A Longitudinal Examination of Depression Among Older Adults: The Role of Working Memory and Sleep

Introduction: The development and regulation of depressive symptoms and the ability to regulate
their development is a complex process. Both working memory and sleep disturbance relates to
depressive symptom endorsement, though the mechanisms relating these variables have not been
examined longitudinally. The current manuscript contains a series of three interrelated studies
that aim to elucidate the relationship between potential emotion regulation resources
longitudinally within the context of the Selection, Optimization, and Compensation of Emotion
Regulation (SOC-ER) model. Study one examined the temporal relationship between working
memory and depression, study two examined working memory and depression following loss of
spouse, and finally, study three examined the relationship between sleep quality, working
memory, and depressive symptoms.
Method: Participants were drawn from the Health and Retirement Study, which is a longitudinal
dataset sponsored by the National Institute on Aging and collected through the University of
Michigan. Data have been collected every two years since 1992 and consists of randomly
selected participants ages 65 years and older (Hauser & Willis, 2004). Working memory was
measured by Serial 7’s, and the 8-question CES-D was used to measure depressive symptoms.
Results: Study one utilized a latent growth model to evaluate the relationship between working
memory and depressive symptoms over time. A significant bidirectional rather than a temporal
relationship between the two was observed. Furthermore, both depressive symptoms and
working memory ability was found to become worse over time. Study two utilized a latent
growth model of the trajectory of depressive symptom development following the loss of a
spouse. Results indicated that the starting point of initial depressive symptom endorsement was
significantly related to working memory ability. Working memory also moderated the
relationship between depressive symptom endorsement and time, where individuals with better
working memory tended to report lower depressive symptoms and demonstrated a lesser increase
in depressive symptoms. Study three utilized a multilevel model that demonstrated depression
increases over time and with age. Regardless of time, better sleep quality and better working
memory both result in lower depressive symptom endorsement, and there were associations
between lower depressive symptom endorsement and both better sleep quality and better
working memory.
Conclusions: Findings strongly support working memory, sleep quality, and spousal support as
emotion regulation resources within the context of the SOC-ER model. Future research should
continue to examine these and similar interrelated factors such as inhibitory control, processing
speed, and vascular burden longitudinally to provide further understanding of changes in
emotion regulation processing among older adults.