Poststroke sleep phenotypes and executive dysfunction: toward a sleep-network-executive framework

Post-stroke sleep disturbance and executive dysfunction frequently coexist, yet the mechanisms linking disturbed sleep to cognitive recovery remain poorly understood. A major challenge is phenotypic heterogeneity: insomnia, sleep fragmentation, obstructive sleep apnea, hypersomnolence, and circadian disruption differ in their biological drivers and cognitive consequences. To develop a testable network-level account, we conducted a targeted synthesis of clinical stroke studies, stroke neuroimaging research, non-stroke sleep-network evidence, and intervention studies, mapping each evidence stream according to its proximity to the proposed pathway. The synthesis indicates that focal lesions, diaschisis, and structural disconnection after stroke can disrupt the integrity and coordination of the salience, default mode, and frontoparietal networks. Clinical studies support an association between disturbed sleep and poorer post-stroke cognition, while neuroimaging findings demonstrate altered large scale network organization in post-stroke cognitive impairment. Evidence from non-stroke populations further suggests that sleep loss, insomnia, and sleep fragmentation can impair network segregation, default mode regulation, frontoparietal recruitment, and salience-network-mediated switching. We therefore propose that sleep fragmentation, insomnia, and arousal instability may act as functional stressors on an already vulnerable post-stroke network architecture. Inefficient switching may permit context-inappropriate persistence of default mode activity and reduce or delay frontoparietal engagement during executive demand, thereby contributing to impaired conflict monitoring, working memory, processing speed, and cognitive flexibility. No study has directly tested this complete pathway within a single stroke cohort. The framework should therefore be regarded as a falsifiable mechanistic hypothesis rather than an established causal model. Future studies should combine objective sleep phenotyping, dynamic and effective connectivity, task based neuroimaging, domain specific executive measures, and longitudinal or interventional designs within the same stroke cohorts.