A mania-related memory bias is associated with risk for relapse in bipolar disorder<sup/>.

OBJECTIVE: Previous work has shown that neuropsychological performance can predict outcome of psychotherapy. The present paper explores whether an affective bias in verbal memory is associated with recurrence of mood episodes in patients with Bipolar Disorder (BD).

METHOD: 76 euthymic adult patients with BD were randomly assigned to either 9 months of Cognitive Behavioral Therapy (CBT) or Support Therapy (ST), and followed up for 2 years after completing therapy. At baseline, affective learning memory (Emotional Auditory Verbal Learning Test; EMO-AVLT) and other measures were assessed. Recurrence of a mood episode was the primary outcome.

RESULTS: The survival analyses revealed that the interaction between therapy condition, more specifically ST, and a recognition bias in favor of mania-related, but not depression-related words predicted recurrence of mania. Recurrence of depression was predicted by neither affective memory bias nor their interaction with treatment.

CONCLUSIONS: A mania-related memory bias emerged as a predictor of mania recurrence, specifically in an unstructured setting such as ST. Perhaps mania-related schemata are more salient or more easily activated in those at high risk for recurrence. Interventions targeting patients' insight into their internal states as potential indicators of prodromal manic symptoms could be the key to improve the outcome of psychological interventions in BD. Additional research in the role of cognitive factors in relapse prevention is warranted.