Feature detection and classification of Action Potentials from detrusor smooth muscle cells

Unlike other excitable cells, the urinary bladder smooth muscle cells exhibit a wide variety of spontaneous Action Potential (sAP) shapes. The complex network of gap-junction coupled smooth muscle cells and innervations makes it difficult to comprehend the origin of such a shape diversity. It is expected that the shape of the sAPs observed intracellularly from a smooth muscle cell give hints about the syncytial properties of the tissue in its neighborhood. In this work, four major features - foot convexity, after hyperpolarization, after depolarization, and extended after hyperpolarization - which define the shape of sAPs were identified. Algorithms were designed to automatically detect the presence of each features. The proposed algorithms were able to detect the features with accuracies greater than 70%. Further, a scheme to label the sAPs based on the presence/absence of the features is also proposed. The labels could be used to classify the sAPs into different groups.