06/2019 – 11/2019
Cortisol – commonly known as the “stress-hormone” – plays an important role in examining the functioning of the hypothalamic-pituitary-adrenal (HPA) axis, which controls the response to stress. It is one of the few hormones required for life and involved in a number of vital functions . Especially when it comes to mental and physical strain, pain or hypotension, the level of cortisol is increased in order to provide energy substrates for metabolism . The occurrence of mental strain in form of stress, e.g. at work, in a relationship or at home activates the HPA system and helps individuals to cope with acute, time-limited stressors. Frequent or chronic activation of the HPA axis in response to stress is thought to be associated with adverse physical and mental health conditions .
The adaption to stress is also stated to be detected in the diurnal cycle of cortisol. Normal cortisol levels are high on waking, rise to a peak level 30 minutes after waking – also known as cortisol awakening response (CAR) – and slowly decrease until bedtime, when they reach the lowest value. Attenuated rhythms are often related to chronic psychosocial stress, as well as poor mental or physical health. However, varying psychological factors and types of stress lead to other outcomes in the cortisol profile . Therefore, it is important to determine different groups of stress responders and link them to the corresponding biological and psychological variables. Another parameter, which is suggested to reflect stress related changes, is the enzyme alpha-amylase. Similar to cortisol, amylase is increased in states of stress .
So far, linkages between cortisol, as well as amylase-patterns and mental health have been evaluated based on statistical methods. Another promising approach to examine these relations is the application of machine learning techniques, e.g. clustering algorithms, to biological and psychological data. Up to now, there have been few publications about this topic. Galatzer-Levy et al. evaluated the relevance of machine learning to the study of stress pathology, recovery, and resilience. In this work, supervised learning methods were used to classify and differentiate diseases . Similar methods were also used for the classification of different stress response patterns based on saliva and blood markers . With the use of modern machine learning methods, it is a lot easier to take into account several features and assess complex relationships in the field of mental health.
The goal of this bachelor’s thesis is therefore to cluster different stress responder types based on diurnal saliva (cortisol and alpha-amylase) and blood markers (Interleukin-6) and to interpret the results of the clustering process with respect to prior knowledge from literature. The data used in this thesis was acquired in a previous study with 75 subjects. Furthermore, another goal is to analyze whether possible links between diurnal cortisol and amylase patterns and other variables exist and whether these correlations can solely be predicted based on survey-assessed variables, such as self-esteem, suppression, socioeconomic status, etc.
- J. M. Smyth et al., “Individual Differences in the diurnal Cycle of Cortisol,” Science (80)., vol. 22, no. 2, pp. 89–105, 1997.
- F. Lang and P. Lang, “Hormonale Regulation,” in Basiswissen Physiologie, Berlin, Heidelberg: Springer Berlin Heidelberg, 2011, pp. 235–268.
- A. S. Desantis, C. W. Kuzawa, and E. K. Adam, “Developmental origins of flatter cortisol rhythms: Socioeconomic status and adult cortisol activity,” Am. J. Hum. Biol., vol. 27, no. 4, pp. 458–467, 2015.
- E. K. Adam, M. E. Quinn, R. Tavernier, M. T. McQuillan, K. A. Dahlke, and K. E. Gilbert, “Diurnal cortisol slopes and mental and physical health outcomes: A systematic review and meta-analysis,” Psychoneuroendocrinology, vol. 83, pp. 25–41, 2017.
- U. M. Nater and N. Rohleder, “Salivary alpha-amylase as a non-invasive biomarker for the sympathetic nervous system: Current state of research,” Psychoneuroendocrinology, vol. 34, no. 4, pp. 486–496, 2009.
- I. R. Galatzer-Levy, K. V. Ruggles, and Z. Chen, “Data Science in the Research Domain Criteria Era: Relevance of Machine Learning to the Study of Stress Pathology, Recovery, and Resilience,” Chronic Stress, vol. 2, p. 247054701774755, 2018.
- L. Abel, R. Richer, A. Küderle, S. Gradl, B. M. Eskofier, and N. Rohleder, “Classification of Acute Stress-Induced Response Patterns,” in Proceedings of the 13th EAI International Conference on Pervasive Computing Technologies for Healthcare – PervasiveHealth’19, 2019, pp. 366–370.