A dual-networks architecture of top-down control was investigated using functional magnetic resonance imaging (fMRI) and resting-state functional connectivity. The study identified two distinct brain networks: the fronto-parietal network, which is involved in initiating and adjusting control, and the cingulo-opercular network, which maintains stable 'set-maintenance' over entire task epochs. These networks are organized in a small-world architecture, characterized by dense local connections and weaker long-range connections between components. The fronto-parietal network includes regions such as the dorsolateral prefrontal cortex (dlPFC), inferior parietal lobule (IPL), and intraparietal sulcus (IPS), while the cingulo-opercular network includes the anterior prefrontal cortex (aPFC), anterior insula/frontal operculum (aI/fO), and dorsal anterior cingulate cortex/medial superior frontal cortex (dACC/msFC). The cerebellum forms a separate but related cluster of regions that is interposed between the two networks. The study suggests that top-down control in the human brain is supported by a large collection of functionally related regions, organized into two distinct networks. These networks function at different timescales, contributing to the adaptability and stability of top-down control. The fronto-parietal network is responsible for rapid adaptive control, while the cingulo-opercular network is responsible for stable set-maintenance. The dual-networks model of top-down control suggests that the brain can rely on at least two fairly parallel control networks: one optimized for rapid adaptive control and the other for stable set-maintenance. This architecture increases the overall resilience of top-down control to damage or other perturbations. The study also highlights the importance of small-world network structures in efficient information processing, which is supported by the principles of complex systems. The findings suggest that the human brain uses multiple controllers and small-world-like architecture to ensure resilient performance.A dual-networks architecture of top-down control was investigated using functional magnetic resonance imaging (fMRI) and resting-state functional connectivity. The study identified two distinct brain networks: the fronto-parietal network, which is involved in initiating and adjusting control, and the cingulo-opercular network, which maintains stable 'set-maintenance' over entire task epochs. These networks are organized in a small-world architecture, characterized by dense local connections and weaker long-range connections between components. The fronto-parietal network includes regions such as the dorsolateral prefrontal cortex (dlPFC), inferior parietal lobule (IPL), and intraparietal sulcus (IPS), while the cingulo-opercular network includes the anterior prefrontal cortex (aPFC), anterior insula/frontal operculum (aI/fO), and dorsal anterior cingulate cortex/medial superior frontal cortex (dACC/msFC). The cerebellum forms a separate but related cluster of regions that is interposed between the two networks. The study suggests that top-down control in the human brain is supported by a large collection of functionally related regions, organized into two distinct networks. These networks function at different timescales, contributing to the adaptability and stability of top-down control. The fronto-parietal network is responsible for rapid adaptive control, while the cingulo-opercular network is responsible for stable set-maintenance. The dual-networks model of top-down control suggests that the brain can rely on at least two fairly parallel control networks: one optimized for rapid adaptive control and the other for stable set-maintenance. This architecture increases the overall resilience of top-down control to damage or other perturbations. The study also highlights the importance of small-world network structures in efficient information processing, which is supported by the principles of complex systems. The findings suggest that the human brain uses multiple controllers and small-world-like architecture to ensure resilient performance.