Neural Signals in Multichannel Neural Recording Systems

We have used a well-known technique in wireless communication, pulse width modulation (PWM) of time division multiplexed (TDM) signals, within the architecture of a novel wireless integrated neural recording (WINeR) system. We have evaluated the performance of the PWM-based architecture and indicated its accuracy and potential sources of error through detailed theoretical analysis, simulations, and measurements on a setup consisting of a 15-channel WINeR prototype as the transmitter and two types of receivers; an Agilent 89600 vector signal analyzer and a custom wideband receiver, with 36 and 75 MHz of maximum bandwidth, respectively.

Furthermore, we present simulation results from a realistic MATLAB-Simulink model of the entire WINeR system to observe the system behavior in response to changes in various parameters. We have concluded that the 15-ch WINeR prototype, which is fabricated in a 0.5-mum standard CMOS process and consumes 4.5 mW from plusmn1.5 V supplies, can acquire and wirelessly transmit up to 320 k-samples/s to a 75-MHz receiver with 8.4 bits of resolution, which is equivalent to awireless data rate of ~ 2.56 Mb/s.