Inducing and Assessing Acute Mental Stress in Controlled Conditions: Topical Review and Guidelines for Effective Experimental Protocols

Mental health is influenced by the fast-paced nature of life. In this scenario, stressful events play an important role. Extensive research has been carried out to develop non-invasive devices for stress detection, which primarily use physiological data and, more recently, artificial intelligence algorithms. When developing either a new device or algorithm, tests in controlled environments are preferred, because of the supervision of possible confounding factors while running the experiments. However, because of the extremely subjective perception of stress, the characteristics of the investigated samples, and the conditions under which the experiment is conducted, the data may not be representative of a perceived stressful condition, leading to biases. Given the importance of reliable experimental protocols for stress induction, especially if cortisol level is not monitored, this work aims to present approaches for inducing and assessing acute mental stress in controlled conditions, analyzing the problem from engineering and psychological perspectives. All the phases of the experimental protocol are discussed, examining both the factors that could induce stress and the assessment tools, like questionnaires and physiological signals. The analysis of the latter will be focused on the exogenous factors that may compromise the measures, providing solutions for their mitigation. With this work, researchers with different backgrounds can improve the efficacy of their studies, limiting biases and misleading results.