Sleep disturbance is increasingly recognised as a modifiable contributor to brain ageing, but its relationship with cerebral perfusion across the cognitive continuum remains incompletely defined. In this BIOCIS study of 1,698 participants classified as cognitively normal, subjective cognitive decline, or mild cognitive impairment, sleep quality was assessed using the Pittsburgh Sleep Quality Index, while global and regional cerebral blood flow were quantified using arterial spin labelling MRI. Poorer sleep metrics were associated with regionally specific cerebral hypoperfusion, involving medial frontal, posterior cingulate, hippocampal, pontine, and global grey matter regions. These associations were most widespread among cognitively normal participants, in whom poorer sleep efficiency, shorter sleep duration, daytime dysfunction, and sleep medication use were linked to lower perfusion across multiple networks. In contrast, subjective cognitive decline and mild cognitive impairment showed more selective associations. Plasma pTau181 further moderated the association between daytime dysfunction and posterior cingulate hypoperfusion in cognitively normal adults. Overall, these findings suggest that sleep-related neurovascular vulnerability may be detectable before overt cognitive decline and may interact with early tau-related pathology, supporting the integration of sleep, biomarker, and perfusion imaging measures for preclinical risk stratification.