Local disruptions in non‐rapid eye movement sleep expression are associated with cerebrovascular pathology and worse memory performance in older adults

Destiny Berisha, Batool Rizvi, Miranda G Chappel-Farley, Abhishek Dave, Novelle Meza, Negin Sattari, Kitty K. Lui, Ivy Y. Chen, Nicholas J. Tustison, Ariel B. Neikrug, Ruth M. Benca, Michael A. Yassa, Bryce A. ManderAlzheimer 2024 · DOI

TL;DR

This study reveals that white matter hyperintensities (WMH), a marker of cerebrovascular damage, are linked to specific disruptions in non-rapid eye movement (NREM) sleep patterns in older adults. These localized sleep disturbances, particularly in lower frequency bands, correlate with poorer memory performance, suggesting a potential mechanism for cognitive decline in aging.

What Problem Does This Paper Address?

As we age, both sleep quality and memory function tend to decline. Non-rapid eye movement (NREM) sleep, crucial for memory consolidation, is particularly affected. Simultaneously, cerebral small vessel disease (CSVD), often manifesting as white matter hyperintensities (WMH) on MRI, becomes more prevalent with age and is known to impact sleep. However, the precise relationship between the location and nature of NREM sleep disruptions and the extent of WMH burden, and how this duo impacts memory, remains unclear. This paper aims to bridge this gap by investigating how localized NREM sleep EEG patterns are associated with WMH burden and memory performance in cognitively healthy older adults.

Methods

The researchers employed a multi-modal approach. They used T2-weighted fluid-attenuated inversion recovery (FLAIR) MRI to visualize and quantify lobar white matter hyperintensity (WMH) volumes using a validated semi-automated method. Overnight polysomnography with high-density 128-channel electroencephalography (EEG) was conducted to capture detailed sleep architecture and brain activity during NREM sleep. A sleep-dependent mnemonic discrimination task was administered to assess memory performance. Crucially, they used threshold-free cluster enhancement (TFCE) with 5000 permutations to perform topographical correlation analyses, allowing them to identify significant associations between WMH burden and specific EEG frequency bands across different brain regions while controlling for multiple comparisons.

Datasets

The study utilized data from fifteen cognitively intact older adults. This cohort underwent MRI, polysomnography, and a memory task. No new datasets were introduced or benchmarked.

Key Insights and Ideas

  • WMH aren't just a general "bad" sign for sleep: The study moves beyond a global association to show that the location of WMH is linked to specific, localized disruptions in NREM sleep EEG.
  • Lower frequencies are particularly vulnerable: The findings highlight that disruptions in slower NREM sleep frequencies (like delta and slow sigma) are more strongly associated with WMH burden, suggesting these fundamental sleep processes are disproportionately affected.
  • Sleep disruptions might be a pathway to memory decline: The paper proposes that the observed localized NREM sleep alterations, driven by WMH, could be a key mechanism explaining why individuals with more cerebrovascular damage experience worse memory performance.

Key Results

  • Older age was associated with reduced posterior total power, slow sigma power, and slow wave activity (SWA) over centro-posterior EEG derivations.
  • WMH burden showed positive associations with frontal and parietal regions, but not occipital or temporal regions.
  • Occipital WMH burden was linked to global reductions in alpha activity, centro-posterior reductions in SWA and Delta power, frontal reductions in total and fast sigma power, global reductions in slow sigma power, and central reductions in theta activity.
  • Better performance on a mnemonic discrimination task (specifically for low similarity lures) was positively associated with higher relative slow oscillation expression over fronto-parietal regions, although this did not survive stringent statistical correction.

Limitations and Future Work

The authors acknowledge the small sample size (n=15) as a limitation. Future studies are recommended to further investigate the role of NREM sleep expression in cognitive decline associated with CSVD burden, potentially with larger cohorts and longitudinal designs to establish causality.

Summary updated

Abstract

Abstract Background Aging is associated with disruptions in non‐rapid eye movement (NREM) sleep and memory decline. Cerebral small vessel disease (CSVD) increases with age and is associated with clinical sleep disturbance, but little is known about its relationship with local expression of NREM sleep. Here, we explore associations between CSVD burden, memory, and local electroencephalography (EEG) measures during NREM sleep in older adults. Methods Fifteen cognitively intact older adults (mean age 71.8±6.3 years, 10 female, Apnea Hypopnea Index (AHI)=7.82±7.95) underwent T2‐weighted fluid‐attenuated inversion recovery magnetic resonance imaging, an in‐lab overnight polysomnography with 128‐channel EEG, and a sleep‐dependent mnemonic discrimination task was administered. Lobar white matter hyperintensity (WMH) volumes were calculated using a validated semi‐automated toolbox. Topographical correlations were run with 5000‐permutation threshold free cluster enhancement. Results Older age was associated with lower absolute posterior total (all p<0.016, r>‐0.61) and slow sigma power (all p<0.018, r>‐0.60) and slow wave activity (SWA) (all p<0.017, r>‐0.61) over centro‐posterior EEG derivations. Age was also positively associated with frontal (τ=0.51, p<0.01) and parietal (τ=0.49, p<0.01) but not occiptal (τ=0.33, p=0.092) or temporal (τ=0.36, p=0.064) WMH burden. Occipital WMH burden was also not associated with AHI (τ=0.038, p=0.843). Occipital WMH burden was associated with global reductions in alpha activity (all p<0.043, τ>‐0.78), centro‐posterior reductions in SWA (all p<0.043, τ>‐0.63) and Delta (all p<0.009, τ>‐0.63), frontal reductions in total (all p<0.043, τ>‐0.71) and fast sigma power (all p<0.004, τ>‐0.57), global reductions in slow sigma power (all p<0.043, τ>‐0.71), and central reductions in theta activity (all p<0.042, τ>‐0.57). Better lure discrimination for low similarity lures was significantly positively associated with higher relative slow oscillation expression over fronto‐parietal regions, though this did not survive TFCE correction (all p<0.033, r>0.55). Conclusions Collectively, these findings indicate that WMH burden in older adults is associated with local disruptions in NREM sleep expression in multiple frequency bands, of which lower frequencies were particularly vulnerable. This in turn may explain weakened memory performance. Future studies should examine the role of NREM sleep expression in cognitive decline associated with CSVD burden.
BibTeX
@inproceedings{berisha2024local,
  author = {Destiny Berisha and Batool Rizvi and Miranda G Chappel-Farley and Abhishek Dave and Novelle Meza and Negin Sattari and Kitty K. Lui and Ivy Y. Chen and Nicholas J. Tustison and Ariel B. Neikrug and Ruth M. Benca and Michael A. Yassa and Bryce A. Mander},
  title = {{Local disruptions in non‐rapid eye movement sleep expression are associated with cerebrovascular pathology and worse memory performance in older adults}},
  booktitle = {Alzheimer},
  year = {2024},
  doi = {10.1002/alz.090118},
}
Figure