Novel methods to evaluate sleep deficiency in persons with Alzheimer's Disease and Related Dementias
Funded Grant
Overview
Affiliation
View All
Overview
description
ABSTRACT Better methods for the assessment of sleep deficiency (SD) in Alzheimer’s disease and related dementias (ADRD) are urgently needed. Over half of persons with ADRD suffer from SD, which includes abnormalities in sleep duration, quality, and/or circadian alignment of sleep with biologic night. SD accelerates cognitive and functional decline, while existing interventions may slow decline. Unfortunately, there are no feasible, accurate methods to assess SD in persons with ADRD. Self-report is not accurate, and gold standard objective methods (i.e., polysomnography and dim light melatonin [DLMO] assays) are burdensome. Actigraphy is an alternative for sleep and rest-activity measures, with the latter often substituted for DLMO assays. However, actigraphy does not measure sleep architecture (e.g., slow wave sleep) and rest-activity measures have low correlations with DLMO. Finally, polysomnography is usually assessed over one night and may not capture habitual sleep patterns. Wearable multi-sensor headbands (MSHBs; with electroencephalography, movement, and heart rate) are easily used over multiple nights and may better assess habitual sleep patterns. Unfortunately, most have proprietary algorithms developed using data from young, healthy adults without sleep disturbances. To advance research and clinical care, algorithms that perform well among persons with ADRD and sleep disturbances and are publicly available are needed. Actigraphy, though suboptimal for measures of sleep on its own, has high resolution accelerometry and light exposure data which, when combined with MSHB data, is likely to improve prediction of sleep stages and circadian alignment. The goal of this proposal is to validate novel algorithms that feasibly evaluate SD in persons with ADRD and sleep disturbances. We will examine 120 persons with ADRD in a 9-day protocol at baseline and follow up over 2 years. At the baseline visit, we will perform the Clinical Dementia Rating [CDR], Vascular/Alzheimer’s Disease Assessment Scale–Cognitive Subscale [VADAS- Cog]), and brain MRI. Over the next 9 days, subjects will wear a MSHB for 7 nights, an actigraph for 7 days and nights, complete a home-based polysomnogram, and have salivary melatonin assays to identify DLMO. We will repeat CDR and VADAS-Cog yearly, and MRI at year 2. Using a MSHB and actigraphy, with polysomnography as the gold standard, we will develop and validate algorithms for sleep stage classification that have equivalent or better performance than proprietary algorithms (Aim 1). Using a MSHB and actigraphy, with DLMO assays as the gold standard, we will develop and validate algorithms to measure circadian alignment that have lower prediction error for phase angle than standard rest-activity measures (Aim 2). Finally, we will determine whether cognitive decline and neurodegeneration are better predicted by multi-night algorithm measures than by single- night polysomnography measures of sleep (Aim 3). Using data common across sleep wearables, we will deliver feasible, open-source methods to assess SD accurately and comprehensively, guiding future interventions to improve sleep and prevent adverse outcomes in ADRD, including neurocognitive decline.
