Influence of evoked response latency and amplitude on stimulus artifact removal

The removal of stimulus artifacts is a challenge not only to evoked potentials but also to other electrically-elicited signals such as M wave and direct cortical responses. Several techniques try to remove stimulus artifacts. However, the influence of the amplitude ratio between signal and artifact has not been reported yet. Neither has been the effect of the delay between stimulus and signal. So, this work aimed to study the relationship between the evoked response's parameters (amplitude; latency) and the associated errors, for several amplitude ratios and delays between stimulus artifact and evoked response simulation models. In order to do so, the evoked responses were represented by a mathematical function with varying amplitudes and latencies. The stimulus artifacts were represented by the concatenation of a rectangular bipolar pulse and three exponential functions. Reasonable results were obtained for these models, whenever the amplitude ratio between evoked response and artifact was between 5% and 40%, and the evoked response's latency was within the range of 3ms to 8ms. These results showed that the technique has the potential to contribute to stimulus artifact removal on somatosensory evoked potentials recorded at the elbow, for both wrist and hand stimulation of the median nerve.