Computational intelligence methods for the identification of early Cardiac Autonomic Neuropathy

Cardiac Autonomic Neuropathy (CAN) is a disease that involves nerve damage leading to abnormal control of heart rate. CAN affects the correct operation of the heart and in turn leads to associated co-morbidities. An open question is to what extent this condition is detectable by the measurement of Heart Rate Variability (HRV). However, if possible we wish to detect CAN in its early stage, to improve treatment and outcomes. HRV provides information only on the interval between heart beats, but is relatively non-invasive and easy to obtain. HRV has been conventionally analysed with time- and frequency-domain methods, however more recent analysis methods have shown an increased sensitivity for identifying risk of future morbidity and mortality in diverse patient groups. A promising non-linear method is the Renyi entropy, which is calculated by considering the probability of sequences of values occurring in the HRV data. An exponent α of the probability can be varied to provide a spectrum of measures. In previous work we have shown a difference in the Renyi spectrum between participants identified with CAN and controls. In this work we applied the multi-scale Renyi entropy, as well as a variety of other measures, for identification of early CAN in diabetes patients, using computational intelligence methods. The work was based on measurements from 67 people with early CAN and 71 controls. Results suggest that Renyi entropy forms a useful contribution to the detection of CAN even in the early stages of the disease, and that it can be distinguished from controls with a correct rate of 68%. This is a significant achievement given the simple nature of the information collected, and raises prospects of a simple screening test and improved outcomes of patients.