Spectral analysis of multiunit action potential trains of muscle sympathetic nerve activity in humans

The application of conventional signal processing methods used to obtain an integrated signal from muscle sympathetic nerve activity (MSNA) reduces the amount of information and may confound the spectral characteristics. We present a novel alternative method of processing the raw MSNA signal using a wavelet transform denoising technique that enables detection of individual action potentials and facilitates spectral analysis. A spike density function (SDF) is generated from the denoised signal by replacing the detected action potentials with delta functions and convolving with a 3 Hz Gaussian filter. This method was validated using data from a sinusoidal neck suction (NS) experiment in humans. The results of the analysis indicate that the oscillations of sympathetic nerve firings closely followed the NS frequency. In conclusion, the SDF representation allows for a novel and insightful analysis of spectral components of action potential trains in raw MSNA.