Data Collection for Stress Identification

According to the WHO (World Health Organization) the stress has become a worldwide epidemic and has been considered as the disease of the century, attaining people of different ages and occupations. Stress is defined as a pressure or tension, of a physical, mental or emotional type, which overwhelms the individual and adversely affects the performance of his physical and intellectual functions. When stress is prolonged, the immune system is affected, leaving the body subject to disease and premature aging. Among the psychophysiological reactions associated with excessive stress are chronic fatigue, ulcers, depression, hypertension and heart attack. In light of these consequences, it is important that measures are taken to early identifying the symptoms of stress, allowing people to be properly treated and resume their normal activities.

The studies presented in literature suggest the possibility of using biological signals as indicators of the presence of stress. In these studies, certain physiological variables are measured and related to this stress on the individual. However, most previous studies focused on acute stress, meaning that short-lived. The objective of this study was to relate certain physiological indicators with the presence of chronic stress, ie that long-lasting and often not perceived by the individual.

To this end, the investigators performed the measurement of blood pressure values, skin temperature, galvanic skin response and heart rate variability, and compared with the results provided by the ISSL (Lipp's Inventory of Stress Symptoms for Adults), PSS (Perceived Stress Scale) and SRRS (Social Readjustment Rating Scale), as the stress condition of the individual. The heart rate variability was checked in both time and frequency domain. All of the physiological signals were monitored by non-invasive and non-intrusive sensors and totally safe for participants.

To investigate the relationship between the physiological variables and the existing chronic stress condition the investigators used an algorithm for analysis of biological signals based on Multilayer Perceptron Neural Network with Backpropagation Algorithm (MLP) and Principal Components Analysis (PCA). To evaluate the performance of the classifier on the identification of chronic stress the investigators check the indicators of precision, sensitivity, specificity and accuracy. The results were then analyzed by Receiver Operating Characteristic Curves (ROC).

Ethical approval was obtained from the Local Ethics Committee (CAAE-28201014.2.0000.5497) for the participants. They had been informed about the methodology and confidentiality of their personal information. All procedures were performed after written informed consent from all participants. Data collection was completed in October 2014.

Inclusion criteria

• Healthy volunteers

Exclusion criteria

• Presence of any disease
• Use of medications to control blood pressure
• Use of anxiety medications and antidepressants