Depression is a debilitating illness with a risk of developing a treatment resistant form. Currently, diagnosis is purely clinical with little features available to identify potentially adverse developments. Clinical features such as early onset age, prolonged episodes, anxiety, somatic symptoms and apathy are all arguments raising fears the onset of resistance to conventional treatments. According to neuroimaging knowledge about the pathophysiological mechanisms, involving front-limbic functional networks supporting the functions of emotional regulation and reward system, recent work has focused on the identification of neuroimaging biomarkers predicting therapeutic response. Among the regions of interest identified, the anterior cingulate cortex, amygdala, hippocampus, and regions participating in the Default Mode Network Training (Medial prefrontal cortex, posterior cingulate cortex, inferior parietal lobe) are most frequently areas associated with the prediction of therapeutic response. Limitations most reported in these studies are the heterogeneity of experimental paradigms (resting state, cognitive or emotional functional tasks), imaging (PET scan, MRI) the heterogeneity of clinical resistance criteria forms studied, different techniques (as that we consider remission (threshold score) or response (50% decrease from baseline score), and the sample size. Knowing that MRI into daily clinical practice in the SHU of Adult Psychiatry, as the balance sheet of the disease, monitoring its evolution, as in assumption rTMS (repetitive Transcranial Magnetic Stimulation) (pretreatment assessment and neuro), the identification of neuroimaging biomarkers in a population of patients with clinical criteria of Mood Depressive Episode, with an acquisition of identical and reproducible image daily routine methodology appears to be a reliable way to correct the heterogeneity of conventionally published studies on the topic. This study aim to identify, in routine care, predictive clinical and neuroimaging markers of poor outcome in major depressive disorder.

The main objective is to assess neuroimaging biomarkers like cortical thickness predictive of poor outcome in major depressive disorder.

Secondary objectives are :

• Identifying neuroimaging biomarker like Cerebral Blood Flow (CBF) predictive of poor outcome in major depressive disorder;
• Identifying morphological biomarker like volumetric abnormalities (Voxel Based Morphometry - VBM), other than the cortical thickness, predictive of poor outcome in major depressive disorder;
• Identifying clinical features (sociodemographic, clinical dimensions such as apathy or anxiety) predictive of poor outcome of depression.

This study expect to assess on a large population of patients:

• Clinical and neuroimaging markers (morphological and perfusion) predictive of poor outcome in major depressive disorder, allowing early identification of patients at risk of poor therapeutic response. These markers should allow a better stratification of patients;
• A better characterization of pathophysiological processes involved in major depressive disorder at different stages of the illness;
• Development of innovative technologies such as treatment with repetitive transcranial magnetic stimulation, or neurofeedback using real time fMRI, on both aspects of evaluation of the effectiveness and optimization of procedures.

Such a study would open up on thinking in terms of therapeutic management. Indeed, recurrent and potentially resistant forms and screened using such predictive neuroimaging biomarkers could serve more specific therapeutic approaches in a preventive approach.

Inclusion Criteria:

• Men and women aged over 18 years;
• Patient with a diagnosis (according to DSM IV criteria measured at MINI) for Major Depressive Episode, and / or recurrent depressive disorder unipolar or bipolar , or chronic and resistant depression (according to the criteria of Thase and Rush);
• Patients will be stratified as the stages of classification and Thase Rush scale widely used to characterize the therapeutic resistance in depression;
• Intensity of EDM with a minimum score of 15 (MADRS);
• Patient in receiving information on the protocol;
• Patient who received information about the protocol and did not express opposition to participate.

Exclusion Criteria:

Related to MRI

• Pacemaker or implantable defibrillator;
• Neurosurgical clips;
• Cochlear implants;
• Metal intra orbital or encephalic foreign bodies;
• Stents placed for less than four weeks and osteosynthesis material posed for less than six weeks;
• Claustrophobia.

Other criteria

• Pregnant or lactating women;
• Hemodynamically unstable acute respiratory failure , a general poor condition or a need for monitoring incompatible with the constraints of MRI ;
• Legal protection ;
• Patients hospitalized without their consent.

Patient with a diagnosis (according to DSM IV criteria measured at MINI) for Major Depressive Episode, and / or recurrent depressive disorder unipolar or bipolar , or chronic and resistant depression (according to the criteria of Thase and Rush)