Analysis of physiological responses from multiple subjects for emotion recognition

Abstract

Psychological disorders, including emotion and behavioral disorders, are common in the modern society. For example, previous studies by the New Zealand Mental Health Survey and the US National Co-morbidity Surveys (NCS), have found that the incidence of depression from ages 18 to 32 to be around 18%. Thus, there is a need to effectively monitor the emotional or affective states of these patients with psychological illness. Whilst real-time and continuous psychological monitoring systems are still not prevalent, the monitoring of physiological signals are made easier with mobile sensors that can be attached to the human body. These physiological signals can then be processed using an embedded processor to provide an alternative means for automatic emotion recognition. However, for such a system to be developed, the relationship between physiological and psychological signals has to be understood. This paper aims to address this by investigating the relationship between the emotional experiences from multiple subjects and their physiological responses, including the skin conductance, heart rate, respiration and movements of the facial muscles. In summary, preliminary evaluations described in this paper demonstrated that the heart rate, respiration, blood volume pulse and electromyography signals have an impact on the recognition rate achievable by the proposed multi-user physiological response-based emotion detection system.