Prediction of Psychological Flexibility with multi-scale Heart Rate Variability and Breathing Features in an “in-the-wild” Setting

Psychological flexibility (PF) has recently emerged as an important determinant in pain related outcomes. It is related to pain adaptation, social functioning and emotional well-being. A recent study indicates PF being a significant predictor of heart rate variability (HRV) and mediating relationship between HRV and pain interference. In recent years, HRV has been studied using non-linear dynamics approaches which better quantify the fractal behavior of the inter-beat interval time series. In this study, we propose the use of multi-scale HRV features for predicting PF. The new features are tested on a dataset collected from 200 hospital workers (nurses and staff) during their normal work shifts. We show that fusion of breathing signal features further improves the performance showing the complementarity of two feature sets. We achieve an overall improvement of 4.54% F1-score over benchmark HRV features. These results indicate the importance of non-linear features for PF measurement. An accurate measurement of PF can help in developing pain and distress intervention methods by unobtrusive measurement of physiological signals using wearable sensors in real life conditions.