The effects of cognitive remediation on cognitive abilities and real-world functioning among people with bipolar disorder: A systematic review

Background: Bipolar disorder (BD) is characterized by cognitive impairments that are known to predict psychosocial functioning and quality of life. While cognitive remediation (CR) was originally developed to directly target cognitive symptoms in traumatic brain injury and psychotic illnesses, the efficacy of CR in BD has begun to emerge only in the last decade. Functional Remediation (FR) is an integrated intervention that has been developed to restore psychosocial functioning by means of ecological neurocognitive techniques that involve psychoeducation about cognitive dysfunctions and their impact on the general functioning. Because of the heterogeneity of treatment targets and mechanisms of actions, here we aim to illustrate the effects induced by existing CR/FR approaches in BD. Methods: In this systematic review, we evaluated cognitive and functional outcomes after CR/FR in studies conducted in BD. Results: Eleven studies met inclusion criteria: 3 RCTs that compared CR/FR to one or more control condition (n = 354), 5 secondary analyses that further examined data from these trials, 2 single-arm studies, and 1 naturalistic study. While features such as the use of computerized training tools and a group-based format recurred across studies, CR/FR paradigms targeting different cognitive and functional domains showed specificity of training focus to outcomes. Effect sizes were in the medium-large range, suggesting that patients with BD respond to treatment at or above the level reported in psychotic patients. Integrated approaches that combined cognitive exercises with group-based experiences were associated with both cognitive and functional improvements. Conclusions: In this review, we found support for the use of CR/FR paradigms in patients with BD with evidence of cognitive and functional improvements. The scarcity of currently published RCTs as well as of data examining mechanisms of action and neural correlates limits the generalizability of our findings.