Stress recognition and classification in network interactions of physiological systems from biosignals: recent trends and novel approaches

Stress is an important complex physiological systems response aimed to prepare the body to handle the challenge(s) presented by an internal or external environment – a stressor or their combination. While a stress response to acute stressor is usually helpful to cope with the given challenge, a chronic exposure to long lasting stressors mostly of the psychological or social kinds can lead to cumulative physiological and psychological effects, resulting in the increasing risk of health problems including cardiovascular disease, metabolic disease, weakening of immune response, digestive problems and various psychological consequences. Dysregulation of the autonomic nervous system, potentially leading to a cascade of physiological effects that worsen individual health conditions is one of the crucial mechanisms of above mentioned associations. Early stress detection and its severity quantification are thus crucial not only to avoid potentially irreversible damage, but also to promote mental health and well-being in various daily-life contexts (e.g. at work or at home).

The widespread use of non-invasive and cost-effective wearable devices, combined with more efficient and available computational resources and enhanced signal processing techniques, is paving the way towards a real-time objective stress recognition using biosignals recording and analysis. Recently, methods from dynamical systems and information-theory, capable to quantify the changes in the dynamical interactions between various components of the controlled systems (according to the so-called “Network Physiology” approach), have been exploited for multilevel stress assessment. At the same time, many studies have demonstrated the feasibility of performing stress classification using artificial intelligence algorithms, including both machine and deep learning (either utilizing supervised or unsupervised approaches). These recent trends have resulted in promising results not only for stress recognition and classification, but also to better elucidate the complexity of physiological responses to stress.

This Research Topic aims at promoting investigation of methodological approaches derived from advances in bioelectronics, biosignal processing and artificial intelligence, as well as in basic and integrative physiology and clinical medicine, to provide insight on stress recognition and classification in network interactions of physiological systems from biosignals.

Therefore, we solicit reviews, systematic reviews, mini-reviews, original research articles, brief research or case reports, opinions, perspectives or general commentaries, and method papers which cover (but are not limited to) the following themes:
• current advances in the development of experimental protocols for stress elicitation;
• innovative hardware equipment for non-invasive recording of multimodal biosignals;
•novel signal processing and time series analysis methods;
• supervised and unsupervised machine learning and deep learning algorithms;
• dynamics networks approaches for assessing the effects of stress on physiological systems and their interactions.

Within these scopes, a particular emphasis will be devoted to the more detailed description and quantification of the physiological responses to stressors of different types and intensities, produced by a complex interplay of nervous, endocrine, immune and other systems involved in the stress response.

Keywords: Stress, biosignals, wearable devices, machine learning, deep learning, time series analysis, autonomic nervous system, dynamic interactions, Network Physiology

Important Note: All contributions to this Research Topic must be within the scope of the section and journal to which they are submitted, as defined in their mission statements. Frontiers reserves the right to guide an out-of-scope manuscript to a more suitable section or journal at any stage of peer review.