Arbitrary units are a composite and useful measure of muscle sympathetic nerve activity

In humans, the muscle sympathetic nerve activity (MSNA) signal is challenging to detect, record and analyze. Several methods exist that attempt to capture the latent construct of MSNA. We directly compared the performance of five MSNA parameters: burst frequency, burst incidence, median burst amplitude, arbitrary units (AU) and fractal dimension (FD). The MSNA signal was recorded in 33 subjects for approximately 30 min before, during and after the application of a graded cold pressor test stimulus at 18 degrees C, 10 degrees C and 2 degrees C in random order with an adequate wash-out period. Using coefficient of variation, Shannon's entropy and principal component analysis, we observed that these five parameters defined two physical and conceptual domains of MSNA-frequency and amplitude. Since AU combines information from both these domains, we observed that it explained maximum inter-subject and inter-experimental segment variation. FD did not explain the inter-subject variability and was identified as a unique parameter in the factor analysis. Epidemiological studies that attempt to quantify MSNA may consistently use AU as the parameter for quantification of MSNA.