Implantable bioelectronic outlet

Abstract

Developing a reliable, long-term electrical interface for implanted bioelectronics is essential for chronic diagnostics, therapeutics, and device maintenance. Here, we introduce the implantable bioelectronic outlet (IBO), a soft, tissue-like electronic interface that enables on-demand, direct ohmic connection between implanted electronics and external devices. IBO is composed of a conducting polymer–coated low-density polymer matrix that is jacketed with a hydrophobic elastomer, allowing repetitive insertion of external electrical contacts without crack propagation. IBO was functional after 1 year of implantation with minimal tissue effects. We demonstrate that IBO enables high-fidelity, bidirectional transmission of signals and power, including low-voltage neurophysiological signals, high-speed digital signals, neurostimulation protocols, and efficient high-current power delivery, validated in small and large animal model studies. The IBO provides a robust and scalable platform for safe, direct, and durable electrical interfacing with implantable bioelectronics.

Implantable bioelectronic outlet is a soft interface for direct electrical connection between implants and external devices.

INTRODUCTION

With the advancement of bioelectronic medicine, there is an increasing need for reliable electrical access to devices implanted in the body (1–3). This access is critical for long-term, longitudinal diagnostics such as assessing functional tissue responses to device implantation and operation (4, 5) and monitoring for tissue health (in the case of transplanted organs or resected tumor zones) (6–8). It can also provide a pathway for maintenance, modification, and charging of implanted electronics without requiring an additional surgical procedure, in contrast to the required repeat surgeries for issues such as battery replacement in many currently used devices (9). Moreover, this electrical interface can directly act as a conduit for delivering electrical stimulation, which is increasingly used therapeutically for neuropsychiatric, cardiac, pain, and inflammatory disorders (10–12).

Establishing this type of electrical connection remains challenging. The status quo for such connections in conventional electronic devices is use of a pluggable connector. However, this is a strongly discouraged approach for implantable devices because a chronic conduit from the external environment into the body predisposes to infection and can lead to local tissue damage. In addition, the materials used in such connectors are rigid, causing substantial mechanical mismatch with surrounding tissue, lack biocompatibility, and cannot be miniaturized. As a result, direct and repetitive physical access to a device inside the body is generally infeasible. To overcome this limitation, most existing approaches circumvent the need for direct electrical contact by external energy conversion at the implant itself. Several strategies have been pursued. Radio frequency (rf) systems, based on electromagnetic (EM) waves, have been used to transfer power and data using coil antennas (13–15). These are mature and readily translatable technologies from existing wireless electronic systems. However, the electronic circuitry required for their operation is bulky and adds complexity to the implant. Furthermore, the reliance on propagation of EM waves, which involves energy loss in the form of heat, substantially restricts implant location. The high ionic conductivity of biological tissue also severely limits rf signal transmission capabilities and demands (16–20). Although mechanical waves in the form of ultrasound wave have better tissue penetration than rf, communication is strongly dependent on the coupling factor between the transmitter and receiver. The power of these waves is limited due to potential tissue damage (13, 21–24). Magnetic field–powered devices could potentially address some of these issues, but due to their low energy-conversion efficacy, they have not been used as a power source for implantable bioelectronics (25). Therefore, there is an immediate need for establishing a reliable electrical contact with implantable devices to simplify the data and power transmission in miniaturized devices.

Here, we present the implantable bioelectronic outlet (IBO), a soft, tissue-like, low-density, conducting polymer–based, multicontact implantable electronic interface that establishes ohmic contact between implanted bioelectronics and the external world. The IBO enables a direct, low–contact-resistance, high-speed, and high-current link for data and power transmission, sufficient for a wide range of biological applications. These include low-voltage electrophysiological signal transmission (~5 to 500 μV, 0 to 10 kHz), high-speed, digitized multichannel neural signal transfer (16 Mb/s), current-based neuromodulation signals, and battery charging (5 V direct current, 500 mA). Using physical processes, we fabricated consistent multilayer IBOs that support multicontact, cross-talk– and leakage-free parallel links. The fabrication processes are scalable, solvent free, and compatible with conformable electronics, eliminating the need for additional layers and facilitating translation to clinical use. We chronically implanted IBOs in rodents for over a year and observed no substantial signs of scarring, inflammation, or immune response around the device, confirming their potential for long-term use. These devices successfully transmitted multichannel, high–sampling-rate neural signals at various stages of the acquisition chain with high fidelity. Last, we integrated IBOs with conformable devices capable of wrapping around the optic nerve to facilitate directional axonal growth and regeneration in large animal models. Collectively, these results demonstrate that IBOs can integrate with a wide variety of implantable devices, with broad potential to enhance the safety and efficacy of next-generation bioelectronic technologies.

RESULTS

We hypothesized that a tissue-like implant with electrically conducting layers could serve as an electrical junction for external interfacing, allowing reliable, repetitive contact with a miniaturized, insertable stylet. This implant could be integrated with existing implantable bioelectronics to provide a direct power and data link to the external world when desired. Furthermore, it could be possible to create a multicontact interface through horizontal or vertical parallel stacks of conducting and barrier materials (Fig. 1A). The key material-level requirements for such a device are as follows: (i) a low-density, porous structure that suppresses crack and tear propagation to ensure a stable and repeatably pierceable layer; (ii) biocompatibility and mechanical softness to mimic surrounding tissue and minimize immune response; (iii) a hydrophobic layer to prevent water retention and eliminate the risk of hydrolysis at >1-V lines; and (iv) high electrical conductivity with a work function similar to that of metals used in medical applications to establish effective ohmic contact (Fig. 1B). To test this hypothesis, we used low-density polymers (LDPs) such as melamine, polyurethane, and cellulose with pore sizes comparable to the diameter of an insertable stylet (150 μm). We then developed a highly conductive poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) formulation with plasticizing additives that enabled homogeneous, bulk coating of the LDPs using a dipping process (Fig. 1C, fig. S1, and Materials and Methods). We chose melamine for our primary LDP, as it is uniquely suited for this application because of the following: (i) Its hydrophilic surfaces aid with facile deposition of conducting polymers such as PEDOT:PSS; (ii) It has very low water permeability ensuring a leak-free scaffolding; (iii) Its open cell matrix allows for uniform coating of conducting polymers and low-force compression during operation; and (iv) It is readily available in various geometries. Cross-sectional scanning electron microscopy (SEM) of the coated LDPs revealed conformal 100- to 200-nm-thick PEDOT:PSS layers along the pore walls without excess buildup (Fig. 1D, left and top right, and fig. S2). To establish electrical insulation across the entire volume of the polymer, it was immersed and repetitively compressed in a low viscosity, diluted polydimethylsiloxane (PDMS) solution and dried to result into a 200- to 400-nm-thick jacket (Fig. 1D, bottom right, and fig. S2). We sandwiched these jacketed conductive layers between isolating layers of the same LDP and bonded them using hydrophobic, stretchable polymers such as PDMS, Ecoflex, or medical-grade silicone to form the IBO. The resulting stacks could be miniaturized using conventional thermocompression processing, yielding well-defined device geometries (Fig. 1E, top, and fig. S3). Given the facile fabrication process of the IBO, in addition to vertical stacking, it was also possible to create lateral patterns of electrically conducting and isolating LDPs compatible with microneedle arrays (Fig. 1E, bottom).

Fig. 1. Structure and material composition of the IBO.

(A) Cross-sectional schematic of a multilayer IBO implanted inside biological tissue and establishing an electrical connection using a multicontact stylet and conformable bioelectronic contact. (B) Schematic of the material architecture and key requirements of the conducting layer, consisting of a low-density porous polymer (LDP) conformally coated with a highly conducting PEDOT:PSS formulation and subsequently jacketed with hydrophobic PDMS to prevent water retention and hydrolysis while maintaining mechanical softness and biocompatibility. (C) Bulk dip coatings of LDP with PEDOT:PSS (dark blue) as conducting material and PDMS (light blue) as insulating jacket material enable homogeneous electrical conductivity throughout the porous network. (D) SEM images of PEDOT:PSS-coated LDPs reveal conformal deposition along pore walls without excess accumulation (left; scale bar, 100 μm). Comparison of PEDOT:PSS-coated (top right; scale bar, 10 μm) and PDMS-jacketed fiber (bottom right; scale bar, 10 μm) of LDP matrix. (E) Optical images of a stack of conducting (blue) and isolating (white) LDPs being compressed (right side) to form a vertical, multicontact IBO (top; scale bar, 2 mm). Optical micrograph of a patterned conducting circular pad in an isolating LDP layer (bottom; scale bar, 1 mm). (F) Lateral patterning of electrically conducting and insulating LDP regions compatible with microneedle-scale access (100-μm stylet shown), enabling scalable fabrication of multicontact interfaces for bioelectronic integration (scale bar, 100 μm).

We characterized crack propagation in these conductive materials and compared them with homogeneous elastomeric films. Stylets with substantially larger (18 gauge, 1.5-mm outer diameter) and comparable (30 gauge, 0.1 mm outer diameter) outer diameters were inserted and retracted. Conductive LDP films exhibited no crack propagation during either insertion or retraction, whereas homogeneous elastomeric films showed tears extending up to twice the stylets’ outer diameter (Fig. 1F and fig. S4, A and B). Even after repetitive insertion, conductive LDP films maintained their integrity, with fibers compressing around the insertion site (fig. S4C). Next, we characterized the insertion force required to penetrate a standard medical needle into the IBO compared to biological tissue alone. The forces in these situations were comparable, and insertion of the needle into the IBO did not result in excess force being applied to biological tissue that was under the IBO (fig. S5). We also performed detailed mechanical characterization to determine the contribution of each material to the overall Young’s modulus of the IBO. Layer-by-layer compressive force characterization revealed that PDMS contributes the most to the overall required force, and therefore increasing PDMS thickness substantially increased the compressive force (fig. S6, A to C). A similar pattern was observed for tensile strain, with an overall higher required force than compression, attributable to the porous structure of the LDPs (fig. S6D). We also characterized the effect of PEDOT:PSS thickness on the electrical and mechanical properties of the IBO. A single dip coating of PEDOT:PSS resulted in optimal mechanical and electrical properties. SEM images following repeated PEDOT:PSS coating showed highly nonuniform filling of the LDP pores. This excess PEDOT:PSS led to greater stiffness of the IBO without substantial improvement in electrical performance (fig. S7). Overall, mechanical properties conducive to creation of IBO can be realized using existing biocompatible polymers in a scalable fashion.

Next, we moved on to characterize the electrical and electrochemical properties of the materials to investigate IBO data, power, and scaling capacity. Given the complex geometry of the LDPs, it was challenging to directly derive the electrical conductivity of the layer. Therefore, we report the electrical and electrochemical impedance of the film (10 mm by 10 mm; 1 mm thick) using gold (Au)–coated contact probes. This also allowed us to characterize the contact resistance of the IBO with metallic contacts (Fig. 2, A to C). For electrical impedance, because the IBO is in series with implanted electrophysiological sensors, the key consideration is that the total impedance of the IBO and its constituent contacts should be negligible compared to the input impedance of the electrophysiological sensors. This configuration ensures minimal losses at the interface for both recording and stimulation. Overall, the conducting LDP impedance was consistently <100 Ω, which is approximately 10³ to 10⁴ times less than a typical electrophysiological electrode (26–30), making IBO suitable for direct interfacing with such sensors, as demonstrated by our ability to interface with an intracranial microelectrode array (Fig. 2A). In addition, the LDP impedance was highly consistent due to the reproducible and repeatable fabrication process (Fig. 2B). This narrow impedance distribution is important for eliminating signal hazard for high-speed digital communication protocols such as low-voltage differential signaling used in integrated circuits (ICs) for acquisition of electrophysiological signals. Furthermore, the electrical impedance value of the material remained constant from DC up to 1-MHz frequencies, indicating a low parasitic capacitance that could be attributed to poor carrier density and mobility (Fig. 2A). Electrochemical impedance spectroscopy (EIS) of the same conducting LDP revealed an ultralow electrochemical impedance (1 to 50 Ω) that was relatively constant across the spectrum (Fig. 2D). These results indicate that the PEDOT:PSS was able to coat the bulk of the polymer in an electrically continuous manner, maintaining its high porosity and creating an extensive effective surface area for charge transport (figs. S7 and S8). This low EIS performance remained stable even after mechanical compression of the LDP, indicating that the porous network and PEDOT:PSS coating maintained continuous electrical pathways under strain (Fig. 2E). In addition to efficient electronic conduction, the elastomer-based jacket of conducting LDP exhibited robust isolation and water impermeability required for multichannel communication and power transfer in electrophysiological signal acquisition (Fig. 2D, red trace). This feature is particularly critical for applications such as neuromodulation and battery charging where applied potentials can exceed the hydrolysis potential. Furthermore, continuous chronoamperometry at high voltages over an extended period of time revealed no substantial change, establishing the robust hydrophobic barrier properties of IBO and its applicability for power applications (Fig. 2F). Overall, conducting LDPs demonstrated low electrical impedance across a wide bandwidth with high reproducibility, while their elastomer jackets provided a robust barrier facilitating high voltage operation.

Fig. 2. Electrical and electrochemical characterization of conducting and isolating LDPs.

(A) Two-probe impedance spectroscopy of 1-cm² conducting LDP demonstrates low and constant ohmic impedance. (B) Narrow distribution of the conducting LDP resistance at different locations (n = 100 contacts). (C) Low contact resistance (2 to 5 Ω) of conducting LDP with Au-coated contacts and the corresponding impedance scaling. (D) Comparison of EIS of the conducting LDP (black trace) and PDMS-jacketed conducting LDP (red trace) in phosphate-buffered saline (PBS) electrolyte. (E) Comparison of EIS of conducting LDP (black trace) with compressed LDP in PBS revealing no changes in the impedance. (F) Chronoamperometry of PDMS-jacketed conducting LDP film in PBS under +5-V direct potential without any sign of hydrolysis, demonstrating the conformal jacketing of LDP.

After establishing the materials and their mechanical and electrical performance, we moved to fabricate implantable devices. To investigate all potential geometrical possibilities, we used a two-pronged approach: (i) horizontal interface in which the conducting LDP–based contacts are in plane; (ii) vertical interface where the conducting LDPs are sandwiched between barrier layers that serve as electronic insulators and space the contacts. To effectively integrate IBO with implantable devices, we developed a bonding approach allowing for direct ohmic contact between IBO and conformable substrates without the need of any rigid interlayer (Fig. 3, A to C). The IBO conducting layers were aligned and gently pressed (0.2 to 0.8 N) with the mating conformable pads. This contact force did not result in any further compression of LDP. Each conformable contact contains a perforation (500-μm diameter) that serves as a conduit to introduce adhesive elastomer from the backside of the conformable substrate into the interface (Materials and Methods). We used 1-mm-thick LDP layers for these devices to facilitate ease of stylet insertion without requiring direct visual guidance. Even at smaller thicknesses (such as 250 μm; Figs. 1 and 2), the electrical connection remained nearly identical to that of a 1-mm-thick LDP, demonstrating that stable electrical contact is preserved across a wide range of thicknesses. Therefore, the minimum contact area required for stable electrical transmission is substantially smaller than the device dimensions required for the manual alignment of the stylet electrode. Consequently, the thickness of IBO used in the devices was primarily governed by ease of manual access, as electrical contact was stable across the relevant thicknesses.

Fig. 3. IBO can directly integrate with conformable electronics and establishes a highly biocompatible chronic interface.

(A) Fabrication process of direct bonding of conformable electronics to IBO using a perforated pad for delivery of adhesive. (B) Optical micrograph of conformable neurostimulating array with IBO as interface pad (scale bar, 2 mm). (C) Photograph of horizontal (left) and vertical (right) IBOs bonded to a conformable substrate to establish multichannel devices without the use of rigid interlayers (scale bars, 2 mm). (D) EIS comparing conformable electrodes with and without the IBO interface shows no substantial change in impedance. (E) Comparison of aligned (black) and misaligned (red) contact impedances demonstrating a cross-talk free multichannel contact. (F) Micrograph of an IBO implanted for 1 year in subcutaneous space with no gross tissue abnormality (top; scale bar, 2 mm). Extracted IBO after 1 year of implantation surrounded by healthy vascularized tissue (bottom; scale bar, 1 mm). (G) Postexplant electrical characterization of aligned (blue) and cross-talk–free misaligned (red) contact impedances for each of the six channels in an IBO device, measured after 1 year of subcutaneous implantation in a rodent.

We characterized the quality of this connection by comparing the EIS of the conformable electrode with and without the IBO interface, finding no substantial change in the impedance (Fig. 3D). We investigated the interchannel cross-talk using two-port impedance spectroscopy between two adjacent contacts, while the entire device was immersed in phosphate-buffered saline (PBS), finding that the impedance was consistently high across the channels, consistent with lack of cross-talk (Fig. 3E and fig. S9). Furthermore, we demonstrated robustness of these characteristics to repeated penetrations by comparing the impedance of within versus across layers over 100 insertions (fig. S10A). Connection with the IBO was also able to withstand substantive bending (fig. S10B). We implanted these IBOs in healthy mice using both incision- and injection-based approaches (Materials and Methods). These devices remained implanted for over a 1-year period (n = 6 mice). Subsequently, they were extracted under anesthesia for evaluation of the implant and surrounding tissue intraoperatively. Gross histological analysis of extracted tissue at the implant site and a control tissue site in the same animal was performed (Materials and Methods). The cutaneous tissue above the implant was grossly normal, without scarring or other observable abnormalities, and the implant’s dark color was visible under high intensity light (Fig. 3F, top). Incision of the tissue in the vicinity of the implant similarly revealed a lack of scarring and even the presence of vascularization (Fig. 3F). Hematoxylin and eosin (H&E) staining of tissue adjacent to the implant compared with control tissue extracted from the contralateral area revealed similar tissue architecture and composition (fig. S11). Furthermore, postmortem electrical characterization of the IBO devices implanted in these mice exhibited similar electrical conductivity, barrier properties, and cross-talk compared to preimplantation (Fig. 3G and fig. S12). Overall, IBO demonstrates high-biocompatibility and long-term stability, with the capacity to establish low-resistance, cross-talk–free ohmic contact and facile integration with conformable devices.

Given these results, we created devices that harness the IBO functionality to address challenges of physiological and electronic signal transmission from implantable devices (Fig. 4A and fig. S13). First, we used IBO as an electrical outlet to charge implantable batteries for neural interface devices. The charging duration and the charging current were comparable to that observed with direct ohmic contact, with no temperature increase under 100 mA of current (Fig. 4B and fig. S14). Next, we interfaced IBO with neural probes composed of passive electrodes. In this configuration, IBO transmits the unbuffered signal acquired by intracranial electrodes to analog front-end amplifiers outside the body. The external connection was made using a Au-coated 30-gauge needle or an Au-based microfabricated array laminated onto needles using adhesive (Fig. 4A). After insertion, the connected wire was loosely taped to the skin, providing stable contact despite movement. We were able to acquire neural data in acute and chronic settings in both unanesthetized rats and developing mice [postnatal day 5 (P5) to P20] with comparable signal quality to conventional electronic rigid connectors (Fig. 4C) (31).

Fig. 4. IBO establishes a reliable data transmission pathway for low-voltage analog and high-speed digital signals and enables power delivery and neurostimulation in both small and large animal models.

(A) Optical micrograph of a multicontact probe conformed around a 18-gauge needle (scale bar, 2 mm). (B) Comparison of charging curve for a lithium-ion battery using an IBO in horizontal contact mode (black) and conventional two pole battery connector (red). (C) Superimposed time-frequency spectrogram and its corresponding electrophysiological recording from a developing mouse (P19) using an IBO directly attached to the neural probe for signal acquisition and transmission. (D) High-speed digitized electrophysiological signal transmission using an IBO attached to a neural-chip output (six lines, 32 channels, 20-kHz sampling rate). Signals were acquired from a 32-channel implantable probe in medial prefrontal cortex of an adult rat (scale bar, 200 ms). Blue trace indicates a delta wave, orange is a cortical ripple, and green is a thalamocortical sleep spindle. Inset shows superimposed and average waveform of action potentials transmitted using IBO (scale bar, 1 ms). (E) Time-frequency spectrogram of neural data transmitted by IBO demonstrating transition through waking and sleep states over the course of a 30-min recording session. (F) Intraoperative image of a three-electrode IBO being inserted into an incision in pigs (left; scale bar, 1 mm). Image of the retracted incision after insertion of IBO and stylet penetration into the IBO (right; scale bar, 1 mm). (G) Superimposed individual (black traces) and average (red trace) voltage measurements during 20-μA pulse stimulation delivered by an implantable pulse generator through IBO to an optic nerve stimulator (electrode size = 500 μm by 6000 μm; n = 50 stimulation waveform).

Next, we integrated IBO into the back end of a neural acquisition IC for digital communication. The current state-of-the-art neural ICs are able to perform 32-channel block time-division multiplexing at 20,000 sample/s per channel. These data samples are quantized into 12-bit digital bits that are streamed out along with five additional digital lines using a serial peripheral interface protocol, resulting in digital data rates close to 16 Mb/s. At such high speeds, consistent, low-impedance data lines are critical for successful communication because small impedance differences result in sufficient phase shift to cause data corruption. IBO was able to successfully establish and maintain this high-speed communication over an extended period of time, allowing for clear identification of rodent sleep states, action potentials, and sleep oscillatory patterns using automatic algorithms (Fig. 4, D and E, and fig. S15).

Last, we used IBO in implantable devices requiring electrical stimulation. Given the translational capacity of the IBO and therapeutic potential of electrical stimulation, we deployed IBO in nerve stimulation and regeneration protocols in the pig large-animal model (Fig. 4F). It has been hypothesized that delivery of defined pulses of electrical stimulation across damaged areas of the optic nerve can aid with directional regeneration of axons (32, 33). Therefore, we used parylene C (Pa-C)–based processes to microfabricate a conducting polymer–based conformable nerve stimulating probe that contained multiple pairs of stimulating electrodes to be placed along the length of the optic nerve (fig. S16, A and B). The conformable probe was bonded using the previously described process (Fig. 3A). To benchmark and validate the performance of the IBO, we compared intraoperative performance with similar probes that were bonded to a printed circuit board and directly connected to the electrical stimulator unit using metallic wires (Materials and Methods). We delivered constant current pulses with 20-μA amplitude and 100-μs duration through both the IBO and conventional wired connection. We measured the resultant voltage across the electrode pairs, which represents the overall impedance of the systems. We observed no voltage increase when IBO devices were used compared to the conventional approach (Fig. 4G and fig. S16C). Overall, these experiments demonstrate the wide applicability of IBO for implantable bioelectronics in areas of electrophysiological data acquisition, circuit modulation, and maintenance of devices.

DISCUSSION

In this study, we introduce the IBO, a tissue-like, soft, low density, multicontact interface designed to provide a long-term, stable, low-resistance, and biocompatible electrical junction between implanted bioelectronics and the external world. The porous, low-density architecture suppresses crack propagation and prevents tearing across repeated access sessions, overcoming a longstanding failure mode of homogeneous elastomeric interfaces. This unique mechanical resilience allowed stable, repeatable connections without damaging surrounding tissue.

The IBO addresses key challenges of signal and power transfer for implantable bioelectronics. Conventional strategies rely on energy conversion or exposed transcutaneous ports that present fundamental limits in power delivery, require bulky supporting circuitry, and incur risks associated with heat generation or tissue damage. Our findings demonstrate that a synthesis of tissue-like interface based on conductive LDPs can overcome many of these constraints, enabling high-bandwidth, high-current, and multichannel signal transmission through an ohmic, pierceable, and chronically stable contact. The lateral patterning of IBO is also amenable to use with multiple concurrent penetrating contacts, allowing integration with microneedle-based stylet systems for minimally invasive, large-scale parallel connectivity. The continuous volumetric coverage of the conducting polymer inside the porous network of LDP ensures reliable contact throughout the film. Similarly, the volumetric coating of the device with hydrophobic elastomer produces a robust barrier capable of withstanding voltages exceeding the hydrolysis threshold, allowing the IBO to safely support high-power operations such as neuromodulation and battery charging. Chronically implanted IBOs were functional and biologically well tolerated after 1 year of implantation, underscoring this advantage.

By integrating the IBO with diverse classes of implantable bioelectronics, we demonstrate the broad applicability of the platform. The IBO supported low-voltage analog neural recordings with fidelity comparable to conventional metallic connectors, even in unbuffered acquisition modes where signal integrity is particularly sensitive to impedance. It also enabled digital data streaming at 16 Mb/s from a high-density neural acquisition IC. Beyond data transfer, the IBO provides a reliable conduit for electrical stimulation and power delivery. The ability to charge implantable batteries through a soft tissue interface without requiring rigid percutaneous connectors offers a promising approach for maintaining chronically implanted devices. Likewise, the IBO’s compatibility with neuromodulation was validated in both rodent brain stimulation experiments and large-animal nerve regeneration studies. From a translational perspective, IBO is best suited for applications that require long-term implants to establish intermittent contact with external electronics. These applications could include: i) fast charging of an implantable battery without the need for coils or complex circuitry, ii) high-speed download of locally stored data with minimal power consumption, or iii) updating and maintenance of operating software and stimulation protocols. Future work will focus on establishing flexible and miniaturized stylets for longer term connections and integration with hydrogel coatings of analgesics or anti-inflammatory compounds to optimize parameters for chronic use (34, 35). Overall, IBO has the potential to substantially advance the capabilities and safety of next-generation active bioelectronics, enhancing our ability to harness these devices to benefit health and treat disease.

MATERIALS AND METHODS

Materials

PEDOT:PSS (Clevios PH 1000) was acquired from Heraeus. Ethylene glycol, 4-dodecyl benzene sulfonic acid (DBSA), (3-glycidyloxypropyl)trimethoxysilane (GOPS), hexane, and PBS tablets were purchased from Sigma-Aldrich. PDMS (SYLGARD 184) was acquired from Dow Corning. Commercially available melamine foam sheets (Procter & Gamble, Cincinnati, OH, USA) were used as the LDP material. Polytetrafluoroethylene (PTFE)–coated stainless steel wire (100-μm diameter) was obtained from Wytech. Pa-C dimer, 3-(trimethoxysilyl)propyl methacrylate (A-174 silane), and Micro-90 alkaline cleaning solution were acquired from Specialty Coating Systems for chemical vapor deposition. Shipley 1827 (positive photoresist) and MF-319 developer were acquired from Microposit. Gold etchant–type TFA and chrome etchant 1020AC were acquired from Transene. AZ 10XT (positive photoresist) and AZ 400K developer were acquired from MicroChemicals. 24K Brush Gold Solution for Au electroplating was purchased from Gold Plating Services. Light-Activated Superglue was acquired from J-B Weld and used for acrylic adhesive.

Preparation of conducting and barrier layers

High-conductivity PEDOT:PSS was prepared by mixing 80% PEDOT:PSS, 20% ethylene glycol, and 0.6% DBSA (all v/v). To fabricate the conducting layer, 1-mm-thick LDP was fully immersed in the prepared high-conductivity PEDOT:PSS solution. The LDP was gently compressed to remove excessive solution and dried overnight at room temperature in a well-ventilated environment to yield conductive LDP. The immersion and drying process described above was repeated multiple times for experiments with thick PEDOT:PSS coating. For the barrier layer, a PDMS (SYLGARD 184) solution was prepared by mixing the base and curing agent at a 10:1 ratio, followed by dilution with hexane [1:1 (w/w)]. LDP was dip coated in the diluted PDMS solution and cured in an oven at 110°C for 1 hour. To increase the thickness of the barrier layer, the hexane dilution ratio was reduced [2:1 (w/w)].

Fabrication of IBO

Device components were patterned into standardized cylindrical geometries using biopsy punches or cut into custom shapes using a razor blade. To fabricate the IBO, the patterned conductive LDPs and pristine LDPs were dipped in the diluted PDMS solution, serving the dual purpose of electrical insulation and interlayer bonding. The LDPs were assembled, aligned, stacked, and subjected to mechanical compression using a heat press to define the final device dimensions. The assembly was cured at 110°C for 1 hour to cross-link the PDMS, bonding the layers and fixing the compressed geometry.

Fabrication of conformable thin-film electrodes

The fabrication process for conformable thin-film electrodes was adapted from previously established protocols (36–38). For substrate preparation, a 2-μm-thick layer of Pa-C was deposited on silicon wafers (100-mm diameter; 500-μm thickness; single side polish) via chemical vapor deposition (SCS Labcoter 3) to serve as the flexible substrate and base insulation layer. For metal interconnects, a 15-nm chromium (Cr) adhesion layer followed by a 150-nm Au layer were deposited using electron-beam evaporation (Angstrom EvoVac deposition system). Metal interconnects were patterned using wet etching. A positive photoresist (Shipley 1827) was spin coated at 3000 rpm onto the metal layers and soft baked at 110°C for 1 min. The resist was exposed to ultraviolet (UV) light using a mask aligner (Karl Suss MA6) and developed in MF-319 developer for 1 min, followed by a deionized (DI) water rinse. The exposed Au and Cr layers were selectively removed using Au and Cr etchants, respectively. Last, the remaining photoresist was stripped with acetone. To insulate the interconnects, a second 2-μm-thick Pa-C layer was deposited. An adhesion promoter (A-174 silane) was introduced during the chemical vapor deposition process to ensure robust bonding between the base and encapsulation Pa-C layers. Subsequently, an antiadhesion layer (5 wt % Micro-90 diluted with DI water) was spin coated at 1500 rpm, followed by the deposition of a third 2-μm-thick Pa-C layer, which served as a sacrificial mask for the PEDOT:PSS patterning. A thick positive photoresist (AZ 10XT) was spin coated (300 rpm for 10 s and then 5000 rpm for 30 s), baked at 110°C for 1 min, and patterned to define the electrode openings and device outline as an etch mask. Resist was exposed to UV light using a mask aligner (Karl Suss MA6) and developed with two separate AZ 400K developer (1:4 diluted with DI water) for 3.5 min each. The exposed Pa-C layers were dry etched using oxygen plasma reactive ion etching (Oxford PlasmaPro 80; 180 W, 50 standard cubic centimeters per minute of O₂, 60 mtorr) for 15 min to expose the underlying Au contacts and define the device geometry. The previously prepared high-conductivity PEDOT:PSS solution was mixed with 1 wt % GOPS as a cross-linker, and the mixture was spin coated over the entire wafer at 650 rpm for 30 s. The film was then cured at 110°C for 10 min, and this coating and curing process was performed twice. The sacrificial Pa-C layer was mechanically peeled off, removing the excess PEDOT:PSS and leaving the conductive polymer exclusively in the predefined electrode openings. Last, the fully fabricated devices were released from the silicon carrier wafer by applying a small amount of DI water to facilitate mechanical peeling.

IBO integration with wires

To facilitate low–contact-resistance interfacing between standard wiring and the PEDOT:PSS-coated LDP in IBO, the tips of the wires were electrochemically coated with Au. The wire tip was electroplated with Au at 4 V using a Au electroplating solution and a platinum (Pt) anode. The Au-plated tip was then mechanically inserted into the porous network of the conductive LDP. To mechanically secure the connection and provide electrical insulation, the entire assembly was dip coated in PDMS and cured.

IBO integration with single-channel conformable thin-film electrodes

For the integration of the IBO with single-channel conformable thin-film electrodes, a mechanical anchoring strategy using PDMS was implemented. A through hole was created in the center of the thin-film electrode’s connection pad. The electrode was laminated onto a layer of uncured PDMS, allowing the PDMS to protrude through the hole and become exposed on the upper surface. A pristine LDP, cut to match the dimensions of the electrode pad, was placed directly over the hole. This allowed the uncured PDMS to infiltrate the porous LDP structure, creating a mechanical interlock upon curing. Subsequently, the anchored LDP was dip coated in the high-conductivity PEDOT:PSS solution. This process simultaneously rendered the LDP conductive and established a robust electrical connection with the underlying Au electrode pad. Last, the integrated device was dip coated in PDMS to encapsulate the structure and ensure mechanical stability.

IBO integration with multichannel conformable thin-film electrodes

For multichannel integration, the IBO was fabricated with a conducting layer patterned to match the geometry of the thin-film electrode connection pads. To create an exposed conductive interface, the IBO was cut to reveal the cross section of the PEDOT:PSS-coated LDP without external insulation. This exposed conductive face was aligned and brought into direct contact with the corresponding pads of the conformable thin-film electrode. To maintain alignment and seal the connection, the interface area was dip coated in PDMS.

Fabrication of stylet electrodes for IBO access

Various stylet electrode configurations were fabricated to establish low–contact-resistance interfaces with the PEDOT:PSS layer of the IBO. For standard access, medical needles were coated with a 15-nm Cr adhesion layer and a 150-nm Au layer using a sputter deposition system (Denton DV-502M). To provide electrical insulation, the needle shafts, excluding the conductive tips, were dip coated in a thin layer of light-curable acrylic adhesive. When a smaller diameter needle was required, 100-μm-diameter PTFE-coated stainless steel wires were used, and their tips were exposed by mechanically removing the coating with a razor blade. For interfacing with laterally patterned IBOs, a microneedle array was assembled using Au-coated 30-gauge needles fixed in a custom mold, which was fabricated using a resin three-dimensional printer (Form 3+, Formlabs) to match the specific IBO geometry. To access vertically stacked IBO layers, a conformable thin-film electrode array patterned to match the specific IBO geometry was laminated onto the surface of an 18-gauge needle using the acrylic adhesive.

Electrical and electrochemical characterization

Electrical measurements were performed using a potentiostat/galvanostat/impedance analyzer instrument (PalmSens4). To evaluate the electrical properties of the material, 1 cm–by–1 cm LDP samples were prepared with a pristine single-layer thickness of 1 mm. EIS was conducted in a PBS solution using a standard three-electrode configuration. A Pt wire served as the counter electrode, and a silver/silver chloride (Ag/AgCl) electrode was used as the reference. Impedance spectra were acquired over a frequency range of 5 Hz to 2 MHz with a sinusoidal excitation amplitude of 0.1 V. For contact resistance characterization, a Au-coated pin connector (2.54-mm pitch) was used, and the resistance values were recorded using a digital multimeter (Fluke). Electrical access to the PDMS-jacketed LDP was achieved by inserting a stylet electrode. This insertion caused localized fracturing of the LDP fibers, exposing the PEDOT:PSS-coated cross sections to establish a direct electrical interface. Because this fiber disruption was confined to the contact point and remained embedded within the bulk PDMS-jacket LDP, the LDP maintained its insulation and structural integrity even after the electrical access.

Rodent surgical procedure

All animal experiments were approved by the Institutional Animal Care and Use Committee at University of California Irvine (UCI; #AUP-23-121). Swiss-Webster mouse pups of both sexes and adult male Long-Evans rats were used for in vivo validation. Animals were kept on a regular 12-hour light/12-hour dark cycle with ad libitum access to food and water. Rats were housed in pairs before implantation but separated afterward. Mouse pups were housed with mother and littermates until weaning at P21. Prior experimentation was not performed on these animals, and experimentation was performed during light-on periods.

Mouse implantation

Surgical procedure in mouse pups followed established protocols (31, 39, 40). Anesthesia was induced with 3.0% isoflurane and maintained with 1.0 to 1.5% isoflurane. Postoperative recovery was expedited by minimizing time under anesthesia. For each injection-based IBO implantation, a microchip syringe injector was used to directly place the IBO in the subcutaneous site. For each incision-based IBO implantation, hair was removed 5 mm caudal to nuchal areas before a small skin incision. Round-tip forceps were used to bluntly dissect a small subcutaneous space in which an IBO was inserted. After placement, the incision around the IBO access area was closed with sutures, followed by application of antibiotic ointment. A subset of mouse pups received intracranial implantation of flexible recording and reference electrodes assembled to IBOs. Scalp incisions and burr-hole craniotomies were done for insertion of the reference electrode in the cerebellum (1.0 mm posterior to lambda, 0.5 mm lateral to midline, 0.5 mm ventral to surface; proportional to age) and insertion of the recording electrode in the primary somatosensory cortex (2.0 mm anterior to lambda, 1.0 mm lateral to midline, 0.5 mm in depth; proportional to age). Stylets were used to advance the flexible electrodes to exact depths and then retracted out of the brain tissue and removed. Twenty-gauge needles were used as the tunnel and acupuncture needles were used as the shuttle to subcutaneously guide the flexible device from IBO sites to electrode sites. Postimplantation, craniotomies were covered with tissue foam and incisions closed with veterinary glue. Acute and chronic recordings were done in an incubation chamber with temperature, humidity, and olfaction control for maximized comfort.

Rat implantation

Surgical procedure in adult rats followed established protocols (31, 41–43). Anesthesia was induced with 2.0% isoflurane and maintained with 0.75 to 1.0% isoflurane. Craniotomy was drilled, and dura was removed over the medial prefrontal cortex (3.5 mm anterior to bregma, 0.8 mm lateral to midline, 2.4 mm ventral to surface) to allow chronic implantation of either flexible electrodes or NeuroNexus silicon probes. Two stainless steel screws were used as reference at the cerebellum. Craniotomies were covered with tissue foam and sealed with a silicone elastomer. For each multilayer IBO implantation, transdermal incision and blunt dissection were done 2 cm caudal to nuchal areas for IBO placement, and subcutaneous tunneling of multisite flexible electrodes was accomplished with an 18-gauge needle. Stylets were used to advance the flexible electrodes to depth. Incisions were sutured with absorbable silk sutures.

Pig implantation

All procedures were performed in accordance with institutional biosafety and ethical guidelines of UCI. The skin and subcutaneous tissue surrounding the orbital cavity were carefully incised to expose the optic nerve. A conformable electrode was gently wrapped around the exposed optic nerve, ensuring intimate contact without compressing the nerve bundle. The electrode wirings were connected to an IBO placed beneath the skin behind the ear. For electrical stimulation, an external stylet electrode connected to a programmable stimulator was inserted percutaneously into the IBO, establishing an ohmic connection with the implanted electrode. Controlled current pulses were delivered through the IBO to the conformable optic nerve electrode to evoke stimulation responses.

Neurophysiological signal acquisition with IBO and preprocessing

Stylet electrodes for IBO access were used to penetrate the insulating layer to form electrical connections between intracranial electrodes and signal acquisition systems (RHD2000, Intan Technology). The recordings were performed after the animal recovered from surgery in the absence of anesthesia. Signals were amplified and digitized at 20 kHz with 16-bit storage format. Recordings were first visually inspected for quality using NeuroScope. All analysis was done using MATLAB (MathWorks). Local field potentials (LFPs) were obtained by resampling raw recordings at 1250 Hz, and neural spiking activity was obtained by high-pass filtering raw recordings at 250 Hz. Time-frequency spectrograms were obtained in LFPs using a Gabor wavelet. Delta waves, thalamocortical spindles, and cortical ripples in LFPs, in addition to extracellular waveforms of action potentials, were determined on the basis of established filter-threshold approaches (44, 45).

Histology

Intraperitoneal injection of pentobarbital (100 mg/kg) was used to euthanize rodents. Tissue sample surrounding IBO implantation site was dissected and extracted to be fixated in 4% paraformaldehyde for 24 hours before transferring to saline solutions. Tissue sample was then transferred to UCI Histology Core for H&E staining.

Supplementary Materials

This PDF file includes:

Figs. S1 to S16