How Laboratory Experiments Can Be Exploited for Monitoring Stress in the Wild: A Bridge Between Laboratory and Daily Life

Abstract

Chronic stress leads to poor well-being, and it has effects on life quality and health. Society may have significant benefits from an automatic daily life stress detection system using unobtrusive wearable devices using physiological signals. However, the performance of these systems is not sufficiently accurate when they are used in unrestricted daily life compared to the systems tested in controlled real-life and laboratory conditions. To test our stress level detection system that preprocesses noisy physiological signals, extracts features, and applies machine learning classification techniques, we used a laboratory experiment and ecological momentary assessment based data collection with smartwatches in daily life. We investigated the effect of different labeling techniques and different training and test environments. In the laboratory environments, we had more controlled situations, and we could validate the perceived stress from self-reports. When machine learning models were trained in the laboratory instead of training them with the data coming from daily life, the accuracy of the system when tested in daily life improved significantly. The subjectivity effect coming from the self-reports in daily life could be eliminated. Our system obtained higher stress level detection accuracy results compared to most of the previous daily life studies.

Making Wearable Technology Available for Mental Healthcare through an Online Platform with Stress Detection Algorithms: The Carewear Project

“…After obtaining more (labeled) data, we could use machine learning techniques to better cope with additional information and further increase the performance. Can et al [3] state that models that are trained using data collected in lab conditions outperform models that are solely based on real-life data. Thus, additional data should be collected in the lab and in real life.…”

“…As also stated by Can et al [3], Larradet et al [4], and Sun et al [5], detecting stress in real life is much more difficult than in lab conditions. Because of this, the current study's goal was a first assessment of the performance of the implemented algorithms through comparisons with comprehensive annotated data in a small healthy sample captured in real life.…”

“…As stated above, self-reports are mostly used in real-life experiments to define the ground truth. Because of this, Can et al [3] state that achieving precise annotations and identification of the perceived stress in real life is a difficult task. They also state that the stress level experienced in the laboratory is different from daily life stress.…”

“…Wearables are a specific type of mHealth application consisting of sensors and devices that can be worn on the body and can collect longitudinal and continuous data on a reliable and noninvasive manner outside of lab settings. However, Can et al [3], Larradet et al [4], and Sun et al [5] state that measuring physiological signals during everyday activity is more difficult than in lab conditions. A first challenge is that the physiological responses of mental stress can be masked by physical activity.…”

Toward Emotion Recognition From Physiological Signals in the Wild: Approaching the Methodological Issues in Real-Life Data Collection

“…Lamichhane et al ( 2016 ) monitored subjects for 5 days and addressed inter-individual differences using a Stress Response Factor in order to improve stress recognition models. Can et al ( 2020 ) compared the machine learning models on laboratory data and on daily life data. When the models were trained the data in-the-lab, the accuracy of the system when tested in-the-wild improved significantly reaching 74% detection rate.…”