Session P81.4

The Estimation Method of Physical Activity Energy Expenditure considering Heart Rate Variability

DH Kim*, JS Cho, HS Oh, YJ Chee, IY Kim

Hanyang University
Seoul, Korea

The estimation methods of Physical Activity Energy Expenditure(PAEE) are double labeled water and pulmonary gas exchange analysis as a gold standard. However, these methods have been usually processed in a laboratory environment, even though some portable systems can measure PAEE in free life. Therefore, many researchers have processed the PAEE estimation using accelerometers. There are many commercialized accelerometers such as Actical, Actigraph, and Actiheart, but these products cannot cover all types of activities correctly because only information on acceleration or heart rate is not enough for correct estimation. So we used the information of Heart Rate Variability(HRV). Most of the commercialized accelerometers and regressive equations of PAEE(Mets, metabolism unit) were specific to a particular activity or posture, so each accelerometer and equation can estimate only a few activities’ energy expenditure correctly. Therefore, to classify each activity using HRV is a meaningful approach for better estimation of PAEE.
We measured the data of ECG and acceleration using a Holter and a triaxial-accelerometer respectively. Each subject wore the Holter, an accelerometer, and a portable gas exchange analyzer together, and data measurements were processed in free life during a day. HRV information was acquired from the ECG, and each datum was analyzed by frequency comparison. All of the process of analyzing was compromised of three steps. First was data segmentation, second was separating the segment, whether static activity (posture such as lying, sitting, and standing) or dynamic activity (including moderate and vigorous activities such as walking, running, and other types of activity in free life), and finally deciding for each segmented data which detail activity it was and then applying a specialized equation that was made for each activity. In the first step, the entire data was segmented evenly by specific time, and each segment’s representative activity, static or dynamic, was decided through the HRV analysis and information of acceleration in the second step. After a large group of activity was divided, each segment was finally assigned a detail activity using information of HRV and designated thresholds of acceleration. At the end of the process, the PAEE was calculated in each category through different regressive equations. We validated our results from the accelerometer using a gas analyzer as a reference.

(Abstract Control Number: 167)