The study investigates the dynamic reconfiguration of human brain networks during learning, focusing on the role of modularity in adapting existing brain functions and selecting new neurophysiological activities. Using functional MRI (fMRI) data from 18 participants learning a simple motor skill, the researchers identified modular structures at multiple temporal scales (days, hours, and minutes). They found that these modular structures changed adaptively over time, with a graded structure where fewer modules were present on longer time scales and more modules on shorter time scales. The study also developed a statistical framework to identify modular architectures in evolving systems, which is applicable to various fields where network adaptability is crucial. Key findings include: - Modular structure was consistently observed over different temporal scales, indicating its generalizability. - The community organization of brain connectivity evolved smoothly over time, displaying coherent temporal dependence. - Network flexibility, measured by the number of times nodes changed module allegiance, was modulated by early learning and varied among individuals. - Network flexibility in one session predicted learning success in subsequent sessions, suggesting its predictive value. The study provides insights into the adaptive modular organization of brain activity during learning and highlights the importance of network flexibility in understanding system performance.The study investigates the dynamic reconfiguration of human brain networks during learning, focusing on the role of modularity in adapting existing brain functions and selecting new neurophysiological activities. Using functional MRI (fMRI) data from 18 participants learning a simple motor skill, the researchers identified modular structures at multiple temporal scales (days, hours, and minutes). They found that these modular structures changed adaptively over time, with a graded structure where fewer modules were present on longer time scales and more modules on shorter time scales. The study also developed a statistical framework to identify modular architectures in evolving systems, which is applicable to various fields where network adaptability is crucial. Key findings include: - Modular structure was consistently observed over different temporal scales, indicating its generalizability. - The community organization of brain connectivity evolved smoothly over time, displaying coherent temporal dependence. - Network flexibility, measured by the number of times nodes changed module allegiance, was modulated by early learning and varied among individuals. - Network flexibility in one session predicted learning success in subsequent sessions, suggesting its predictive value. The study provides insights into the adaptive modular organization of brain activity during learning and highlights the importance of network flexibility in understanding system performance.