A Sensor-enabled Digital Trier Social Stress Test in an Enterprise Context

Area of Science:

• Psychophysiology
• Human-Computer Interaction
• Occupational Health

Background:

• The Trier Social Stress Test (TSST) is a standard method for inducing stress.
• Traditional TSST administration occurs in lab settings, limiting workplace application.
• Digital adaptation of TSST is needed for remote and scalable stress assessment.

Purpose of the Study:

• To analyze multi-modal sensor data from an enterprise-based digital TSST.
• To evaluate the efficacy of a semi-automatic, digitally administered TSST.
• To determine the role of State Trait Anxiety Inventory (STAI) scores in digital TSST assessment.

Main Methods:

• Collected noninvasive sensor data (Galvanic Skin Response, Photoplethysmogram) during a digital TSST.
• Administered the digital TSST in a semi-automatic mode within an enterprise.
• Utilized support vector machine classification with STAI scores as ground truth.

Main Results:

• The digital TSST effectively induced stress, measurable via sensor data.
• State Trait Anxiety Inventory (STAI) scores are crucial for assessing digital TSST effects.
• A support vector machine classifier achieved an F1 score of 0.723 using STAI scores.
• An improved F1 score of 0.847 was obtained using change in STAI scores as ground truth.

Conclusions:

• The digital TSST is a viable tool for stress induction and measurement in occupational settings.
• Integrating STAI scores enhances the accuracy of digital TSST outcome assessment.
• The digital TSST offers a scalable and adaptable approach to workplace stress research.