Multisignal approach for stress and workload analysis

Every day, high stress and mental workload negatively impact people and, depending on the circumstances, could put them in danger or make them hazardous. When perceived, stress sets the mind and the body in a fight or flight situation, i.e., blood is redirected from the minor vessels and retained in the major vessels so that the muscular system can react and allow quick movements. Body temperature lowering, faster heartbeats, or increased sweating, especially in the palm of the hands, are some of the main consequences of this phenomenon. On the other side, mental workload refers to the amount of mental effort necessary to accomplish a task: people required to do more than their abilities or resources allow may perform worse or even fail. Health issues like chronic stress and depression can emerge in case of long-term exposure, possibly escalating into physical sickness; therefore, it is crucial to understand how these factors affect people. This thesis aims to explore the field of stress and workload assessment by analyzing features extracted from meaningful and easy-to-get biological signals, both in terms of intrusiveness and cost, through feasible methods. The Electrodermal Activity (EDA), the Photoplethysmography (PPG), and the body temperature were recorded from the subjects while taking the BiLoad Test. In particular, EDA is the footprint of skin electrical activity, proportional to skin moisture level and highly dependent on the emotional state. On the other hand, PPG tracks the Blood Volume Pulse in time, including respiration and heartbeat components. It is possible to extract from this signal the Heart Rate and Heart Rate Variability, which contains information about the sympathovagal balance, an indicator of sympathetic and parasympathetic nervous system activation. To generate the mental states this study wants to investigate, the BiLoad Test was developed, via the Matlab GUI. It is composed of two subsequent well-known tests adopted from literature: the first one is the Stroop Color and Word Test, mainly used for psychological assessment, chosen to generate stress in people; the second, the N-back Test, challenges people to endure an increasing mental workload through different levels. The g.HIamp 144 channels (a biosignals amplifier) and sensors from g.tec allowed the biological signals recording from about thirty subjects aged between 24 and 41 years while carrying out the test. Signals and test performances - reaction times and answers of each set - were elaborated and post-processed using Matlab. Each biological signal revealed a strong dependency on the participants' emotional arousal induced by the test, allowing the choice of the optimal features to use as indicators. A possible development of this study could be implementing a real-time, automatic detection system for moments of stress and high workload, especially for safety and health applications.