36. ANOTHER DUALISM: COGNITIVE VERSUS PERSONALITY PREDICTORS OF RUMINATION IN LATER LIFE DEPRESSION

Introduction

Rumination, a predisposing and perpetuating feature of depression, has been linked to personality and cognitive variables. However, little is known about their comparative influence on rumination. This study examined personality and cognition as predictors of rumination in early-onset, later life depression (LLD), for those with and without a history of depression. Exploratory analysis investigated this relationship in the subset of patients in a current depressive episode (state depression).

Methods

Participants were 76 adults aged 55-79 divided into two groups: “ever depressed” (N=35) including those with current or remitted depression, and “never depressed” (N=41) including those with no history of depression. A subset of the “ever depressed” group in a current depressive episode (N=18) were considered in the exploratory state depression analysis. Personality predictors were the five domains assessed by the NEO Personality Inventory – Revised (NEO-PI-R), openness to experience, conscientiousness, extraversion, agreeableness, and neuroticism. Cognitive predictors were selected based on theory and existing support in the literature for their association with rumination, and included measures of executive function (Trail Making Test B, D-KEFS Color-Word Interference Condition 3, Wisconsin Card Sorting Test Perseverative Responses), attention (WAIS-IV Digit Span Forward), and global cognitive ability (Mini Mental Status Exam).

The omnibus analysis was conducted using hierarchical linear regression modelling for predictors of rumination, with variable blocks composed of 1) demographic factors (age, sex, education), 2) depression history (ever/never depressed), 3) all cognitive and personality variables, and 4) interaction terms (depression history x variable) for WCST Perseverative Errors and NEO-PI Neuroticism, the only variables with significant interaction effects in Block 3. Exploratory analysis considered whether results were influenced by depression severity for those in a current depressive episode, with hierarchical regression variable blocks of 1) demographic factors (age, sex, education), 2) MADRS score, and 3) WCST Perseverative Responses and NEO-PI Neuroticism Factor Score, along with their interaction terms based on depression severity (MADRS total score x variable).

Results

Depression history was a significant predictor of rumination in the primary analysis and remained significant following inclusion of all cognitive and personality variables. Neuroticism (β=.39, p LESS THAN .001) and WCST Perseverative Responses (β=0.18, p LESS THAN .05) significantly predicted rumination, though their interaction terms with depression history were nonsignificant. After inclusion of the interaction terms, MMSE emerged as a significant predictor of rumination (β=.17, p LESS THAN .05), despite initial nonsignificance.

For the exploratory analysis, MADRS score significantly predicted rumination in those with state depression, accounting for 55.7% of variance beyond the covariates considered (R2=.69, F(1, 13)=23.2, p LESS THAN .001). After inclusion of Neuroticism Factor Score, WCST Perseverative Responses, and their interaction terms with MADRS score, none of the included predictor variables were significantly associated with rumination levels.

Conclusions

Depression history was predictive of rumination in LLD after accounting for various sociodemographic variables, affirming prior research that indicates trait depression confers an increased risk of rumination in older adults. Neuroticism was significantly predictive of rumination, indicating personality may increase vulnerability to depressive symptoms regardless of depression history. Of the cognitive domains considered, only cognitive flexibility (WCST Perseverative Responses) was associated with rumination. Equivocal findings regarding the contribution of MMSE performance to rumination levels may suggest more complex underlying relationships between personality and cognition in LLD, though this association requires further study.

Exploratory analysis of the state depression group did not replicate the salience of those predictors that were significant in the omnibus analysis, suggesting predictors of rumination may vary for state- and trait-level depression. Overall, results implicate personality factors, namely neuroticism, as more predictive of rumination than cognition in older adults, though the relationship may be distinct for state and trait depression.