Challenges and opportunities in processing muscle sympathetic nerve activity with wavelet denoising techniques:...

An important issue in analysis of muscle sympathetic nerve activity (MSNA), particularly those measures made in humans, is the problem that background noise of varying levels from recording to recording may interfere with accurate assessment of neural discharge patterns and overall activity. In this study, the utility of wavelet denoising approaches for processing MSNA signals was examined with emphasis on 1) determining whether this approach could improve the signal-to-noise (SNR) in the integrated neurogram, and 2) detecting intra-burst single action potential spikes. The utility of wavelet denoising was examined in simulated data and in original human data with three recordings of varying SNR (low, moderate and high) obtained from two healthy individuals. Only in the high SNR signal was the noise removed without concurrent loss of signal. Using a threshold-detecting algorithm individual depolarization spikes were detected in denoised recordings of high original SNR (>3:1) from four individuals and the interspike interval characteristics of these were quantified on a burst-by-burst basis. Compared with baseline (15+/-1 spikes/burst) a reflexive increase in spike count (29+/-4 spikes/burst) was observed during a held maximal inspiration (P<0.01) with concurrent reductions in inter-spike interval (P<0.01). The findings indicate that within multiunit bursts of sympathetic neural activity in the band-pass filtered neural signal, there are particular frequency components that appear to be shared between the signal and noise. This may limit the utility of wavelet denoising to enhance detection of neural bursts in the integrated neurogram of MSNA. However, opportunities exist with this approach to detect variations in action potential contributions within each burst of MSNA. This latter observation suggests that this denoising approach provides a new probe to explore MSNA discharge patterns.