Temporal Relationship Between Depressive Symptoms and Cognition in Mid and Late Life: A Longitudinal Cohort Study

To examine the bidirectional temporal relationship between depressive symptoms and cognition in relation to risk, reaction, and prodrome.

Cross-lag analysis of longitudinal data collected online at baseline and 12-month follow-up.

A United Kingdom population cohort of 11,855 participants aged 50 years and over.

Patient Health Questionnaire-9 (depressive symptoms), cognitive measures: Paired Associate Learning, Verbal Reasoning, Spatial Working Memory, and Digit Span.

Depressive symptoms predicted a decline in paired associates learning [β = -.020, P = .013, (95% confidence interval [CI], ‒.036, -.004)] and verbal reasoning [β = -.014, P = .016, (95% CI ‒.025, -.003)] but not vice versa. Depressive symptoms predicted [β = -.043, P < .001, (95% CI ‒.060, -.026); β = -.029, P < .001, (95% CI ‒.043, -.015)] and were predicted by [β = -.030, P = < .001, (95% CI ‒.047, -.014); β = -.025, P = .003, (95% CI ‒.041, -.009)], a decline in spatial working memory and verbal digit span, respectively.

Depressive symptoms may be either a risk factor or prodrome for cognitive decline. In addition, a decline in attention predicts depressive symptoms. Clinical implications and implications for further research are discussed.