Sleep spindles are linked to white matter connectivity differentially in older adults with and without cognitive impairment

Abstract

Abstract While sleep disturbance increases with age and neurodegeneration, the relationship between sleep micro-architecture and changes in key white matter tracts is unknown. We aimed to determine how sleep spindles and slow oscillations relate to structural connectivity within selected white matter tracts in older adults at risk for dementia. In this cross-sectional study, 67 participants (mean age = 68.1; 47 females) with subjective cognitive decline (n = 21) or mild cognitive impairment (n = 46) underwent neuropsychological assessment, magnetic resonance imaging and polysomnography. White matter connectivity of anterior thalamic radiation and superior longitudinal fasciculus subdivisions II and III was quantified using whole-brain tractography, with left and right tracts combined. Primary sleep micro-architecture outcomes included slow (11–13 Hz) and fast (13–16 Hz) spindle density and slow oscillation (0.25–1 Hz) power. Secondary outcomes included: overall spindle density (11–16 Hz) and amplitude. Linear regressions assessed associations between sleep micro-architecture and structural connectivity, adjusting for age and sex, with cognitive classification as a moderator. Multiple comparisons were controlled for using the Benjamini–Hochberg false discovery rate procedure, with q < 0.1 considered statistically significant. The mild cognitive impairment group exhibited lower frontal and central fast spindle density compared to the subjective cognitive impairment group (P < 0.05), with no other group differences in sleep macro-architecture or slow oscillatory activity. After false discovery rate correction, there were no significant associations between slow spindle density, fast spindle density or slow oscillation power and tract connectivity. An interaction between central fast spindle density and superior longitudinal fasciculus III connectivity was observed prior to false discovery rate correction, explaining 8% of unique variance, but did not remain significant after correction (q > 0.10). Exploratory analyses reveal a significant interaction between superior longitudinal fasciculus II connectivity and cognitive classification on spindle amplitude, with greater connectivity associated with higher spindle amplitude in individuals with mild cognitive impairment (P = 0.001), but not in those with subjective cognitive decline (P > 0.05). Overall, white matter connectivity of the anterior thalamic radiation and superior longitudinal fasciculus was not robustly associated with sleep spindle density or slow oscillatory power in this at-risk cohort. However, superior longitudinal fasciculus II connectivity may differentially relate to spindle amplitude in mild cognitive impairment, potentially reflecting altered thalamocortical network engagement in early neurodegeneration.