Heart rate variability classification based on personalized frequency domain indices

Abstract

Heart rate variabilities have been utilized as means to assess the autonomic activity. However, large individual variations prevent from consistent quantitative interpretation. This paper introduces personalized indices to concur the difficulty. Drug induced (Atropine and Propranolol) distinct physiologic states have been made and ECG data were obtained from 14 normal subjects. Personalized relative indices rLF, rHF, rLF/rHF, rMF and average/variance of HR have been utilized for the classification. An improved dichotomous classification scheme achieved average classification sensitivity of 0.91 and specificity 0.97 showing the effectiveness of proposed personalized indices.