PUF-addressed ME multisite stimulator, 2021-2022
Applications
Individually addressed ME stimulation
Yu and colleagues’ IEEE Journal of Solid-State Circuits paper appeared online in December 2021, with a March 2022 issue date. It reports a shared-transmitter, individually programmable stimulation platform with physical unclonable function (PUF) IDs. Primary affiliations include Rice and Baylor.
This is the paper-specific 330-kHz, PUF-addressed configuration. The 2022 endovascular ME-BIT is related hardware, but its larger film, transmitter frequencies and endovascular packaging are not substituted here. The 2025 off-the-shelf spinal network is another configuration, not this ASIC.
Published components
| Part | Specification |
|---|---|
| Implant | 6.2 mm³; 30 mg |
| ASIC | TSMC 180-nm CMOS; 1 × 0.8-mm die, 0.8 mm² |
| ME film | 2 × 3 mm; table gives 0.2-mm thickness; approximately 330-kHz resonance |
| Power storage | Off-chip energy-storage capacitor |
| Individual address | Eight-bit CMOS PUF ID with temporal majority voting |
| Downlink | Amplitude-shift keying; 64 resonance cycles per bit, 5.16 kbps |
| Output | Voltage-controlled monophasic/biphasic pulses; electrodes shorted after each stimulus |
| Programmability | 4-bit amplitude and pulse width; 5-bit delay up to 0.8 ms |
The introductory amplitude range starts at 0.3 V, while functional measurements give 0.25-3.5 V. The introductory pulse-width range is 0.15-1.2 ms. These specifications are not replaced with the different endovascular paper’s payload or pulse-width range.
Bench performance is not implant depth
Two bench implants at 15 and 25 mm from one transmitter were individually programmed without changing the other’s output. Ex-vivo tests used 2-cm porcine tissue, with additional air gap when transmitter separation exceeded 2 cm; reliable operation reached 3.5-cm total separation. That is not 3.5 cm of implanted tissue.
The paper demonstrates up to 40-mm transmitter separation in air. Its 60-mm depth is a COMSOL human-tissue/IEEE-exposure simulation, not an animal implantation result. Film-only characterization gives less than 20% voltage loss below 60° rotation, but complete-device tests at 30-mm separation tolerate 50° in one plane and 40° in the other. The abstract’s 60° must not become an all-plane operating guarantee.
Efficiency and addressing limits
The 90% figure is stimulation-circuit efficiency above 1.5 V, not end-to-end wireless power efficiency. The functional section gives a peak measured transfer efficiency of 1.03% at the coil center under ideal alignment, while comparison-table conditions list other transfer figures. The SoC’s 9-µW idle consumption is separate from stimulation power.
PUF IDs select a device; the accessible paper does not turn an eight-bit address into encryption, authentication or a neural-data uplink. Simulated bit distributions and supply regulation are not a clinical reliability dataset.
Biological evidence and open work
The application separates synchronized Hydra contractions from acute rat sciatic stimulation. Proposed spinal cord stimulation and cardiac pacing appear as target applications, not demonstrated pig/human therapies in this paper. Chronic packaging, long-term tissue response and full model geometry are not established here.
No full model is supplied. Die/film dimensions and total volume do not locate every component, contact, encapsulation boundary or animal-specific stereotrode. The Rice lab brief links the research groups.
Primary source
- Published primary manuscript, JSSC 2021-2022, DOI 10.1109/JSSC.2021.3129993, system/circuit sections, Figures 19-26 and Tables I-II.