Developmental differences in generalizable neural representations driven by multiple emotional and cognitive tasks

From the perspective of development, childhood is one of the most critical stage during brain development: neural system and cognitive behavior undergo a prolonged and intricate developmental process. A central question in developmental cognitive neuroscience pertains to how our brain develop highly specialized yet interacting neural modules to support a wide spectrum of cognitive and emotional functions. It is still inconclusive how these neural systems interplay and work together to promote cognitive and emotional maturation. The early maturational perspective believed that as the anatomical structure of a specific cortical area matures, each neural module will "perform their duties" to support the development of corresponding cognitive functions. Later, the interactive specialization theory argued that there is a special brain function module with the properties of a general developmental architecture to support the development of different cognitive abilities, which can co-activate in multiple neurobiological models. Recently, researchers proposed a multi-demand system model, where the frontal-parietal network system supports various cognitive functions through diverse neural activation modes, fostering cognitive flexibility, and playing a role in coordinating and integrating different levels of neural computing resources across cognitive domains during children's brain development. Based on the interactive specialization and multi-demand system model, the present study put forward the scientific questions: whether the multi-demand frontal-parietal system have a general neural representation pattern under different cognitive subdomain tasks, and how this pattern supports the development of children's multiple cognitive domains through a hierarchical distributed neural representation organization. Integrating traditional developmental psychology with non-invasive functional magnetic resonance imaging in cognitive neuroscience, we used multiple task paradigm (attention network test, numerical N-Back working memory and emotion matching tasks) across cognitive domains and innovative hierarchical distributed neural representation modeling to explore a general neural representation framework and its developmental rules for multiple cognitive domains. By building hierarchical distributed neural representation modeling method across multiple cognitive domains, we systematically investigate the developmental patterns of neural information representation in children and adults. The results indicated that both children and adults exhibited the phenomenon of the multiple-demand frontoparietal system (including the intraparietal sulcus and frontal eye area) jointly participating in a variety of emotional and cognitive tasks, that is, co-activation across tasks; it is worth emphasizing that the multiple-demand frontoparietal system in children showed lower levels of generalizability of neural representations across tasks, whereas the anterior cingulate gyrus, dorsolateral prefrontal cortex, and anterior insula, which were used as control analyses, did not show differences in generalizability between the groups. We speculate that the multi-demand frontoparietal system may serve as a potential universal "hub" during development. Through compositional information coding organization, it can enable hierarchical neural representation and computation driven by different task goals, thereby supports the development of emotional and cognitive functions with age. This study breaks through the current research framework of developmental cognitive neuroscience from the perspective of a single-task paradigm and is expected to provide new insights into the working principles of brain development across emotional and cognitive domains, as well as to inspire the new artificial intelligence algorithms.