Jeremy Holleman, Dan Yeager, Richa Prasad, Joshua R. Smith and Brian Otis
Intel and University of Washington

This paper presents the NeuralWISP, a wireless neural interface operating from harvested RF energy. The NeuralWISP is compatible with commercial RFID readers and operates at a range up to 1m. It includes a custom low-noise, low power amplifier IC for processing the neural signal and an analog spike detection circuit for reducing digital computational requirements and communications bandwidth. Our system monitors the neural signal and periodically transmits the spike density in a user-programmable time window. The entire system draws an average 20µA from the harvested 1.8V supply.