Detecting Depression Using Auditory and Linguistic Indications Poojan Patel 1 , Navin Agarwal 2 , Prof. Rakhee Kundu 3 Computer Engineering Deaprtment, Mukesh Patel School of Technology and Management- NMIMS University 1 poojan17.nmims@gmail.com 2 official.navin.agarwal@gmail.com 3 rakheekundu@rocketmail.com Abstract— Depression is known to result in neuro-physiological and neurocognitive changes that affect control of motor, semantic, and cognitive functions. In this paper, biomarkers are derived from all of these techniques, drawing first from previously developed neurophysiological motivated auditory and outer coordination and timing features. In addition, a unique indicator of lower vocal tract constriction in articulation is assimilated that relates to vocal projection. Linguistic features are analyzed for content using a skimpy coded rhetorical embedding space, and for circumstantial clues related to the individuals current or previous depression condition. Keywords— Depression classification, Multimodal, decision tree, speech synthesis, Task functional magnetic resonance imaging, Public Heath Questionnaire. 1. INTRODUCTION According to world health organization [1] depression affects more than 300 million people every year and it is one of the most common mental disorder.Depression affects between 5% and 10% of individuals in primary care but is only diagnosed in around 50% of cases. Similar type of problems was also found in general hospital settings, where there is substantial under identification and unmet need for psychological health services. Depression is different with respect to the usual mood fluctuations and short-lived emotions to challenges in everyday life. When the condition of emotional imbalance last for a longer or a moderate period of time also it needs special attention as it might be depression. A person suffering from depression can affect his personal and professional life adversely and leave a negative impact everywhere. The worst thing that a person suffering from depression can do is harm an individual or even attempt suicide. Nearly 800000 people attempt suicide yearly and major suicidal attempts are done by age group 15-29 years. In the current paper we are speaking about the major technique used to detect depression with the dataset provided by the Audio/Visual Emotion Challenge (AVEC) 2016 and Audio/Visual Emotion Challenge (AVEC) 2017 along with the data developed during the Distress Analysis Interview Corpus (DAIC) 2. BACKGROUND The Distress Analysis Interview Corpus (DAIC) is a collection of sub structured clinical interviews. Designed to simulate usual protocols for identifying people at risk for post-traumatic stress disorder (PTSD) and depression, these interviews were generated as part of a bigger effort to create a computer that interviews people and identifies verbal and nonverbal indicators of mental illness (De Vault et al., 2014). The corpus contains four types of interviews: Face-to-face interviews between participants and a human interviewer Teleconference interviews, conducted by a human interviewer over a teleconferencing system; Wizard-of-Oz interviews, conducted by an animated virtual interviewer called Ellie, controlled by a human interviewer in another room; Automated interviews, where participants are interviewed by Ellie operating as an agent in a fully automated mode. CIKITUSI JOURNAL FOR MULTIDISCIPLINARY RESEARCH Volume 5, Issue 11, November 2018 ISSN NO: 0975-6876 http://cikitusi.com/ 43