Large-scale wearable data reveal digital phenotypes for daily-life stress detection

Abstract

Physiological signals have shown to be reliable indicators of stress in laboratory studies, yet large-scale ambulatory validation is lacking. We present a large-scale cross-sectional study for ambulatory stress detection, consisting of 1002 subjects, containing subjects’ demographics, baseline psychological information, and five consecutive days of free-living physiological and contextual measurements, collected through wearable devices and smartphones. This dataset represents a healthy population, showing associations between wearable physiological signals and self-reported daily-life stress. Using a data-driven approach, we identified digital phenotypes characterized by self-reported poor health indicators and high depression, anxiety and stress scores that are associated with blunted physiological responses to stress. These results emphasize the need for large-scale collections of multi-sensor data, to build personalized stress models for precision medicine.

Prediction of stress and drug craving ninety minutes in the future with passively collected GPS data

“…Our findings complement a body of published work by other investigators who have used machine learning with GPS-based, EMA-based, or sensor-based inputs to predict drug use, 23 smoking, 24,25 exercising, 26 diet-related behaviors, [27][28][29][30] and mood changes, [31][32][33][34][35] on time scales ranging from hours to days. A closely related body of work used similar inputs for automated detection of current (not future) cigarette cravings, 36 food cravings, 37 stress, [38][39][40][41] drinking, 42 manic episodes, 43,44 and mood. [45][46][47][48][49][50] Prediction or detection accuracy in these studies was greatest for targets that had clear, enduring signatures, such as the transition from a depressive state to a manic state, which, with digital phenotyping, was detected with sensitivity and PPV of 0.97 on a whole-day time frame (that is, detection was counted as correct if the model flagged a transition before a whole day had elapsed).…”

“…[45][46][47][48][49][50] Prediction or detection accuracy in these studies was greatest for targets that had clear, enduring signatures, such as the transition from a depressive state to a manic state, which, with digital phenotyping, was detected with sensitivity and PPV of 0.97 on a whole-day time frame (that is, detection was counted as correct if the model flagged a transition before a whole day had elapsed). 44 More elusive, however, was the detection of mental states such as stress, for which sensitivity and PPV were often below 0.50, 41,50 and the prediction of future states or events, for which sensitivity, specificity, and PPV (when PPV was reported) tended to be in the 0.70s at best, occasionally reaching the 0.80s. 23,25,27,30,35,40 These are mostly averages across all participants in a given study; many of the published reports do not provide information on how accuracy varied across people or time.…”

Combining physiological data and subjective measurements to investigate cognitive load during complex learning

“…Former research has investigated the relationship between learners' cognitive load and their physiological behaviour. The physiological measures that have been used to investigate cognitive load are among others heart rate by electrocardiography (ECG), brain activity by electroencephalography (EEG), eye activity (e.g., blink rate, pupillary dilation), EDA, heat flux and ST (Antonenko, Paas, Grabner & van Gog, 2010;Haapalainen et al 2010;Scharinger, Soutschek, Schubert & Gerjets, 2015;Smets et al, 2018;Zagermann, Pfeil & Reiterer, 2016). Although a lot of physiological data, such as brain and eye activity, has been proven to be highly effective for measuring cognitive load, these types of physiological data often requires expensive sophisticated equipment that is highly obtrusive in measuring cognitive activities, especially in ecological valid contexts (Chen et al, 2016;Scharinger et al, 2015).…”

Quantified Complementary and Alternative Medicine: Convergence of Digital Health Technologies and Complementary and Alternative Medicine

“…Further, mood states could be assessed using different sensor technologies. Even if this may be an indirect measure of one's psychological state, we believe we now have better tools at hand to quantify, for example, stress in real time [3,4].…”