Evaluation of normalised Renyi entropy for classification of cardiac autonomic neuropathy

Early detection of cardiac autonomic neuropathy (CAN) is vital to provide timely health care. There are a variety of measures that can be considered for detection of CAN from the heart beat signal, which should provide a relatively noninvasive test. In this work we consider the use of two measures based on the RR interval - the standard deviation and the Renyi Entropy. We find that caution is indicated in the use of entropy for two reasons. First, these two measures are related, so that the researcher who provides Renyi entropy as a measure may in fact be measuring only the standard deviation. Second, there are several methods for calculating Renyi entropy, so the method needs to be carefully selected to provide meaningful results.