Wide-Band Shared LNA for Large Scale Neural Recording Applications

Abstract

Optimizing the LNA front-end of a neural recording channel for noise allows achieving a bandwidth that can be exploited to interface several electrodes in a frequency-division multiplexing approach, resulting in an optimization of power and area consumption for the overall neural recording system. This paper presents a wide-band LNA front-end whose bandwidth is maximized for the given noise level, so that it can be shared among different electrodes. The resulting architecture achieves a remarkable NEF value, as demonstrated by the simulations performed in a commercial 130nm CMOS technology.