0804 Blood Lipids Moderate the Associations Between the Proportions of Specific Sleep Stage Durations and Inhibitory Control in Older Black Adults

Abstract Introduction Sleep disturbances impact cognitive performance. Each sleep stage may uniquely impact different cognitive functions, but findings have been inconsistent. Cardiovascular disease (CVD) influences sleep and cognition, particularly executive function; yet, whether cardiovascular risk moderates sleep durations and executive function remains unclear. Therefore, we tested if CVD risk biomarkers moderated sleep stage durations/percentages and executive function in older Blacks, who are disproportionately affected by sleep disturbances and CVD. Methods Eighty-three Black adults, aged ≥55 years, participated in an eight-day-long observational study (NIA-funded MOSAIC study, n=51) or a six-month-long randomized controlled trial (NIA-funded PRAISE study, n=32). Data from both studies were harmonized in accordance with a standardized protocol. At the baseline visit, fasting serum cardiovascular risk biomarkers (total cholesterol, high-density lipoproteins (HDL)) were collected and quantified, from which total cholesterol:HDL ratios (TC:HDL-c) were calculated. Each participant was instructed to wear a Fitbit Inspire 2 wrist-worn device at least 4 consecutive days; thereafter, averaged durations/percent durations of each sleep stage were obtained from the Fitbit device. Executive function was characterized using the Dimensional Change Card Sort (DCCS, which tested cognitive flexibility) and Flanker (which tested inhibitory control) tasks in NIH Toolbox. An age-/education-adjusted T-score value was generated from each neurocognitive task. Hierarchical moderated regression analysis tested whether cardiovascular risk markers moderated the associations between sleep architecture durations/percentages and executive function in older Black adults. Results %N1/%N2, %REM, and TC:HDL-c were each independently associated with Flanker T-scores (p’s< 0.0071). The interactions between light non-REM percentages (%N1/%N2; ΔR2=0.11, ΔF=10.73, p=0.0016) and %REM (ΔR2=0.12, ΔF=12.06, p=0.0008) and TC:HDL-c were significant. Specifically, those with low TC:HDL-c ratios showed better inhibitory control performance when higher proportions of REM sleep and lower proportions of N1/N2 sleep were obtained, whereas the reverse was true in those with high TC:HDL-c ratios. In contrast, TC:HDL-c ratios did not moderate the links between sleep durations/percentages and DCCS T-scores. Conclusion The proportions of light non-REM and REM sleep durations, but not N3 or TST, were associated with inhibitory control performance, which varied by levels of cardiovascular risk in older Black adults. Future work should investigate these associations in other neurocognitive domains. Support (if any) R01AG075007 (PRAISE); R01AG067523 (MOSAIC); T32HL166609