Stroke induces widespread disruptions to brain function, extending beyond focal lesions to alter the multiscale temporal dynamics that govern neural processing. These dynamics-operating across milliseconds to months-form a hierarchical architecture essential for communication, integration, and adaptive plasticity. While stroke is known to impair local neural activity, its effects on this temporal hierarchy and their consequences for functional recovery remain poorly understood. We conducted a comprehensive investigation of intrinsic neural timescales (INT) in 15 ischemic stroke patients using longitudinal fMRI at five time points over six months, comparing them to age-matched healthy controls. INT quantifies how long neural populations retain information, providing a quantitative measure of fundamental processing dynamics. To further elucidate the mechanistic basis of stroke-induced changes, we performed computational modelling of parsimonious excitable neuronal network dynamics, offering insights into alterations in brain activity from a dynamical systems perspective. Our analyses revealed some key findings: Stroke patients exhibited significantly prolonged INT across multiple cortical regions, indicating slowed temporal dynamics that persisted throughout recovery. The typical hierarchical organization of INT, where sensory areas have shorter timescales than higher-order association areas, was disrupted, particularly in the early post-stroke period. Recovery trajectories diverged at two months post-stroke: patients with poor outcomes maintained abnormally long INT in cognitive control networks (dorsal attention, language, and salience systems), whereas those with better recovery showed progressive normalization toward healthy INT patterns, restoring the brains dynamic balance across multiple timescales. Stroke-induced INT prolongation can be modelled as a critical slowing down driven by a shift in the distance to criticality caused by increased neuronal excitability. Within the framework of criticality, where neural systems operate near the boundary between order and disorder to optimize information processing, stroke-induced changes in neural excitability appear to push brain dynamics toward the critical regime and potentially into a supercritical state. This shift, characterized by excessive temporal persistence and reduced flexibility, may underlie both the observed INT prolongation and its association with poor recovery outcomes. Our findings highlight the importance of temporal dynamics in stroke recovery and suggest that INT may serve as a biomarker for predicting long-term functional outcomes. This perspective provides a novel way to conceptualize stroke-induced changes in brain dynamics, framing them within the broader context of brain self-organization and adaptive processes. By integrating these insights with neurorehabilitation strategies, such as non-invasive brain stimulation, INT could inform targeted interventions to restore neural excitability and intrinsic timescales, thereby improving recovery trajectories.