1 Sleep, Cognition and Neuroimaging Laboratory, Concordia University, Montreal, Canada.
2 Centre de Recherche de l'Institut Universitaire de Gériatrie de Montréal (CRIUGM), Montréal (Québec), Canada.
3 Nuffield Department for Primary Care Health Sciences, University of Oxford, Oxford, UK.
4 Unitat de Trastorns Cognitius, Cognition and Behavior Study Group, Hospital Universitari Santa Maria Universitat de Lleida, Lleida, Spain.
5 Translational Research in Respiratory Medicine (TRRM), Hospital Universitari Arnau de Vilanova-Santa Maria, Biomedical Research Institute of Lleida (IRBLleida), Lleida, Spain.
6 Alzheimer's Disease and Other Cognitive Disorders Unit, Neurology Service, Hospital Clínic de Barcelona, Fundació de Recerca Clínic - Institut d'Investigacions Biomèdiques August Pi i Sunyer (IDIBAPS), Barcelona, Spain.

Introduction: Changes in sleep physiology can predate cognitive symptoms by decades in persons with Alzheimer's disease (AD), but it remains unclear which sleep characteristics predict cognitive and neurodegenerative changes after AD onset.

Methods: Using data from a prospective cohort of mild to moderate AD (n = 60), we analyzed non-rapid eye movement sleep spindles and slow oscillations (SOs) at baseline and their associations with baseline amyloid beta (Aß) and tau and with cognition from baseline to 3-year follow-up.

Results: Higher spindle and SO activity predicted significant changes in Aß and tau at baseline, lower Alzheimer's Disease Assessment Scale Cognitive Subscale (better cognitive performance) score, and higher Mini-Mental State Examination score from baseline to 36 months. Spindles and SOs mediated the effect of phosphorylated tau 181 (pTau181)/Aß42 on cognition, while pTau181/aß42 moderated the effect of spindles and SOs on cognition.

Discussion: Our findings demonstrate that spindle and SO activity during sleep constitute predictive and non-invasive biomarkers of neurodegeneration and cognition in AD patients.

Highlights: Sleep spindles predict long-term cognitive performance in AD. Sleep spindle and SOs can be predictive, non-invasive biomarkers for AD. Sleep may be one of the most important modifiable risk factors for AD progression. Sleep microarchitecture is a novel therapeutic target for preserving brain heath. Sleep physiology can provide novel therapeutic targets to slow AD progression.