Characterization of apnea events in sleep breathing disorder by local assessment of the fractal dimension of heart rate

Aim of this work is to describe a novel method for estimating overnight modulations of fractal dimension (FD) of R-R intervals (RRI), and to evaluate if these FD estimates may characterize different apnea types in sleep studies. Polysomnographies were performed in 10 patients with severe obstructive sleep apnea syndrome. Both obstructive (OS) and mixed (MX) apneas were found in all patients. The FD of RRI interpolated series was estimated “locally” over a running window of 15 s, overlapped by 14 s, by an algorithm based on the Mandelbrot's definition of the FD of geometric curves. The FD estimator appeared able to follow rapid changes in RRI dynamics. Moreover, it was significantly higher in OS than in MX apneas for 9 out of 10 patients (p<;0.05). Thus, the proposed running FD estimator might enhance the capability of the simplest sleep monitoring devices to characterize sleep breathing disorders.