HRV-based Stress Level Assessment Using Very Short Recordings

Abstract

The increased stress level of an individual often influences the medical test results. Heart rate variability (HRV) can be used to characterize stress level. Parameters in the time domain are preferred, as they make the assessment possible from relatively short (3 - 5 minute-long) recordings. Electrocardiographic signal (ECG) in Einthoven I lead and photoplethysmographic signal (PPG) at the fingertip were recorded in parallel. This paper analyzes the differences between the lengths of successive heart cycles (NN); the ratio of these differences exceeding 50 ms (pNN50) is considered to be a good measure of stress. The parameter triplet pNNtri has been defined dividing the differences of successive NN intervals into three categories: pNN0_20 (0 – 20 ms), pNN20_50 (20  – 50 ms) and pNN50 (greater than 50 ms). pNNtri characterizes the actual stress level significantly better (especially for senior persons) than pNN50 alone. NN intervals determined from PPG differ from those determined from ECG recorded in parallel.