Neurobehavioural predictors of depression relapse

• Applicant: Professor Dr. Henrik Walter
• Subject Area: Clinical Psychiatry, Psychotherapy, Child and Adolescent Psychiatry ; Cognitive, Systems and Behavioural Neurobiology
• Term: from 2014 to 2019
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 255342426

We aim to a) identify neuroimaging predictors of a high risk of Major Depressive Disorder (MDD) relapse after antidepressant medication (ADM) discontinuation; and b) examine the effect of medication withdrawal on the remitted depressed state. This is part of an endeavour to use behavioural and neurobiological measures to stratify existing clinical psychopathological entities with respect to treatment outcomes. Current pharmacological depression treatment options lead to eventual remission in up to 70\% of patients. Because the risk of relapse after discontinuation is high (30-60\% in 6 months), guidelines recommend treatment continuation for various periods. However, physicians then face a similar problem again: (i) patients discontinue psychotropic medication independently at very high rates,particularly after achieving remission; and (ii) these recommendations do not take individual variability into account. Markers for safe ADM discontinuation would help identify at-risk patients in whom continuation or further therapy could be recommended on stronger, individually valid, grounds. By providing an objective end-point to treatment this may enhance concordance with treatment. Furthermore, although the neurobiology of affective function after remission has been examined previously, the contribution of ADM remains poorly understood and characterised. We propose a 6-month follow-up study of 76 patients who have been in remission for a minimum of 6 months and intend to discontinue their ADMs independently of this study. We will test the ability of three neuroimaging biomarkers in predicting early relapse, including both novel predictors derived from computational neuroscience, and established ones. Several versions of the former have established validity in predicting response to treatment. The latter has been suggested to be one important characteristic of depression vulnerability. Second, subjects will undergo an established planning task that measures the impact of aversive outcomes on planning. This will be slightly modified to additionally quantify helplessness. All subjects will undergo scanning twice, and will be divided into two groups of equal size. In group 1W2, scan 1 will occur just prior to medication withdrawal, and scan 2 between 5-20 ADM half-lives after withdrawal. In group 12W, both scans will occur before withdrawal: scan 1 approx 5-20 ADM half-lives before, and scan 2 just prior to withdrawal. We will use scan 1 as the main predictor for relapse. We will use the interaction between groups and scans to examine the effect of medication withdrawal. In a subsidiary analysis, we will also use changes between scan 1 and 2 in group 1W2 to predict relapse. In all cases, we will test for incremental predictive power above and beyond clinically available measurements.

• DFG Programme: Research Grants
• International Connection: Switzerland
• Participating Person: Professor Quentin Huys, Ph.D.