In-body current path manipulation with minimal attenuation

Abstract

Wireless power transfer–based neuromodulation has emerged as a promising alternative to battery-powered implants. However, its practical application is hindered by limited therapeutic efficacy resulting from low power transfer efficiency, shallow penetration depth, and safety concerns. In this study, we report an in-body current path manipulation and concentration for advanced targeted neuromodulation, overcoming the limitations of conventional technologies. By implanting a focusing electrode, we were able to direct the triboelectric current, which has a low frequency and high impedance generated by human movement, toward the target area, with the concentrated current exhibiting minimal attenuation regardless of the electrode size, implantation site, and depth. Applying our technology to modulate damaged neural systems confirmed therapeutic efficacy and validated safety, demonstrating its potential for next-generation targeted neuromodulation.

Controlling electric current path inside the body enables minimally invasive, battery-free neuromodulation with minimal loss.

INTRODUCTION

Implantable medical devices have ushered in an era of therapeutic interventions by enabling precise stimulation to promote neural regeneration and manage chronic diseases (1, 2). However, their reliance on batteries poses various limitations, including the need for timely removal of depleted batteries from within the body, increasing risks of complications and patient discomfort (3). Moreover, ensuring sufficient energy capacity requires the battery to meet a threshold size, which limits implantation sites and necessitates the use of lead wires, thereby increasing the risk of infection. In response to these concerns, batteryless neuromodulation using wireless power transfer (WPT) has emerged as a promising alternative (4). However, existing WPT methods, such as inductive coupling (IC), radio frequency (RF), and far-field electromagnetic waves, suffer from low power transfer efficiency, limited transmission distances, and safety concerns. For instance, IC experiences exponential attenuation, requiring large devices for power transmission beyond depths of 1 cm, while exhibiting notably low transmission efficiency (5). Furthermore, the safety of RF for the human body needs to be verified, and it is hindered by performance degradation within the body, primarily due to the body shadowing effect (6–8). Other approaches, including ultrasound, magnetoelectric, and capacitive coupling, are similarly constrained by low efficiency and alignment requirements (9). Under these constraints, WPT-based neuromodulation technology faces critical challenges that hinder its clinical application due to inconsistent and unstable neural modulation. Therefore, a next-generation neuromodulation technology is highly required, one that can reliably deliver electrical signals to targeted nervous systems regardless of depth or distance, while ensuring minimal invasiveness and confirmed biosafety, thereby overcoming the limitations of battery-dependent and WPT-based methods.

The human body has recently been used as a transmission medium in body area networks (BANs) to reduce signal loss by leveraging its conductive properties (10, 11). However, frequencies in the kilohertz to gigahertz range commonly used in BAN experience reflection and absorption at cellular membranes, leading to signal attenuation (text S1) (12). Considering these phenomena, we selected low-frequency triboelectric currents (on the order of a few hertz) as the targeted neuromodulation energy source to minimize attenuation. Building on the principle that electric currents follow the path of least resistance, we subsequently designed in-body triboelectric current path manipulation for a targeted neuromodulation as an alternative to traditional technologies (Fig. 1A). In this technology, triboelectric generator (TG) and the human body form a closed loop through an earth ground, allowing triboelectric currents to flow into the body. Low-frequency triboelectric currents predominantly flow through the extracellular fluid (ECF), which has low impedance (Fig. 1B and fig. S1) (13, 14). In addition, the high output impedance of TG further reduces tissue-induced current attenuation (fig. S2 and text S2), and their relatively long wavelength (approximately 75 million times longer than the 2-m transmission length) mitigates the impact of reflections during electrical propagation, ensuring stable transmission. However, because triboelectric currents traverse the body in a random and unpredictable, we implemented a low resistance focusing strategy that directs triboelectric currents toward a focusing electrode (FE), which has lower resistance than ECF, thereby enabling a controlled flow path (fig. S3). This means that triboelectric current is directed toward the target area by implanting FE at the desired site, enabling precise modulation of the tissue. We also observed minimal current attenuation even with a small FE size, ensuring that the technology is both minimally invasive and straightforward while achieving superior neuromodulatory outcomes (fig. S4).

Fig. 1. Current path manipulation within the body.

(A) Schematic representation of the components and mechanism of in-body current path manipulation. TG transfers triboelectric current externally through a hydrogel patch on the skin, while, inside the body, the low-resistance FE concentrates triboelectric current onto itself. (B) Frequency-dependent electric current flow characteristics in the human body and a circuit diagram for low-frequency scenarios. The low-frequency electric current predominantly flows through ECF. CPE, cytopathic effect. (C) Illustration of FE implantation depth and triboelectric current attenuation as a function of implantation depth in pig models (n = 3). (D) Experiment setup for evaluating focused electric current in pig experiment and focused triboelectric current measurements depending on the FE implantation site (n = 3). Triboelectric current was transferred to hydrogel on the hind leg skin, and FE was implanted at varying locations: hind leg, back, front leg, and neck (deep implant). (E) Focused triboelectric current measurements in different animal models, including rat and mouse. Schematics in (A) and (E) were created in BioRender. Kang, D. (2025) https://BioRender.com/z93p821.

RESULTS

Control of in-body triboelectric current pathway to targeted area

To generate triboelectric currents for targeted neuromodulation, we developed a high-performance TG capable of externally delivering stable output to the human body. In TG, we used cross-linked poly(vinylidene fluoride-trifluoroethylene) [P(VDF-TrFE)] and perfluoroalkoxy (PFA) as the tribopositive and tribonegative layers, respectively (fig. S5) (15–18). For tribopositive material, a cross-linked P(VDF-TrFE) was synthesized through the formation of imine bonds, using polyethylene glycol (PEG) diamine as the cross-linker (fig. S6). The chemical structure of cross-linked P(VDF-TrFE) was confirmed using Fourier transform infrared results and x-ray diffraction analysis (figs. S7 and S8) (19–21). Among the components of this material, the nitrogen in poly(ethylene glycol) bis(amine) (PEG diamine) affects triboelectric behavior due to the formation of negative local dipoles, enhancing its tribopositive behavior (22). We confirmed his enhanced tribopositive behavior in cross-linked P(VDF-TrFE) (figs. S9 to S13). Although tribopositive behavior of P(VDF-TrFE) improved with increased cross-linking density, a density of 1.44% was selected as optimal for tribopositive materials, primarily due to its superior mechanical properties (figs. S14 and S15). A multistacked TG was fabricated using optimized material. Long-term output stability was verified through continuous operation for 8 hours daily over 20 days, during which no performance degradation was observed. These results demonstrate the durability of our TG, confirming its potential for both short-term and long-term neuromodulation applications (fig. S16).

To evaluate focused electric current in FE for the targeted neuromodulation, we used a pig model due to its physiological similarities to humans. An earthing pad connected to the building ground was positioned beneath the animal model to establish a potential difference and enable electric current flow (7, 23). The oscilloscope was isolated from the building ground to prevent undesirable closed circuits (see Materials and Methods and fig. S17). At shallow (<1 cm), deep (>1 cm), and the peripheral nerve (~4 cm) FE implant depths, we measured focused triboelectric current in FE, with their average current attenuation exhibiting the values of −0.308, −0.320, and −0.324 dB, respectively (Fig. 1C and figs. S18 to S20). Moreover, we evaluated the focused electrical current in the FE across different implantation sites (the hind leg, back, front leg, and neck, the focused electrical current), which yielded values of −0.320, −0.322, −0.324, and −0.326 dB, indicating minimal loss (Fig. 1D and figs. S21 to S24). In addition to the pig model, we also used mouse and rat models to confirm that the focused electrical current at FE exhibits negligible attenuation (Fig. 1E and fig. S25). To confirm that the focused electrical current originates from TG, we compared its optimal impedance values with those of the TG output (fig. S26). Moreover, to demonstrate the miniaturization capability of our technology, we measured triboelectric current with FE of varying sizes. Regardless of FE size, the focused electrical current exhibited a value of approximately −0.320 dB (fig. S27). To validate the applicability of alternative conductors, we used gold, Ag/AgCl gel, and graphite (each with impedance lower than ECF) as FEs and measured current attenuation; all showed minimal loss of approximately −0.32 dB, comparable to Mg (fig. S28). Moreover, when two or more FEs were implanted in the body, the FEs were electrically connected in parallel, allowing simultaneous current flow. This indicates that our technology can modulate multiple sites concurrently (fig. S29). All experiments were conducted without considering alignment between TG and FE, and consistent results were obtained regardless of alignment.

To enable minimal invasiveness and eliminate the need for subsequent removal surgeries, we used biodegradable magnesium (Mg) as the material for FE (figs. S30 and S31) (24, 25). Given the critical importance of biocompatibility, we evaluated the cytotoxicity levels via the 3-{4,5-dimethylthiazol-2-yl}-2,5-diphenyltetrazolium bromide assay by culturing human fibroblasts (American Type Culture Collection, CRL-1502) on Mg surfaces. After 3 days, the relative cell viability for cells grown on each material exceeded 90%, indicating no cytotoxicity from the FE (fig. S32). To further validate biosafety, we conducted blood tests following FE implantation after 7 days and 30 days postcomplete degradation. The results showed no signs of toxicity, indicating no adverse effects on the body (figs. S33 to S35).

Mechanism of triboelectric current concentration within the body

To elucidate the principle of minimal attenuation when focusing on FE implanted into the body, we evaluated bioelectrical impedance characteristics to analyze the current pathways and the components enabling current flow. First, mice were surgically dissected into three tissues: epidermis + dermis, adipose tissue, and muscle, followed by evaluation (figs. S36 to S38) (26). Among these, to consider hydrogel-based electrical current transmission in our technology, we attached a commercial hydrogel patch to the exposed epidermis and performed electrochemical impedance spectroscopy (EIS) through the patch to evaluate the impedance of the epidermis + dermis (27). From the Bode plot of each skin layer, we analyzed the dominance of resistive reactance at the frequency of the TG output by observing the low phase angle shift (Fig. 2, A to C). This indicates that triboelectric current primarily flows through ECF, which has insufficient impedance to attenuate triboelectric current (figs. S2 and S39 and table S1). To ensure the reliability of the results, we conducted cyclic voltammetry (CV) analysis and confirmed the absence of nonideal electrochemical reactions under the EIS measurement conditions (figs. S40 to S42). Moreover, EIS measurements at different immersion times (0 and 20 min) confirmed that phosphate-buffered saline (PBS) had a negligible effect on the tissue, with only minimal changes observed (figs. S43 and S44 and table S2). We next performed in vivo experiments to measure the electrical impedance from the hydrogel attached to the mouse skin to the cuff electrode wrapped around the peripheral nerve at low frequency (Fig. 2D and fig. S45). The observed low phase angle variation provides evidence that the current is transmitted through the ECF; further, fluorescence imaging confirmed that this transmission is enabled by the interconnection of each layer through the ECF (figs. S46 and S47).

Fig. 2. Principle in in-body pathway manipulation of minimal attenuation.

(A to C) Bode plots of various skin layers from a mouse in PBS: (A) epidermis + dermis (with hydrogel), (B) adipose tissue, and (C) muscle. The lines represent fits to the data using the Levenberg-Marquardt method. (D) Quantitative analysis of electrical impedance, including capacitance, resistance, impedance, and phase, in a mouse at 4 Hz (n = 20). (E) In-body triboelectric current path manipulation and concentration imaging based on a voltage-sensitive dye according to the implantation position (top, back; bottom, left leg) of FE. Scale bars, 5 cm (overall view, left) and 2 mm (fluorescence images, right).

Last, we examined the effect of FE on triboelectric current concentration within the body. Before analyzing the current concentrated in the FE, we confirmed that Mg, used as the FE, acts as a resistive element in the body and exhibits extremely low resistance. This low impedance was stably maintained even after 7 days of implantation (fig. S48). Given its lower resistance than ECF, it facilitates control over the current pathway within the body. Subsequently, fluorescence imaging confirmed that the triboelectric current was directed toward the FE, depending on its implanted location (Fig. 2E). In the absence of the FE, the current exhibited a random distribution rather than a targeted flow (fig. S49). However, upon FE implantation, the current pathway shifted, initially focusing on the FE before reaching the ground, confirming the alteration in the current trajectory. Building on these findings, we represented the mechanism of in-body triboelectric current path manipulation through a circuit diagram, offering a comprehensive overview of our technology (fig. S50).

Targeted neuromodulation based on in-body triboelectric current path manipulation

To verify the neuromodulatory efficacy using our technology, we evaluated its capacity to regenerate damaged peripheral nerves in a mouse model (text S3) (28, 29). First, we analyzed changes in neuronal activity induced by focused triboelectric current using fluorescent imaging with Fluo-4 Direct, an intracellular calcium dye insensitive to extracellular calcium. Compared to the wild-type model, the model in which FE was wrapped around the peripheral nerve and stimulated via a focused triboelectric current exhibited stronger fluorescence. This finding indicates that our technology effectively enhances neuronal activity to a heightened level (Fig. 3A). Furthermore, to demonstrate the potential for nerve regeneration using this technology, we established the following experimental protocol (Fig. 3B): (i) A mouse model with peripheral nerve injury was designed by severely compressing the peripheral nerve. (ii) Triboelectric currents from TG were transmitted through a hydrogel attached to the skin of the mouse for 30 min for 7 days. (iii) Last, we evaluated the therapeutic outcomes of in-body triboelectric current path manipulation through histopathological, behavioral, and electrophysiological studies. On the basis of the protocol, we implanted FE exclusively on the right nerve in a mouse model with bilateral peripheral nerve injuries, administered neuromodulatory treatment, and compared regeneration across both sides. Histopathological studies, including semithin and electron microscopy (EM) examinations, showed that myelination was enhanced in the right peripheral nerve compared to the left nerve (fig. S51). Myelination indicates that focused triboelectric current is able to accelerate the regeneration of injured nerves. Moreover, nerve conduction studies (NCSs), an electrophysiological test of peripheral nerves, were performed to evaluate whether restored myelin affects complex motor nerve functions. NCS results showed that the right peripheral nerve wrapped with FE was better than the left nerve in terms of motor nerve conduction velocities (MNCVs) and compound muscle action potentials (CMAPs) (fig. S52). The varying degrees of nerve regeneration observed within the same subject, depending on the presence of the FE, provide compelling evidence that the FE focuses triboelectric current within the body.

Fig. 3. In-body triboelectric current path manipulation–based neural modulation.

(A) Fluorescent images showing neuronal activation in response to in-body current path manipulation. Application of triboelectric current from TG resulted in strong fluorescence around the FE. (B) Images of the peripheral nerve in noninjured and injured mouse models, along with the experimental setup for validating neuromodulation using our technology. Scale bar, 2 mm. (C) Images of semithin sections (magnification ratio, 100 times) and EM (scale bars, 2 μm) displaying the toluidine blue–stained cross section of the peripheral nerve. (D) Percentage of the myelinated axons and the unmyelinated axons (wild type, n = 1000; injured model, n = 1000; FE model, n = 1000; TG model, n = 1000; TG + FE model, n = 1000). (E) Histogram depicting the distribution of inner diameters of myelinated axons (wild type, n = 1000; injured model, n = 1000; FE model, n = 1000; TG model, n = 1000; TG + FE model, n = 1000). (F) Scatter plots illustrating the correlation between g-ratio and axon diameter (wild type, n = 1000; injured model, n = 1000; FE model, n = 1000; TG model, n = 1000; TG + FE model, n = 1000). (G) Behavior assessment using rotarod in five models (wild type, n = 15; injured model, n = 15; FE model, n = 15; TG model, n = 15; TG + FE model, n = 15). (H to K) Recorded nerve conduction velocity and action potentials after the in vivo electrotherapy for the compression peripheral nerve injury mouse model (wild type, n = 15; injured model, n = 15; FE model, n = 15; TG model, n = 15; TG + FE model, n = 15). (H) MNCV, (I) CMAP, (J) SNCV, and (K) SNAP were evaluated in five models. Data are presented as mean values ± SD. P values were determined using a two-sided t test; n.s., nonsignificant; *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

To demonstrate that the therapeutic effect stemmed from in-body triboelectric current path manipulation rather than other factors, we established five models (wild type, injured, FE, TG, and TG + FE) and performed a comparative analysis across these groups (fig. S53). The wild-type model consisted of mice with no injury or treatment. The injured model served as a control group with peripheral nerve injuries and no additional intervention. The FE model involved the implantation of FE in the injured peripheral nerve without the application of triboelectric current from TG. For the TG model, triboelectric current was applied via TG without the implantation of FE for 30 min daily for 1 week. Last, the TG + FE model had the FE wrapped around the injured nerve and applied triboelectric current from TG for 30 min daily for 1 week. Histopathologic studies, including semithin and EM examinations performed across five models, demonstrated that the TG + FE model had increased myelination, indicating greater therapeutic efficacy compared to the other models (Fig. 3C). Quantitative analysis further revealed that the TG + FE model had a higher percentage of myelinated axons compared to the injured, FE, and TG models. Notably, the TG + FE model showed recovery to levels of myelination similar to those of the wild-type model (Fig. 3D). Statistical analysis based on histopathological studies of each group also showed that the TG + FE model had an increased number of myelinated axons and a larger average axonal diameter. The distribution curve of the TG + FE model shifted to the right compared to the injured, implanted, and electrified models, showing that the injured peripheral nerves were recovering (Fig. 3E). In addition, the TG + FE model showed a lower g-ratio than the injured, FE, and TG models, further suggesting enhanced myelination (Fig. 3F). In behavioral assessments using the rotarod test, the TG + FE model was found to have longer latencies, which were almost twice as long as those observed in the injured, FE, and TG models (Fig. 3G and movie S1). In the NCS experiments, the TG + FE model showed higher MNCVs and CMAPs in the motor nerve test compared to the other models (Fig. 3, H and I, and fig. S54). In the sensory nerve test, sensory nerve conduction velocities (SNCVs) and sensory nerve action potentials (SNAPs) were restored to levels similar to the wild type (Fig. 3, J and K). These results showed that the restored myelin by in-body triboelectric current path manipulation effectively modulates nervous system function. Through a comparative analysis of nerve regeneration across the five models, we confirmed that the observed therapeutic outcome resulted from our technology as a whole, rather than from the individual contributions of FE or TG components.

Biosafety of in-body triboelectric current path manipulation

Typically, it is perceived that electrical currents passing through the human body may induce adverse physiological effects, such as muscle damage and burns. Contrary to this common belief, safely regulated electrical currents have been effectively used in various bioelectronics. For example, bioimpedance analyzers apply a low-level current to measure body fat and muscle mass, while transcutaneous electrical nerve stimulators alleviate pain sensations by transmitting electrical impulses through the skin (30, 31). In addition, numerous bioelectronics use currents to the human body for health monitoring and therapeutic purposes, and they are considered highly safe, having received clearances from the US Food and Drug Administration (table S3). Furthermore, the Institute of Electrical and Electronics Engineers (IEEE) provides the standards for the safety limits on the exposure of the human body to electric/magnetic fields (32). According to IEEE standards, currents below 0.5 mA are imperceptible to humans and are deemed to have a negligible impact on the human body. In comparison, the output of TG demonstrated lower current levels. This performance in our technology aligns with established safety standards, thereby ensuring the safety and reliability of our technology (text S4).

Beyond securing safety through comparison of in-body triboelectric current path manipulation with established safety standards and previously approved bioelectronics, we also experimentally verified its safety by supplying triboelectric power to both cellular cultures and mouse models for 30 min daily over a period of 7 days. First, to assess the impact of triboelectric power at the cellular level, we cultured fibroblast cells and subjected them to this treatment (fig. S55). Using differential staining techniques in live/dead assays, live cells can be distinguished by their green fluorescence. Concurrently, dead cells are identified by red staining, a consequence of membrane permeability alterations. The live/dead cell assay results, showing an absence of dead cells, indicate that the application of triboelectric power does not exert detrimental effects on fibroblast cells (fig. S56). Furthermore, hematoxylin and eosin (H&E)–stained sections from various organs in mouse models were obtained to analyze the effect of triboelectric current on the body (Fig. 4A). Upon examination of skin, heart, liver, lung, spleen, and kidney tissues, the results revealed an absence of inflammatory responses and fibrosis. The effect on the muscle was also evaluated by assessing changes under muscle condition and volume through magnetic resonance imaging (MRI) of the hind leg in the mouse model (Fig. 4, B and C). MRI images of the hind leg muscles demonstrate no impact or destruction by our technology with no meaningful changes in muscle volume in both the left and right hind legs. These observations indicate an absence of damage or adverse effects on these tissues, highlighting the biocompatibility and safety of in-body triboelectric current path manipulation.

Fig. 4. Elucidation of harmlessness of in-body current path manipulation.

(A) H&E-stained tissue images illustrating the inflammatory responses in skin, heart, liver, lung, spleen, and kidney tissues for the control group and TG + FE models. The images compare the control group with the TG + FE models, demonstrating no inflammatory effects induced by in-body triboelectric current path manipulation. (B) MRI images of the hind leg of a mouse from the control group and in-body current path manipulation, describing the assessment of any discernible differences under tissue condition or damage attributable to in-body current path manipulation. Scale bars, 5 mm. (C) Analysis of changes in the muscle volume of the left and right legs (n = 15), indicating the absence of significant harm or muscle degradation due to in-body current path manipulation. Data are presented as mean values ± SD.

DISCUSSION

In this study, we present in-body triboelectric current path manipulation and concentration capable of realizing advanced targeted neuromodulation. Using pig models, we demonstrated that triboelectric currents are focused on FE, regardless of its implantation depth, position, or size, with minimal attenuation, enabling stable neuromodulation. In addition, it overcame the limitations of conventional methods for targeted neuromodulation, such as alignment issues and biosafety concerns, proving this technology to be an innovative approach. Considering its superior neuromodulatory efficacy, high penetration capability with minimal attenuation, biosafety, and minimally invasive property, our technology holds promise for the development of a next-generation neuromodulation system applicable to deep tissues.

MATERIALS AND METHODS

Materials

Cyclohexanone, PEG diamine (2000 Da) was purchased from Sigma-Aldrich. Magnesium (Mg) was purchased from GFM in Republic of Korea. P(VDF-TrFE) with a trifluoroethylene composition of 30% was supplied by Piezotech.

Preparation of cross-linked P(VDF-TrFE)

A P(VDF-TrFE) solution was prepared by dissolving 2 g of P(VDF-TrFE) copolymer and PEG diamine (content varying according to different cross-linking densities) in 20 ml of cyclohexanone using a magnetic stirrer. After removing air bubbles under vacuum, it was coated on a glass substrate using an automatic bar coater and left to naturally dry for 24 hours in a fume hood. After volatilizing most of the solvent, the film was placed in a vacuum oven for drying at 50°C for 6 hours and cross-linked at 240°C in a vacuum oven for 1 hour. After naturally cooling down to room temperature, the cross-linked P(VDF-TrFE) film was peeled off from the glass substrate. The thickness of the film was ~40 μm.

Fabrication of TG

TG was constructed using a multilayer stacking of contact-separation layers. Cross-linked P(VDF-TrFE) and PFA were chosen as the tribopositive and tribonegative materials, respectively, to enable effective contact electrification. Silver was deposited on a polyethylene terephthalate (PET) sheet to function as the electrode layer. This PET sheet, with the electrode layer, had cross-linked P(VDF-TrFE) and PFA attached to it. Subsequently, each sheet, attached cross-linked P(VDF-TrFE) and PFA, was arranged to face the other across a 5-mm gap, creating a multilayer stacking structure.

Measurement and characterization

Electric currents were measured and recorded using an oscilloscope (Rohde & Schwarz, MXO44) with a low-noise current amplifier (FEMTO, DLPCA-200). X-ray photoelectron spectroscopy and x-ray diffraction analysis were performed using Thermo Fisher Scientific ESCALAB250 and Bruker AXS D8 Discover. Stress-strain curve was obtained using a universal testing machine (Instron, 5844). Kelvin probe force microscopy measurements (Park Systems, XE-100 with PT/Cr-coated silicon tips) were performed to measure the surface potentials of materials. To measure the current attenuation, a closed circuit was established using an earthing pad. The earthing pad, which is conductive and connected to the building ground, was placed beneath the animal models to create an earth grounding system. Moreover, to eliminate noise, we used the fast Fourier transform method. After evaluating the focused current in this experimental setup, the following Eq. 1 was used for current attenuation calculation

$$\mathit{\text{Current attenuation}}\mathit{=}\mathit{20}\mathit{\text{log}}({\mathit{I}}_{\mathit{\text{FE}}}\mathit{/}{\mathit{I}}_{\mathit{\text{TG}}})$$ (1)

where $I_{\text{FE}}$ and $I_{\text{TG}}$ represent focused triboelectric current at FE and triboelectric current from TG, respectively.

$$\mathit{\text{PTE}}\left(\mathit{\text{dB}}\right)\mathit{=}\mathit{10}\mathit{\text{log}}({\mathit{P}}_{\mathit{\text{out}}}\mathit{/}{\mathit{P}}_{\mathit{\text{in}}})$$ (2)

where $P_{\text{in}}$ is the input power delivered to the transmitter (Tx) and $P_{\text{out}}$ is the output power received at the load.

Electrochemical impedance analysis

EIS measurements of each skin layers were conducted using a three-electrode system with a potentiostat (Gamry References 600+). A 1× PBS solution, with a concentration comparable to that of bodily fluids, was used as the electrolyte. In the three-electrode configuration, an Ag/AgCl electrode served as the reference electrode, a Pt wire as the counter electrode, and each skin layers as the working electrode. The frequency sweep was performed from 1 Hz to 1 MHz, with measurements taken at 10 points per decade. R&SLCX LCR meter (Rohde & Schwarz, Germany) was used to conduct in vivo EIS experiments.

In vivo experiment

All studies on mice and rats were approved by the Institutional Animal Care and Use Committee (IACUC) of Samsung Medical Center (2022-05-28-1) and Sungkyunkwan University (2022-05-28-1). Furthermore, all studies on pigs were also approved by IACUC of BIOSTEP (23-KE-0493).

The 8-week-old institute for cancer research (ICR) mice were anesthetized through the inhalation of 2 to 2.5% isoflurane. Each mouse was then laid in a prone position, and the skin on the outer side of the left thigh was carefully shaved. Using fine scissors, a minor incision was made at the midthigh level, which was followed by carefully separating the biceps femoris from the gluteus muscles to reveal the peripheral nerve. For inducing a controlled compression injury, a HALSEY needle holder (with smooth jaws and a length of 12.5 cm) was used. The needle holder, which has three stages of compression, used its second stage to exert pressure on the peripheral nerve for a duration of 5 s. The operation was completed by suturing the skin and muscle back together and allowing a recovery period of 2 days for the surgical wound.

For in vivo electrostimulation, anesthesia was induced in 8-week-old ICR mice via isoflurane inhalation (Hana Pharm, 200711400). Subsequently, the subject was secured in a prone orientation, and the dorsal integument was depilated. Electrophysiological interfaces, namely FE and TG, were aseptically implanted adjacent to the peripheral nerve following sterilization with low-temperature ethylene oxide gas. The surgical apertures were then approximated with cutaneous sutures. Prophylactic antisepsis was achieved through the application of a povidone-iodine preparation pad (Green Pharm, 201801264). An electroconductive hydrogel patch was affixed to the depilated dorsal region and linked to the TG via a conductive filament to administer electrostimulation.

Fluorescence imaging

The tissue was collected from an ICR mouse. Fluorescein isothiocyanate (FITC)–dextran (Sigma-Aldrich, 46944) and Hoechst 33342 (Thermo Fisher Scientific, 62249) were administered intraperitoneally, followed by a waiting period of 15 to 20 min before euthanization and tissue biopsy. The collected tissue specimens were sectioned into ultrathin slices using a cryo-ultramicrotome (RMC Products PowerTome). FITC-dextran, which emits green fluorescence (excitation, ~495 nm; emission, ~519 nm), was used to visualize ECF distribution, while Hoechst 33342, a blue fluorescent nuclear stain (excitation, ~350 nm; emission, ~461 nm), was applied to label cell nuclei. Fluorescent images were acquired using a 5× objective under a Zeiss LSM 900 confocal microscope

In vivo fluorescence imaging was conducted using an IVIS Spectrum system (Xenogen, USA). The FluoVolt Membrane Potential Kit (Thermo Fisher Scientific, F10488) was prepared following the standard preparation guidelines, diluted in PBS, and administered via intraperitoneal injection under isoflurane anesthesia. Following administration, mice remained under continuous anesthesia, allowing the dye to circulate for 20 to 30 min before imaging. Fluorescence signals were acquired using an IVIS Spectrum system with an excitation wavelength of ~465 nm and an emission wavelength of ~540 nm, corresponding to the FITC filter set. Mice were positioned in the imaging chamber under anesthesia, and fluorescence images were captured at baseline and postadministration time points. Fluorescence intensity was analyzed using Living Image software 4.72 version (PerkinElmer, USA).

Calcium imaging was conducted using an epifluorescence microscope equipped with an FITC filter. Tissue samples were incubated with Fluo-4 Direct at the recommended concentrations, following the manufacturer’s protocol (Invitrogen, F10471). The incubation was performed for 30 min at 37°C in a humidified incubator, followed by an additional 30 min at ambient temperature. Subsequently, tissues were rinsed three times with a neural maintenance medium. Calcium transients were captured with an excitation wavelength of 488 nm and an emission wavelength of 530 nm. Fluorescence imaging was carried out using an epifluorescence microscope fitted with either a 10×/0.15 HC PL Fluotar or a 40×/0.60 HC PL Fluotar objective.

Complete blood count analysis

The complete blood count test was carried out by collecting whole blood into tubes containing EDTA, which serves as an anticoagulant to preserve the condition of the blood. These samples are then analyzed using an automated hematology analyzer (ProCyte DX hematology analyzer, IDEXX Laboratories Inc., Westbrook, ME, USA). This instrument accurately assesses various blood parameters, including the counts of red and white blood cells, platelet levels, hemoglobin concentration, and hematocrit values, thus offering a thorough evaluation of health.

Serum chemistry analysis

Serum chemistry analyses were performed using the Fuji DRI-Chem NX500i system (FUJIFILM). Serum samples, prepared by centrifuging whole blood at 2000 rpm for 10 min, were used to assess various serum chemistry parameters.

Semithin section and EM

Peripheral nerves were initially preserved in a mixture of 2% paraformaldehyde and 2.5% glutaraldehyde within 0.1 M phosphate buffer, followed by a rinse in the same buffer. For examination under light microscopy, the sections received toluidine blue staining, whereas, for electron microscopy analysis, they were treated with uranyl acetate and lead citrate for contrast enhancement. In addition, the samples were processed with 1% osmium tetroxide, underwent dehydration in ethanol, transitioned via propylene oxide, and lastly encased in epoxy resin suitable for both microscopy methods. The evaluation for pathological changes was performed using semithin slices for light microscopy and ultrathin slices for electron microscopy, using specific stains and a transmission electron microscope (Hitachi, HT7700) set to 100 kV.

Nerve conduction study

NCS was conducted using a Nicolet VikingQuest device (Natus Medical). For the MNCV assessment, stimulation cathodes were situated at two points: the sciatic notch and a location 6 mm distal along the course of the peripheral nerve from the sciatic notch. Recording electrodes were affixed to the belly of the gastrocnemius muscle. Moreover, a ground electrode was positioned at the tail base of each animal. An impartial evaluator, unaware of the treatment groups, measured MNCV and CMAP amplitudes, using supramaximal stimuli. For the SNCV, both the stimulating cathode and the recording electrode were placed on the tail, spaced 30 mm apart, with a ground electrode similarly positioned. SNCV and SNAP amplitudes were determined by the same independent examiner.

Behavioral test

During the adaptation phase, the animals were subjected to training sessions three times per day at a speed of 10 rpm for three consecutive days. Their ability to stay on the rotarod (Panlab, LE8205) was measured by noting the time until they fell at a set speed of 20 rpm, with this process repeated three times during a period of 3 min. The rotarod assessment was carried out thrice for every animal, and the mean of these three values was computed to determine the outcome. To ensure objectivity, two researchers, who did not know the details of the treatments administered, independently evaluated the results of each trial.

Cell viability assays

Live/dead assays were conducted using the LIVE/DEAD Cell Imaging Kit (Invitrogen, R37601). Cells received a treatment involving a 1:1 ratio of 2× working solution and were then allowed to incubate for 20 min at ambient temperature. To visualize the results, fluorescent microscopy images were obtained using ×20 magnification objectives on either a Zeiss LSM 700 or 780 confocal microscope.

H&E staining

The tissues were preserved by immersing them in 4% formaldehyde for a duration of 24 hours at 4°C. After this preservation process, the samples underwent a rinsing process in PBS at ambient temperature. Following the rinse, the samples were subjected to dehydration through ethanol, clarified using xylene, and then encased in paraffin wax. Slices with a thickness of 5 μm were meticulously prepared from the encased tissues and placed onto microscope slides to be ready for staining. For the evaluation of tissue damage and inflammation, these thin slices were treated with H&E stains.

MRI test

A Biospec 7.0-T 30-cm horizontal bore scanner with Paravision 5.1 software, manufactured by Bruker Biospin MRI GmbH in Germany, was used to perform in vivo monitoring of mouse hind limbs and muscle damage. The imaging setup involved a Bruker four-element 1H volume coil array as the receiver and a Bruker 72-mm linear-volume coil as the Tx. To localize the leg, axial, midsagittal, and coronal scout rapid acquisition with fast low-angle shot imaging techniques were used. High-resolution T1-weighted images were acquired in the cross-sectional view using the following parameters: repetition time of 5000 ms, echo time of 32 ms, field of view of 30 mm by 30 mm, matrix size of 250 by 250, and slice thickness of 0.5 mm without any gap.

Supplementary Materials

The PDF file includes:

Supplementary Texts S1 to S4

Figs. S1 to S56

Tables S1 to S3

Legend for movie S1

Legend for data S1

Other Supplementary Material for this manuscript includes the following:

Movie S1

Data S1