The present study aims to construct a multi-scale database, consisting of human facial expression video recordings, speech, and physiological signals, after experimental induction of emotional states. The collected data will be used to develop computational methods capable of recognizing non-verbal signs, which can accurately detect facial signs of clinical depression. The method aspires to increase recognition sensitivity of depressive signs, in order to aid in early diagnosis and relapse prevention.
The study is divided in two parts. Study-1, which recruits mentally healthy adult individuals 18 – 60 years old, in order to assess the within-group performance and reliability of the algorithm, with regard to emotion recognition. Study-2 will apply the experimental protocol to an age-matched group of persons carrying a primary diagnosis of depressive disorder.