DustNet: A Wireless Network of Ultrasonic Neural Implants

Area of Science:

• Biomedical Engineering
• Neuroscience
• Materials Science

Background:

• Percutaneous wires for neural recording risk infection and degradation, limiting chronic prosthetic use.
• Wireless, implantable solutions are needed for robust, long-term neural interfaces.

Purpose of the Study:

• To develop DustNet, a fully wireless, ultrasonically-powered, spatially distributed neural recording system.
• To enable high-throughput communication between multiple neural implants over a single ultrasound link.

Main Methods:

• Developed DustNet, a network of miniaturized, ultrasonically-powered neural recording implants.
• Implemented a time-division multiple-access (TDMA) protocol with 16-level amplitude modulation for efficient data transmission.
• Utilized a 0.7 mm³ implant with a CMOS IC fabricated in 28nm process.

Main Results:

• Demonstrated system functionality with 4 implants at 90 mm depth in oil.
• Achieved a maximum data rate of 200 kb/s with individual implants transmitting at 50 kb/s.
• Showcased low power consumption (7 µW) and potential for scaling to 400 kb/s.

Conclusions:

• DustNet offers a promising solution for chronic, wireless neural recording.
• The system enhances prosthetic control by enabling natural movement reconstruction.
• Ultrasonic power and advanced communication protocols pave the way for next-generation neural interfaces.