Characterization Of Physiological Signals Under Cognitive Stress

A fusion of wearable device sensors and an electroencephalography system was implemented to identify changes under controlled situations related to stress in the student’s population, to assess whether the physiological signals can detect variations in the stress level by a descriptive statistical analysis. The signals used are: movement, blood volume pulse, electrodermal activity, heart rate, temperature and electroencephalography. The analysis was made in volunteers between 18 and 25 years old, who are young adults with at least 10 years of schooling and being apparently healthy. The recording protocol consisted of 3 consecutive stages taken in a single session, which consisted of two math tests (stage 1 and stage 3) with a relaxation period (stage 2) in between. The physiological signal data were segmented per stage and descriptive statistical metrics were calculated per volunteer. For the appropriate proper selection of metrics, they were compared using the Repeated Measures one-way ANOVA method with α = 0.05, using Tukey's multiple comparisons tests. It was found that the heart rate and temperature signals, along with the descriptive statistical metrics, range, and kurtosis, are the best suited for the study to find the differences. Volunteers in stage 3 maintained a state of relaxation equal to or less than that of stage 2.