nerve with the help of a straight needle attached to a 10-0 loop thread⁴⁶. The process was monitored under a dissection microscope to ensure the correct position of the active sites inside the nerve fascicles (Fig. 4b). During the experiments, the animal body temperature was maintained with a heating pad.

Nerve stimulation was performed by applying trains of biphasic current pulses of a fixed duration of 100 µs per phase and increasing amplitude from 0 to 150 µA in 1 or 3 µA steps at 3 Hz for 33 s (Stimulator DS4, Digitimer) through the different EGNITE microelectrodes. Simultaneously, the CMAPs were recorded from GM, TA and PL muscles using small needle electrodes (13 mm long, 0.4 mm diameter, stainless steel needle electrodes A-03-14BEP, Bionic) placed in each muscle⁶². The active electrode was placed on the muscle belly and the reference at the level of the tendon. Electromyography recordings were amplified (×100 for GM and TA, ×1,000 for PL; P511AC amplifiers, Grass), band-pass filtered (3 Hz to 3 kHz) and digitized with a PowerLab recording system (PowerLab16SP, ADInstruments) at 20 kHz.

Data analysis

The amplitude of each CMAP was measured from baseline to the maximum negative peak. The voltage peak measurements were normalized to the maximum CMAP amplitude obtained for each muscle in the experiment. A selectivity index (SI) was calculated for each active site as the ratio between the normalized CMAP amplitude for one muscle, CMAP_i, and the sum of the normalized CMAP amplitudes in the three muscles, following the formula SI_i = nCMAP_i/∑nCMAP_j, at the minimum stimulation current amplitude that elicited a minimal functionally relevant muscular response (defined as at least 5% CMAP amplitude for one of the muscles with the respect to the maximum CMAP amplitude of that muscle that had been previously determined). Then, the active sites with highest SI for each of the three muscles were selected as the SIs for each muscle in a given experiment.

Chronic intraneural biocompatibility

Following a previously reported procedure^50,63, the sciatic nerve of anaesthetized Sprague–Dawley female rats (250-300 g, ∼18 weeks old) was exposed and the devices for in vivo biocompatibility with and without EGNITE were longitudinally implanted in the tibial branch of the sciatic nerve (n = 6–8 per group). Briefly, the nerve is pierced at the trifurcation with a straight needle attached to a 10-0 loop thread (STC-6, Ethicon); the thread pulls the arrow-shaped tip of the bent electrode strip. The tip is cut to take away the thread, and the tips of each arm are slightly bent to avoid withdrawal of the device. A longitudinal implant was chosen because it allows a better study of the foreign body response inside the nerve⁵⁰.

Nerve and animal functional assessment

Animals were evaluated during follow-up postimplantation by means of nerve conduction, algesimetry and walking track locomotion tests⁶². For conduction tests, the sciatic nerve of the implanted and contralateral paws was stimulated by needle electrodes at the sciatic notch and the CMAP of the PL muscle was recorded as above. The latency and the amplitude of the CMAP were measured. For the algesimetry test, rats were placed on a wire net platform and a mechanical non-noxious stimulus was applied with a metal tip connected to an electronic Von Frey algesimeter (Bioseb). The nociceptive threshold (force in grammes at which the animals withdrew the paw) of implanted versus contralateral paws was measured. For the walking track test, the plantar surface of the hindpaws was painted with black ink and each rat was left to walk along a corridor. The footprints were collected, and the sciatic functional index calculated⁶².

Histology

After 2 or 8 weeks, animals were perfused with PFA (4%), and the sciatic nerves were harvested, postfixed, cryopreserved and processed for histological analysis. For the evaluation of the FBR, sciatic nerves were cut in 15-μm-thick transverse sections with a cryostat (Leica CM190). Samples were stained with primary antibodies for myelinated axons (anti-RT97 to label Neurofilament 200K, 1:200; Developmental Studies Hybridoma Bank) and macrophages (anti-Iba-1, 1:500; Wako). Then, sections were incubated for 1 h at room temperature with secondary antibodies donkey anti-mouse Alexa Fluor 488 and donkey anti-rabbit Alexa Fluor 555 (1:200, Invitrogen). Representative sections from the central part of the implant in the tibial nerve were selected, images taken with an epifluorescence microscope (Eclipse Ni, Nikon) attached to a digital camera (DS-Ri2, Nikon) and image analysis performed with ImageJ software (National Institutes of Health). The amount of Iba-1-positive cells in the whole area of the tibial nerve was quantified and the thickness of the tissue capsule was measured as the mean distance of each side of the implant to the closest axons.

Statistical analysis

For statistical analysis of data, we used one- or two-way ANOVA followed by Bonferroni post hoc test for differences between groups or times. GraphPad Prism software was used for graphical representation and analysis. Statistical significance was considered when P < 0.05.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

• SEO Powered Content & PR Distribution. Get Amplified Today.
• PlatoData.Network Vertical Generative Ai. Empower Yourself. Access Here.
• PlatoAiStream. Web3 Intelligence. Knowledge Amplified. Access Here.
• PlatoESG. Carbon, CleanTech, Energy, Environment, Solar, Waste Management. Access Here.
• PlatoHealth. Biotech and Clinical Trials Intelligence. Access Here.
• Source: https://www.nature.com/articles/s41565-023-01570-5