Classification and analysis of electrical signals in urinary bladder smooth muscle using a modified vector quantization technique

Voiding of the urinary bladder depends on contraction of the smooth muscle of its wall, known as detrusor muscle, and this in turn relies on electrical signals generated in the muscle cells. The exact shapes of these signals contain important information about bladder biophysics, but are poorly understood. We present an analysis of detrusor signals using a vector quantization technique based on a modified k-means clustering algorithm for the automatic detection and classification of the signals. We find that our procedure is able to sort the signals from a mixed pool into three predefined classes with an overall sensitivity of 0.9 and a specificity of 0.97. The various features of the signals belonging to an example class are evaluated for inter-feature correlation, and these correlations appear to be consistent with certain hypotheses about the mechanism of generation of the signals. Our work offers a novel approach to analyzing intracellularly recorded signals and inferring muscle biophysics at the cellular level.