Core cognitive impairments and their neural correlates in mood disorders

Abstract

Background: Mood disorders such as bipolar disorder (BD) and major depressive disorder (MDD) are associated with impairments in a wide range of cognitive functions compared to healthy controls (HCs). However, it is unclear whether the this reflects a general cognitive impairment, or whether specific primary impairments are responsible for wider dysfunction. Processing speed (PS), sustained attention (SA), and executive function (EF) may be particularly impaired in mood disorders and may affect wider cognition, creating a hierarchy of cognitive dysfunction. Mixed findings in the literature are partly due to heterogeneity of neuropsychological assessment across studies. Studies rarely account for relationships between cognitive functions or investigate the presence of a cognitive hierarchy. Cognitive impairments are also thought to relate to structural brain abnormalities in mood disorders, but whether specific functions are related to distinct brain areas is not clear and much of the research on brain-cognition associations is limited by univariate analysis. Methods: A systematic review of k=103 studies was conducted to examine the presence and magnitude of PS and SA impairments in people with BD and MDD compared to HCs. Data were meta-analysed for each neuropsychological test score separately and subgroup analysis was performed across mood states, where possible. Hierarchical regression was used to examine the role of PS, SA, and EF in memory in euthymic BD (n=62), BD depression (n=43), and MDD (n=41) compared to matched HCs (n=142), controlling for age and premorbid IQ. Ex-Gaussian distributional parameters were obtained from continuous performance test reaction times to conceptualise SA. Network graphs were used to illustrate interrelationships between cognitive functions within each group. Canonical correlation analysis was used to assess multivariate associations between abnormal cortical thickness and core cognitive impairments in euthymic BD (n=56), controlling for age, sex, and premorbid IQ. Results: The meta-analysis suggested that BD and MDD show impairments in PS and SA across most neuropsychological tests. Impairments were present in both symptomatic states and in euthymia in most cases, however, some outcome measures were not impaired in euthymia. PS and EF appeared to explain memory impairments in euthymic BD, whereas SA appeared to contribute to memory impairments in MDD. Memory did not seem to be impaired in depressed BD, however, SA appeared to contribute to memory performance. Impairments in

PS, attention, and EF were associated with abnormal cortical thickness in the PCC, superior temporal, parahippocampal, right entorhinal and right lateral occipital areas in euthymic BD. Discussion: PS and SA appear to be impaired in mood disorders, however, more research is needed to investigate the nature of these impairments in BD and MDD in different mood states, controlling for clinical confounds. PS and EF may be primary impairments in euthymic states, whereas SA plays a role in cognitive functioning in depressed states. Our results highlight potentially important relationships between cognitive functions; further research is needed to unravel the precise cognitive profile in each diagnostic group and how this varies between mood states. Core cognitive dysfunction may be associated with abnormal cortical thickness in several brain regions in BD, including some regions implicated in the default mode network. Future research should further explore brain-cognition associations using multivariate analysis and should account for covariance between different brain structural morphological features, such as cortical thickness, cortical volume, and surface area