Assessing sleep architecture and cognition in older adults with depressive symptoms attending a memory clinic

Andrea Ricciardiello,Andrew C McKinnon,Loren Mowszowski,Haley M LaMonica,Zoe Menczel Schrire,Carla Haroutonian,Aaron Lam,Ian B Hickie,Angela D'Rozario,Sharon L Naismith
DOI: https://doi.org/10.1016/j.jad.2023.12.032
2024-03-01
Abstract:Background: While depression is intrinsically and bidirectionally linked with both sleep disturbance and cognition, the inter-relationships between sleep, cognition, and brain integrity in older people with depression, especially those with late-onset depression are undefined. Methods: One hundred and seventy-two older adults (mean age 64.3 ± 6.9 years, Depression: n = 66, Control: n = 106) attending a memory clinic underwent a neuropsychological battery of declarative memory, executive function tasks, cerebral magnetic resonance imaging and overnight polysomnography with quantitative electroencephalography. Results: The time spent in slow-wave sleep (SWS) and rapid eye movement (REM) sleep, slow-wave activity, sleep spindles, hippocampal volume and prefrontal cortex thickness did not differ between depression and control and depression onset groups. However, sleep onset latency (p = 0.005) and REM onset latency (p = 0.02) were later in the Depression group compared to controls. Less SWS was associated with poorer memory (r = 0.31, p = 0.023) in the depression group, and less SWS was related to better memory in the control group (r = -0.20, p = 0.043; Fishers r-to-z = -3.19). Limitations: Longitudinal studies are needed to determine if changes in sleep in those with depressive symptoms predict cognitive decline and illness trajectory. Conclusion: Older participants with depressive symptoms had delayed sleep initiation, suggestive of delayed sleep phase. The association between SWS and memory suggests SWS may be a useful target for cognitive intervention in older adults with depression symptoms. Reduced hippocampal volumes did not mediate this relationship, indicating a broader distributed neural network may underpin these associations.
What problem does this paper attempt to address?