While studies suggested that locus coeruleus (LC) neurodegeneration contributes to sleep-wake dysregulation in Alzheimer's disease (AD), the association between LC integrity and circadian rest-activity patterns remains unknown. Here, we investigated the relationships between 24-hour rest-activity rhythms, cognitive trajectories, and autopsy-derived LC integrity in older adults with and without cortical AD neuropathology. This retrospective study leveraged multi-modal data from participants of the longitudinal clinical-pathological Rush Memory and Aging Project. Indices of 24-hour rest-activity rhythm fragmentation (intradaily variability) and stability (interdaily stability) were extracted from annual actigraphic recordings, and cognitive trajectories were computed from annual cognitive evaluations. At autopsy, LC neurodegeneration was determined by the presence of hypopigmentation, and cortical AD neuropathology was assessed. Contributions of comorbid pathologies (Lewy bodies, cerebrovascular pathology) were evaluated. Among the 388 cases included in the study sample (age at death = 92.1 ± 5.9 years; 273 women), 98 (25.3%) displayed LC hypopigmentation, and 251 (64.7%) exhibited cortical AD neuropathology. Logistic regression models showed that higher rest-activity rhythm fragmentation, measured up to ~7.1 years before death, was associated with increased risk to display LC neurodegeneration at autopsy (odds ratio [OR] = 1.46, 95% confidence interval [CI95%]: 1.16-1.84, pBONF = 0.004), particularly in individuals with cortical AD neuropathology (OR = 1.56, CI95%: 1.15-2.15, pBONF = 0.03) and independently of comorbid pathologies. In addition, longitudinal increases in rest-activity rhythm fragmentation partially mediated the association between LC neurodegeneration and cognitive decline (estimate = -0.011, CI95%: -0.023--0.002, pBONF = 0.03). These findings highlight the LC as a neurobiological correlate of sleep-wake dysregulation in AD, and further underscore the clinical relevance of monitoring rest-activity patterns for improved detection of at-risk individuals. ANN NEUROL 2024;95:653-664.