Can we change depressive beliefs? Modulation of belief updating by ketamine in treatment resistant depression

Treatment resistant depression (TRD) is a tremendous health problem that has encouraged novel treatment strategies involving sub-anesthetic ketamine infusions. While ketamine induces rapid and sustained clinical improvement, its cognitive effects are less well known. Here we tested if ketamine alters the negative expectancies about the future that are insensitive to disconfirming information, and play a key role for the maintenance of depression. Twenty-six TRD patients performed a belief-updating task before and after ketamine treatment, and were compared to a group of thirty healthy controls. The task involved estimating likelihoods of experiencing adverse life events before and after receiving information about the actual occurrence of events in the general public. A computational reinforcement-learning (RL) model tested how much belief updating was explained by learning from positive (better than thought) and negative (worser than thought) estimation errors. Global clinical improvement was assessed by the Montgomery Asberg Depression scale (MADRS). Following a single ketamine infusion patients updated their beliefs about the future more after favorable, and less after unfavorable information, controlled for sequential testing effects by comparison to healthy controls. The emergence of optimism biases in belief updating was underpinned by asymmetric learning from positive rather than negative estimation errors, and formally mediated precocious clinical responsiveness at one week of treatment observed in 19% of the patients. These findings provide novel insight into the cognitive mechanisms of ketamine action in TRD, and through it open the window toward promising perspectives for augmented psychotherapy in mood disorder.