EN
BACKGROUND AND AIMS: A common paradigm in electrophysiology is a study of responses of neural tissue to voltage or current stimulation. Even a short stimulation can elicit artifacts lasting for tens of milliseconds after stimulation, of amplitude comparable to the responses to be measured. The ability to automatically detect and remove stimulation artifacts from physiological recordings would improve the reliability of biological conclusions obtained in the experiments. In this work we show how to assess and subtract such artifacts from multi-electrode array (MEA) recordings. METHODS: In acute brain slice preparations of the rat somatosensory cortex we investigated in-vitro evoked extracellular responses using 60-channel MEAs (Multichannel Systems). We applied voltage stimulations at different locations under artificial cerebro spinal fluid (ACSF) (a) without tissue, (b) with tissue and (c) with tissue after application of sodium channel blocker (TTX). RESULTS: We have considered several models of artifact dependence on the distance from the stimulating electrode and on time from the onset of stimulation. We found that the best model for prediction of the artifact on every electrode for physiological recordings is its value recorded on slices after TTX application. The slice need not be the same. A proxy from recordings with just the ACSF can be used to construct a still acceptable model. Using Independent Component Analysis and Current Source Density reconstruction we investigate the structure of the artifact and example physiological responses. CONCLUSION: With the help of extracellular recordings of slices after application of TTX it is possible to reliably estimate stimulation artifacts interfering with physiological responses and thus improving the quality and precision of data obtained in such experiments.