Leon Schmid

Advisors

Luca Abel (M. Sc.), Robert Richer (M. Sc.),  Prof. Dr. Björn Eskofier

Duration

11/2023 – 03/2024

Abstract

Stress is a multifaceted occurrence that has diverse effects on individuals, influencing both their mental and physical well-being [1, 2]. Although acute stress responses may be adaptive in young healthy individuals and thus do not usually pose a health risk, with unremitting threat, the long-term impact of stressors may damage health, especially for older or less healthy individuals [3]. A well-established method to measure an acute psychological stress response are inflammatory markers and cortisol [4]. In particular, stress-induced inflammation can be detected most reliably via blood tests, i.e. an invasive measurement [5]. However, research has shown that an acute stress response also manifests itself in physical movements and posture such as defensive freezing [6] or a slower gait and slumped posture in individuals suffering from depression [7]. A non-invasive nature would enable researchers to more accurately observe how individuals react physiologically to stress in more naturalistic settings as stress-inducing factors arise [5]. Conventionally, motion analysis is conducted through optical or inertial measurement unit (IMU)-based motion capture methods. While these techniques provide the most precise results, they are associated with steep expenses, restricted accessibility, and can disrupt natural behavior when inducing stress. Therefore, we use the openTSST framework, developed at the Machine Learning and Data Analytics Lab, which provides an easy-to-use interface to OpenPose as a web-based platform for video-based motion analysis and feature extraction pipeline with a special focus on the analysis of psychosocial stress [8].

The goal of this bachelor’s thesis is, therefore, to use openTSST on video data collected in a study conducted within the EmpkinS collaborative research center. So far, 45 participants performed the Trier Social Stress Test (TSST) and the control condition on two consecutive days in a randomized order. During this thesis, the dataset will be extended with additional participants. The focus of this thesis is mainly on the feature extraction pipeline of openTSST and the subsequent data analysis. Thus, previous features will be optimized, and new features will be developed, which will hopefully lead to a better description of how an acute stress response affects movement or posture. For this purpose, the existing openTSST framework will be extended accordingly. It will also be investigated what difference it makes whether the study participants perform the TSST while standing or sitting and if or how the applicability of our framework changes. To gain a comprehensive impression of which parameters can influence a psychosocial stress reaction, it will also be analyzed what influence the gender has. Furthermore, machine learning techniques may be used to try to achieve a reliable distinction between the stressed and the control condition.

 

Full Thesis

 

References

[1] D. B. O’Connor, J. F. Thayer, and K. Vedhara, “Stress and Health: A Review of Psychobiological Processes,” Annu. Rev. Psychol., vol. 72, no. 1, pp. 663–688, 2021, doi: 10.1146/annurev-psych-062520-122331
[2] Richer, R., Koch, V., Abel, L., Hauck, F., Kurz, M., Ringgold, V., Müller, V., Küderle, A., Schindler-Gmelch, L., Eskofier, B. M., & Rohleder, N. (2024). Machine learning-based detection of acute psychosocial stress from body posture and movements. Scientific Reports, 14(1), 8251. https://doi.org/10.1038/s41598-024-59043-1.
[3] Schneiderman, Neil, Gail Ironson, and Scott D. Siegel. “STRESS AND HEALTH: Psychological, Behavioral, and Biological Determinants.” Annual Review of Clinical Psychology 1 (2005): 607–28, doi:10.1146/annurev.clinpsy.1.102803.144141.
[4] Ulrich-Lai, Yvonne M., and James P. Herman. “Neural Regulation of Endocrine and Autonomic Stress Responses.” Nature Reviews Neuroscience 10, no. 6 (June 2009): 397–409. https://doi.org/10.1038/nrn2647.
[5] Slavish, Danica C., Jennifer E. Graham-Engeland, Joshua M. Smyth, and Christopher G. Engeland. “Salivary Markers of Inflammation in Response to Acute Stress.” Brain, Behavior, and Immunity 44 (February 1, 2015): 253–69. https://doi.org/10.1016/j.bbi.2014.08.008.
[6] Nicolas Rohleder. “Stress and inflammation – The need to address the gap in the transition between acute and chronic stress effects”. eng. In: Psychoneuroendocrinology 105 (July 2019), pp. 164–171. issn: 1873-3360. doi: 10.1016/j. psyneuen.2019.02.021.
[7] Ron Feldman, Shaul Schreiber, and Ella Been. “Gait, Balance and Posture in Major Mental Illnesses: Depression, Anxiety and Schizophrenia”. In: Austin Med Sci 5.1 (2020). p. 1039.
[8] T. Geßler, “openTSST – An Open Web Platform for Large-Scale, Video-Based Motion Analysis During Acute Psychosocial Stress”, Bachelor’s Thesis in Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg