Abstract: PURPOSE: Since the beginning of life, visual perception provides important means to access the environment which are vital stimuli for brain maturation. However, the developing pattern of visual processing networks and the dynamic modularity on whole-brain level during childhood remain unclear. This study aims to shed light on the detailed developmental trajectory of visual networks on different cognitive levels, i.e., object, human face, and word.
METHODS: We utilized millisecond-resolution high-density electroencephalogram (HD-EEG) to achieve visualization of real-time dynamic functional networks of visual recognition. Three categories of visual stimuli, Chinese character, Human face, and Object (photos of commonly found things) were involved. Among a large cohort of typically developed children (n=120, aged 3-13) and health adults (n=35), the neural responses were surveyed through multiple task-activated network analyses based on functional connectivity (FC). Multiple measurements of functional network properties, including node hubness, centrality and modularity, were involved.
RESULTS: The developmental trajectory of visual network is not monotonously upgrading to the matured model but maintain trade-off patterns on FC, cortical hubness, and network modularity, mainly including: (1) Decreasing cortical activation is replaced by increasing functional connectivity on the whole brain. (2) Cortical engagement allocation shifted from the occipital visual areas to higher-level cortices such as left dorsal lateral prefrontal cortices (dlPFC) and parahippocampus. (3) Visual word (Chinese character) network hubs showed two different developmental trajectories, evolving into either strengthening hubs in left dlPFC or attenuated hubness but with enhanced cortical engagement in left parahippocampus. (4) Modularity measurements showed a trade-off between global hub and local hub, with global hub modulated by long-distance connectivity to engage in a high-level task. Local hub was supported by neighbour connection, which emerged early than global hubs and demonstrated plasticity to adapt to increasing processing complexity.
CONCLUSIONS: Besides the maturation of specialized cortices and increasing expertise, our findings highlight the importance of whole-brain network development for the high-level cognitive functions which is featured by dynamically balancing resources to meet the fast-changing brain structure and outer environment. The trade-off patterns help children’s brain achieve a dynamic optimization for efficiency and energy-saving, especially in high-level tasks such as word reading. This study provides understanding on typical as well as atypical brain development and necessitates a rethinking of how cognitive networks develop during childhood.

Key words: Visual processing, Brain development, Functional network, Trade-off