Theta rhythm supports hippocampus-dependent integrative encoding in schematic memory networks

Integrating new information into existing schematic structures of knowledge is the basis of learning in our everyday life activity as it enables structured representation of information and goal-directed behaviour in an ever-changing environment. However, how schematic mnemonic structures aid the integration of novel elements remains poorly understood. Here, we showed that the ability to integrate novel picture information into learn structures of picture associations that overlap by the same picture scene (associative network) or by the conceptually related scene information (schematic network) is hippocampus-dependent, as patients with lesions at the medial temporal lobe (including the hippocampus) were impaired in inferring novel relations between elements within these mnemonic networks but not in retrieving individual pictures in a subsequent memory test. In addition, we observed more persistent and widespread scalp electroencephalographic (EEG) theta oscillatory pattern (3-6Hz) while healthy participants encoded novel pictures related to schematic memory networks, suggesting that theta may reflect distances between elements within a representational network space. Finally, we found high similarity values for neural activity patterns elicited by novel and related events only within associative networks, thereby suggesting that neural reactivation may promote the integration of new information into existing memory networks only when direct associations within the network link their elements. These findings have important implications for our understanding of the neural mechanisms that support the development and organization of structures of knowledge.