AbstractBackgroundSleep disturbances are increasingly acknowledged as a preclinical marker for Alzheimer’s disease (AD), however, the mechanisms underlying the association between sleep disturbances and dementia remain unclear. Sleep disturbances may affect AD through brain morphology changes and vascular damage, as sleep disturbances are known to be linked with altered grey‐ and white‐matter integrity. Moreover, sleep and brain changes may be linked with cortisol levels. Sleep disturbances have been proposed to parallel cortisol dysregulation, and older individuals with preclinical AD have altered awakening cortisol, a feature that is intensified among people also suffering from sleep disturbances.MethodThe study is based on 146 memory clinic patients from the Karolinska University Hospital memory clinic (Sweden). Participants are in the age range 47‐86 years and diagnosed with either subjective cognitive impairment or mild cognitive impairment. Self‐reported sleep disturbances are measured by the Karolinska Sleep Questionnaire. Brain imaging (MRI or CT) was used to semi‐quantify brain atrophy using four visual rating scales: Medial Temporal Atrophy, Global Cortical Atrophy, Koedam (posterior brain atrophy) and Fazekas (white‐matter lesions) scales. Lastly, saliva cortisol has been sampled 6 times daily to measure diurnal cortisol patterns.ResultOrdinal logistic regressions show significant associations between sleep disturbances and brain parameters. Insufficient sleep (OR = 1.30, p = 0.048) is associated with greater odds of white‐matter lesions, and apnea (OR = 1.26, p = 0.036) and snoring (OR = 1.38, p = 0.045) are associated with greater odds of posterior brain atrophy. Increased daytime sleepiness is associated with both reduced awakening cortisol (β = ‐0.03, p = 0.025) and a flattened cortisol curve (β = ‐0.04, p = 0.019). Apnea is associated with reduced awakening cortisol (β = ‐0.06, p = 0.013). However, cortisol was not found to significantly interact with sleep in predicting brain outcomes.ConclusionThese results show that various sleep disturbance parameters (short sleep duration, apnea, and snoring) are associated with neuroimaging correlates, while daytime sleepiness and apnea are associated with markers of dysregulated cortisol patterns. The results suggest that neuroimaging correlates and cortisol measures may clarify the mechanisms underlying the associations between sleep disturbance and cognitive impairment, and the combination of objective sleep measures and neuroimaging markers in future studies may further reveal the mechanisms.