Sleep Variability and Cognitive Decline in Aging

Introduction An estimated 6.07 million older adults in the US in 2020 had Alzheimer's disease, the most common cause of dementia. With the number of patients doubling every five years worldwide, dementia is a growing public health problem. There is no cure for dementia, but there is promising early evidence of prevention through improving lifestyle factors like sleep. Shorter and longer total sleep time (TST), longer wake after sleep onset (WASO), and lower sleep efficiency have been associated with worse cognitive decline and higher dementia risk. Previously published studies usually assessed mean sleep measures rather than night-to-night changes (i.e., intraindividual variability or IIV). Sleep IIV has been associated with increased pain, neurological problems, depression, and anxiety, but relationships of sleep IIV with cognition have not been studied in older adults. We conducted a cross-sectional study of older adults using wearable sleep sensors and neurocognitive tests. We hypothesized that (1) older age will be associated with increased cognitive impairment, worse mean sleep measures, and increased sleep IIV, (2) worse mean sleep measures will be associated with increased cognitive impairment, and (3) higher sleep IIV will be associated with increased cognitive impairment. Methods The study sample included 60 older adults, age range of 68-100 years, residing independently in San Diego County. Mean and IIV (standard deviation) of sleep measures (TST, WASO, sleep efficiency, bed-time, and wake-time) were collected via wrist-worn fitness trackers. The cognitive measures included global cognitive screening (Montreal Cognitive Assessment or MoCA), attention (Wechsler Adult Intelligence Scale-IV– Digit span), and verbal memory (Hopkins Verbal Learning Test – Revised). Spearman's correlations were used to examine age relationships with sleep and cognitive measures. Mann-Whitney U tests were used to compare sleep measures between subgroups based on cognitive impairment. General linear models were used to examine associations between sleep and cognitive measures. Results Older age was significantly correlated with earlier mean bed-time (rs=-0.26, p=0.049), later mean wake-time (rs=0.40, p=0.001), more impaired MoCA scores (rs=-0.42, p=0.006), and greater deficits in verbal memory (rs=0.40, p=0.002), but not with sleep IIV measures or attention scores. The subgroup with impaired attention had significantly higher mean WASO (p=0.034, d=0.28), than the subgroup without attention deficits. The subgroup with impaired verbal memory had significantly earlier mean bed-time (p=0.007, d=0.35) and later mean wake-time (p=0.025, d=0.29), compared to the participants without verbal memory impairments. No differences in sleep measures were found for subgroups based on MoCA score impairment. Poor attention was significantly associated with higher mean and higher IIV WASO (B=0.003, p=0.002, ηp2=0.167 and B=0.003, p=0.001, ηp2=0.192 respectively) as well as with lower mean and higher IIV sleep efficiency (B=-0.012, p=0.032, ηp2=0.087 and B=0.013, p=0.031, ηp2=0.088 respectively), adjusting for race, sex at birth, years of education, and age. Sleep measures (mean and IIV) were not associated with global cognitive impairment or impairment of verbal memory. Conclusions Age was associated with sleep-wake cycle differences (bed-time and wake-time) rather than overnight sleep duration or disturbances. Age was correlated with worse MoCA scores and verbal memory but not attention. This is consistent with previous studies that showed an age-related decline of certain cognitive domains (e.g., memory), but not others (e.g., attention). Both mean and IIV measures of sleep disturbances were associated with impaired attention independent of age. Despite using a homogenous sample of modest size, this study suggests a benefit from passive sleep-sensing to identify patients with high mean or IIV sleep disturbances for cognitive testing or interventions. Our results highlight sleep as a target to improve certain domains of cognition in older adults. Future studies are needed to understand how sleep interventions might improve cognition and how sleep impacts other domains of cognition. This research was funded by This work was supported by the • IBM Research AI through the AI Horizons Network and IBM-UCSD Artificial Intelligence for Healthy Living Center. • National Institute on Aging T35 grant AG26757 (PI: Dilip V. Jeste, MD and Alison Moore, MD, MPH), and the Stein Institute for Research on Aging and the Center for Healthy Aging at the University of California, San Diego. • National Institute of Mental Health K23 grant K23 MH119375-01 (PI: Ellen Lee, MD) • National Institutes of Health Grant NIH UL1TR001442 of CTSA (PI: Gary Firestein, MD).