Emotions can override…the more powerful fundamental motives that drive our lives: hunger, sex, and the will to survive. People will not eat if they think the only food available is disgusting. They may even die, although other people might consider that same food palatable. Emotion triumphs over the hunger drive! A person may never attempt sexual contact because of the interference of fear or disgust, or may never be able to complete a sexual act. Emotion triumphs over the sex drive! And despair can overwhelm even the will to live, motivating a suicide. Emotions triumph over the will to live!
The role of emotions is crucial, not only in everyday life, but especially in the clinical manifestation of several neuropsychiatric disorders and mental health problems.
The key research question is: what is the most effective way to measure emotional states? Actually, an incredible amount of information can be obtained by examining facial expressions and physiological reactivity. These measurement methods are non-invasive, readily available, relatively inexpensive and, without doubts, more objective than self-report questionnaire technique, that, at most and not in real time, can reveal just subjective moods.
This project has been motivated by the need for more objective and quantitative tools allowing to measure and identify emotions in both clinical and non clinical contexts.
Actually, the project aims at integrating, using statistical methods, data derived from the analysis of facial expression, biosignals, questionnaires and patients characteristics, in order to yield significant insight into the mechanism of emotional response, thus deepen theoretical knowledge especially in the field of anxiety and eating disorders.
Therefore, the goal of the project is to realize a non-invasive tool able to simultaneously evaluate psychophysiological response and changes in facial expression, thus identifying significant patterns of affective states.
To achieve this goal, a multidisciplinary and highly qualified research team, including psychologists, statisticians, bioengineers and computer scientists, was assembled, to address emotion recognition issues, by combining different methodological expertise.
In particular, image analysis techniques will be used to extract face features, while biomedical signals will be analyzed to acquire more information on the experienced emotion during each experimental session. Statistical shape analysis will be used to derive classification rules of emotional states on the basis of landmark positions and all the collected data will be integrated in a Bayesian framework.
The focus will be on three specific case-studies related to three important pathological classes of patients. In particular, we aim at distinguish between fear and angry in adult patients affected by anorexia nervosa and in children at risk for social anxiety disorder; moreover we will discriminate between fear and disgust in patients affected by obsessive compulsive disorder.
The ultimate project's aim is to develop a reliable and robust research tool for emotion recognition, generalizable to other clinical populations, allowing to objectively quantify and identify emotional response.