Discriminating affective state intensity using physiological responses

In this paper we explore if physiological signals obtained from a person through wearable sensors permit the correct interpretation of human affective states. A crucial aspect within this field of research is the capability of inducing real-life emotions in laboratory environments. To this end we designed a very strict and regulated experimental protocol. We consider two affective states: a relaxation state and a stressful one. To induce these two states we adopt audio tracks of natural or day-life sounds and math calculations. Moreover, to study different intensities of effectively induced relaxation, we consider two different audio players: a traditional pair of headphones and the Spherison Sound6D pillowⒸ, a special device that provides a complete spherical immersion in what users are listening to. We consider as physiological signals the Galvanic Skin Response (GSR) and the Photoplethysmography (PPG), as they are sensitive measures for emotional arousal. After an extensive analysis on our experimental data, we demonstrate that GSR and PPG signals can successfully distinguish relaxation and stressful states. Moreover, the same physiological signals can discriminate affective state intensity, especially when relaxation is induced adopting the Sound6D technology.