Adaptive RF Wave Focusing for Efficient Power Delivery to Deep Neuroelectronic Devices
DOI: Pending DOI
Synopsis
This work addresses a key bottleneck in neuromodulation: powering implantable devices at depth without batteries. We develop wave-based strategies to efficiently deliver energy through complex, lossy biological tissues to miniature neuroelectronic implants. By shaping electromagnetic fields and leveraging the physics of wave propagation in heterogeneous media, we enable deeper, more reliable, and minimally invasive powering schemes. This approach aims to expand the design space of neuromodulation systems, supporting distributed, long-term implants and opening new opportunities for clinical and wearable neurotechnologies.
Background
Neuromodulation is rapidly evolving toward distributed, minimally invasive systems capable of interfacing with the nervous system at multiple spatial scales. However, a central limitation remains the reliable powering of implantable devices, particularly when targeting deep anatomical regions. Conventional approaches based on batteries constrain device size, lifetime, and biocompatibility, while wired solutions limit clinical translation and patient comfort. Wireless power transfer has emerged as a promising alternative, yet its efficiency is fundamentally challenged by the heterogeneous, lossy, and dispersive nature of biological tissues, which strongly attenuate and distort propagating fields [1]. Recent advances in bioelectronics call for new modalities that go beyond the standard inductive coupling strategies that impose limitations on the device miniaturization and implantation depth. In this context, controlling wave propagation within complex media offers a powerful framework to overcome these limitations. By tailoring the spatial and temporal structure of electromagnetic fields, it becomes possible to enhance energy delivery to deeply embedded, millimetric devices while minimizing off-target exposure. This work builds on these concepts to explore how wave-based control can unlock new capabilities for powering next-generation neurotechnologies, aligned with the needs of the neuromodulation community.
Methods
We investigate battery-free operation based on backscatter communication, where the implant both harvests energy and modulates the reflected signal. This bidirectional link imposes stringent efficiency constraints, as propagation losses are incurred twice. To address this, we develop wave-based optimization strategies for energy harvesting in complex biological media, leveraging metasurface-inspired and wavefront-shaping approaches to maximize local field intensity at the device location while maintaining safe exposure levels.
In parallel, we implement adaptive mid-field focusing techniques to localize and track the implant. By controlling the phase and amplitude of multi-channel transmitters, we iteratively converge on the optimal wavefront that maximizes the backscattered response, enabling simultaneous localization and efficient power delivery. These methods are validated through full-wave numerical simulations (using COMSOL Multiphysics), as quasi-static approximation cannot be appuies above ca. 10 MHz frequencies [2-3]. Experiments in heterogeneous phantoms replicating tissue electromagnetic properties. A programmable multi-antenna platform ensures precise spatiotemporal control for reproducible focusing conditions.
Results
Numerical results first quantify the gains achievable at the receiver side using physics-based design strategies. By jointly optimizing operating frequency, modal regime, and current distribution within the implant, we demonstrate an approx. 3 times improvement in harvested power compared to conventional miniature antennas for wireless implantable devices [4]. The proposed approach enables predictive design: for a given device size and implantation depth, the framework identifies the optimal configuration and associated performance limits. Prototypes following these guidelines show 3-4 times gains over state-of-the-art implantable devices of comparable dimensions, confirming the relevance of co-designing the electromagnetic response with the propagation environment [5].
In parallel, wavefront shaping in the mid-field regime reveals substantially larger system-level gains. Adaptive focusing based on multi-channel control and backscatter feedback enables constructive interference at the implant location while reducing off-target exposure. For a representative device (2 mm size, largest dimension) located 5 cm deep in tissue, simulations predict power delivery improvements of up to four orders of magnitude compared to conventional approaches that illuminate the medium without spatial optimization [6]. In terms or recovered power, we show that it is possible to recover up to 10 mW of power for this scenario. Gains scale with device size and channel count, while remaining constrained by safety limits and tissue losses. These results highlight that the proposed methods are particularly advantageous for deep, millimetric, battery-free implants, whereas shallow or larger devices remain well served by inductive or optical (near-infrared) techniques. Numerical predictions are supported by ongoing experimental validation in tissue-mimicking phantoms, with results to be presented at the conference.
Discussion
These results position wave-based adaptive midfield power transfer as a viable route to overcome current depth–size trade-offs in neuromodulation. By explicitly accounting for the propagation physics in complex tissues, the proposed framework shifts the design approach from device-centric to system-level optimization. This is particularly relevant for emerging applications requiring distributed, minimally invasive, and chronically implantable nodes [7]. At the same time, practical constraints remain. Performance depends on accurate knowledge or estimation of the propagation environment, and real-time adaptive focusing introduces system complexity. Safety, robustness to motion, and integration with clinical workflows will be critical for translation. Ongoing work aims to refine these aspects while engaging with the neuromodulation community to align capabilities with unmet clinical needs.
Disclosures
This study was supported by the European Research Council (ERC) under the EU Horizon Europe programme (Grant Agreement No. 101165786, Project “BESSEL”
References
[1] Soares, I. V. et al. Wireless powering for long-lasting deep-body bioelectronic devices. IEEE Antennas Propag. Mag. 67, 74–86 (2025).
[2] Gaugain, G. et al. Quasi-static approximation error of electric field analysis for transcranial current stimulation. J. Neural Eng. 20, 016027 (2023).
[3] Wang, B. et al. Quasistatic approximation in neuromodulation. J. Neural Eng. 21, 041002 (2024).
[4] Gao, M., Nikolayev, D., Sipus, Z. & Skrivervik, A. K. Physical insights and design principles for efficient wireless implantable bioelectronics. Cell Rep. Phys. Sci. 102627 (2025)
[5] Nikolayev, D., Joseph, W., Zhadobov, M., Sauleau, R. & Martens, L. Optimal radiation of body-implanted capsules. Phys. Rev. Lett. 122, 108101 (2019).
[6] Soares, I. V. et al. Wireless powering efficiency of deep-body implantable devices. IEEE Trans. Microw. Theory Techn. 71, 2680–2692 (2023).
[7] Tian, X. et al. Implant-to-implant wireless networking with metamaterial textiles. Nat. Commun. 14, 4335 (2023).