A spectrum-equalizing analog front end for low-power electrocorticography recording

Abstract

Electrocorticography (ECoG) has spectral characteristics that allow for possible efficiency improvements over traditional broadband recording of low frequency (1-150 Hz) data. In this paper we propose a spectrum-equalization technique for ECoG recording that can potentially decrease the dynamic range requirements of digitization by more than two orders of magnitude, corresponding to more than 6 bits of ADC resolution. We additionally propose relaxed noise requirements for recording based on the signal spectrum and content. We demonstrate a fabricated analog front end and ADC in 65nm CMOS utilizing this equalization technique and with the proposed reduced noise requirements. This paper will show that the fabricated signal chain reproduces ECoG data with fidelity.