Galvanic Skin Response Measurement Data Processing for User-Related Information Extraction

Emotions play a key role in everyday life of human beings. For this reason, in the last years, researchers have exploited physiological changes caused by emotional stimuli in order to develop methods to automatically recognize the emotional involvement of individuals. Among physiological signals, the Galvanic Skin Response (GSR) is one of the most interesting in emotion research. In this work, the whole process of acquisition, elaboration, and analysis of the GSR signal is described in detail. GSR data were measured by using a wearable wrist-worn device. Participants were enrolled and contributed to two experimental tests, where they were submitted to physical and auditory stimuli. The collected data were analyzed in the time domain by computing the number of GSR peaks for each signal, and in the frequency domain by calculating the amount of spectral content within specific frequency sub-bands. GSR signals acquired after physical efforts of different intensities were analyzed in both the time and frequency domain. The results of the time domain analysis show a growth of the GSR peaks number as the intensity of physical efforts increases. While the analysis in the frequency domain shows a shift of the GSR spectral content towards higher frequencies when subjects perform physical exercises of increasing intensities. GSR data recorded during the listening of acoustic stimuli of different pleasure levels were examined in the time domain only. The analysis of the results shows that pleasant and neutral sounds do not produce evident effects on the number of GSR peaks, while the listening of an unpleasant sound elicits a slight increase in the GSR peaks rate. Such an increase is higher when the sound is shorter, probably due to the expectation phenomenon.