Stimulation artifact rejection in closed-loop, distributed neural interfaces

Abstract

A prototype neuromuscular interface with distributed sensing and stimulation functions was developed to evaluate power isolation and maximize stimulation artifact rejection. Evaluation was performed in saline with representative stimulating and sensing electrodes. Stimulation pulses paced at 100 Hz were applied while sensing frequencies from 2.5 to 500 Hz. Two isolation topologies were evaluated, one in which a common ground connection was shared among the modules with AC-coupled power and one in which power and ground were both AC-coupled. The fully AC-coupled design provided 23 dB better common-mode rejection of stimulation artifact than the shared ground design. The effect of reference impedance between the fully isolated sense module and the saline tank was also found to significantly impact stimulation artifact rejection. These results inform system-level design for closed-loop distributed systems wherein the application necessitates concurrent stimulation and biomarker sensing functions.