A pattern recognition method to distinguish gradual unilateral diaphragm paralysis in the cat.

This work deals with the application of a pattern recognition method to distinguish the degree of diaphragm paralysis after gradual unilateral sections of phrenic nerve rootlets in anesthetized, spontaneously breathing cats. The data set consisted of the features that characterize breathing pattern and of phrenic nerve amplitude. The method called for stratification of 6-dimensional vectors into three classes: intact, partial, and complete unilateral phrenicotomy, which offers the possibility to construe the classification rule on the basis of the information contained in a set of feature vectors with the known class-membership. This method deals with the use of a distance function as a measure of similarity between two feature points. The results show that the degree of diaphragm paralysis could be recognized with the probability higher than 90%. Distinguishing the severity of diaphragmatic dysfunction and the compensatory strategies of the respiratory system, knowing only a handful of basic values describing breathing pattern, might have a practical meaning in respiratory emergencies.