Access Type

Open Access Thesis

Date of Award

January 2013

Degree Type


Degree Name




First Advisor

Lisa J. Rapport


Individuals with Major Depressive Disorder (MDD) frequently experience cognitive decrements in addition to mood impairments. Ironically, antidepressant medications used to treat depression may have adverse effects on cognitive functioning. It is imperative to understand the relative cognitive costs of antidepressants when considering the treatment of MDD patients. Furthermore, observational studies of depression are challenged by problems of systematically confounded variables. Researchers are often faced with difficulties in managing this issue and opt to either ignore the problem, alter their sample, or use inappropriate statistical methods (e.g., Analysis of Covariance) due to a limited understanding of acceptable solutions. It is important to provide researchers with the access to general knowledge of methods to manage systematic covariates in order to aid in effective and appropriate decision-making.

Participants included 178 adults with MDD (89 unmedicated-uMDD, 89 medicated- mMDD) and 153 healthy comparison (HC) adults who were evaluated for performance in several domains of cognition. The three groups were significantly different on age (mMDD group was oldest) but equivalent on education. The MDD groups were equivalent in years of illness, differed on depression severity (uMDD more severe than mMDD).

Univariate analyses of variance (ANOVAs) indicated that performance among the groups was significantly different for measures of processing speed, attention, and some aspects of the learning and memory domain. However, the groups were generally equivalent on measures of working memory and executive functioning domains. Post hoc analyses using simple contrasts indicated that the significant were generally driven by worse performance for mMDD groups compared to uMDD and HCs. Exploratory analyses accounting for age and education included Analysis of Covariance (ANCOVA), Matched-samples, Regression-residuals, and Propensity Score Analysis (PSA). Overall, the pattern of findings was constant across methods, with relatively stable effect sizes for significant results. However, effect sizes for general (significant and non-significant) results varied by method.

Antidepressants used to treat MDD improve dispositional symptoms of depression but have adverse effects on some areas of cognition. Comparing the effect sizes may be useful to clinicians and researchers in distinguishing cognitive effects of medication separately from those associated with MDD. Furthermore, employing the proper methods to address systematic confounds (e.g., age and education) are essential to maintain sound research and to promote effective antidepressant treatment. Statistical solutions that are theoretically acceptable to use in order to address systematic covariates are discussed. Both the beneficial effects on depressive symptoms and adverse effects on cognition should be considered by practitioners and consumers when making treatment decisions.

Included in

Psychology Commons