Implant-to-implant wireless networking with metamaterial textiles

Abstract

Implanted bioelectronic devices can form distributed networks capable of sensing health conditions and delivering therapy throughout the body. Current clinically-used approaches for wireless communication, however, do not support direct networking between implants because of signal losses from absorption and reflection by the body. As a result, existing examples of such networks rely on an external relay device that needs to be periodically recharged and constitutes a single point of failure. Here, we demonstrate direct implant-to-implant wireless networking at the scale of the human body using metamaterial textiles. The textiles facilitate non-radiative propagation of radio-frequency signals along the surface of the body, passively amplifying the received signal strength by more than three orders of magnitude (>30 dB) compared to without the textile. Using a porcine model, we demonstrate closed-loop control of the heart rate by wirelessly networking a loop recorder and a vagus nerve stimulator at more than 40 cm distance. Our work establishes a wireless technology to directly network body-integrated devices for precise and adaptive bioelectronic therapies.

Fig. 1. Wireless networking of implantable devices with wearable metamaterials.

a Illustration of a wireless implantable medical devices network interconnected by metamaterial textiles. b–d Comparison of wireless implant-to-implant communication systems: b radiative interconnection is limited to transmission distance and implantation depth, c wireless relays cover longer distances but require additional transceivers, d metamaterial textiles enable efficient long-distance and deep-tissue transmission. e Simulated magnetic field distribution ∣H_x∣ generated by a dipole implanted at a 4 cm depth with a metamaterial textile (top) and in the absence of metamaterial textile (bottom) above tissue. f Comparison of the transmission coefficient as a function of distance between the implanted antennas in (e).

Fig. 2. Design and characterization of metamaterial textiles.

a Photograph of the metamaterial textile in a folded state. Scale bar, 1 cm. b–d Structure of metamaterial textiles including a phased surface loaded with passive elements (b), an SSP waveguide (c), and a matching section (d). e Amplitude and phase responses of the five ports labeled in (b) as a function of frequency. f Dispersion curves for SSP structure design with h = 16, 18, 20, and 22 mm, respectively. g Transmission coefficient as a function of h for varying number of matching units (N). h Photograph of a metamaterial textile network integrating two SSP transmission lines (TL) and four phased surface terminals. i Comparisons of full-wave simulations of wireless implant-to-implant transmission in a computation human body model (28.95 dB enhancement with the metamaterial textile). Antennas are placed 4 cm below the skin (black dots). j Transmission coefficient with (top) and without (bottom) metamaterial textiles as a function of the length of the metamaterial textile and the depth of implantation, where dashed lines are contour lines for transmission levels with a spacing of 5 dB.

Fig. 3. Wireless communication performance of metamaterial textiles.

a Photograph of the fabricated metamaterial textiles. Scale bar, 5 cm. b Configuration of implant-to-implant wireless communication. Space below the metamaterial textile is assumed to be homogeneous muscle tissue. L distance between two implant antennas, z depth, θ azimuth angle. c–e Measured transmission spectra for different lengths of the metamaterial textile (c), depths of Bluetooth antennas in water (d), and rotation angles of the receiver antenna (e). f–h Violin plots for comparison of transmission coefficient measured in different configurations with and without the metamaterial textile at air–water interface over 2.4–2.5 GHz ISM band. Box plots inside the violins indicate the quartiles of corresponding transmission spectra. Endpoints show minimum and maximum values; white dots represent median values; whiskers denote 1.5 of the interquartile range. i Bluetooth RSSI recorded in Supplementary Movie 1. j Comparison of RSSI values. Error bars show mean ± s.d. of RSSI values across the indicated period in (i). Source data are provided as a Source Data file.

Fig. 4. In vivo wireless networking and closed-loop control of implantable devices in a pig.

a Three-dimensional computed tomography reconstruction of implants wireless networking in a pig model showing two devices (loop recorder and VNS node) implanted at the thorax and neck, respectively. Scale bar, 5 cm. b Computed tomography cross-section image showing the metamaterial textile, loop recorder, and VNS node. The distance between two implants and depths of implantation from the metamaterial textile are labeled. Scale bar, 5 cm. c Block diagram of the wireless closed-loop sensing (loop recorder) and stimulation (VNS node) system. When the detected heart rate (HR) increases above the threshold hr_t, VNS node will turn on the stimulation until the heart rate recovers. d Recorded ECG waveforms of the loop recorder with R-peak (RR) intervals labeled (left) and output stimulation signals of the VNS node (right). e RSSI received during Bluetooth interconnection of the two implanted nodes. f Comparison of ∣S₂₁∣ (left) and RSSI (right) with and without the metamaterial textile. Error bars show mean ± s.d. of the transmission spectra within 2.4–2.5 GHz (left) and RSSI values in e (right). g Calculated heart rate during two dose injection cycles. Dashed black lines show the heart rate threshold hr_t. h 5-min change in heart rate after it plateaued during two trials as a function of time (error bars indicate mean ± s.d. of the previous 1-min heart rate). i Comparison of ECG signal segments 5 min after the heart rate plateaued. Fewer peaks detected in 5 s indicate lower heart rate. Peaks are marked with squares. Source data are provided as a Source Data file.