A 2.2 μW analog front-end for multichannel neural recording

In this paper an analog front-end for the multi-channel implantable recording of neural signals is presented. It is comprised by a two-stage AC-coupled low-noise amplifier (LNA) and a one stage AC-coupled variable gain amplifier (VGA). The proposed architecture employs highly power-noise efficient current reuse fully differential OTAs in the LNA stage and a fully differential folded cascode for the VGA stage. Simulation results in AMS 0.18μm validate the proposed architecture under process corners variations with an estimated power consumption of 202μm and 3.lμVrms in-band noise.