Ultra low power amplification and digitization system for neural signal recording applications

Abstract

Scope and Method of Study: The scope is to develop a tunable low power fully integrated band pass filter and a low power second order sigma-delta ADC modulator for implantable neural signal amplification and digitization applications, with subthreshold circuit design techniques in different CMOS processes. Since biopotentials usually contain low frequency components, the neural filters in this project have to be able to achieve large and predictable time constant for implantable applications. Voltage biased pseudo-resistors are vulnerable to process variations and circuit imperfections, and hence not suitable for implantable applications. A current biased pseudo-resistor is implemented in the neural filters in this work to set the cutoff frequency, and a Taylor series is used to study its linearity.

Findings and Conclusions: The filters with proposed current biased pseudo-resistors were fabricated in two different CMOS processes and tested. The test results verify that the filters with current biased pseudo-resistors are tunable, and not vulnerable to process variations and circuit imperfections. The filters with current biased pseudo-resistors meet the design requirements of fully integrated, implantable applications. The sigma-delta ADC modulator was designed and simulated in a half micron SOS CMOS process. The simulation results of the ADC confirm the possibility of an ultra low power ADC for neural signal recording applications.