Exploration of neurocognitive interaction between language and memory via graph theory

Since complex human behaviors rely on the interaction between neurocognitive functions and the brain operates as a network, modern cognitive neuroscience studies aim to map cognitive onto cortical networks. Furthermore, a complete description of neurocognitive interaction should be based on intrinsic and extrinsic states to capture its underlying flexibility and dynamic architecture. We explored the language - declarative memory interaction and their cerebral substrate language and memory network (LMN) within this framework, following substantial evidence of their interconnection. Temporal lobe epilepsy (TLE) patients can especially serve as a model for understanding this interaction since their epileptogenic zone is located in temporal regions that serve as an integrative hub, and they show both naming and long-term memory deficits. We assumed that the language and declarative memory integration is based on the dynamic LMN that reconfigures with task demands and brain status. Therefore, we explored two types of LMN dynamics: (1) a state reconfiguration (intrinsic resting-state vs. extrinsic sentence recall task) showing key subprocesses the LMN is supporting with its adaptation; and (2) a reorganization of state reconfiguration (TLE vs. healthy controls) showing the additional processes needed to support language-memory interaction when the standard interface is unfunctional. To that end, we applied graph theory to study how LMN responds to environmental demands by segregating into modules that support common processes and how are different LMN modules inter-connected to provide integration. We will present our findings and discuss them by proposing a dynamic language and memory interaction model. Banjac, S. (LPNC), Roger, E., Pichat, C., Cousin, E., Mosca, C., Lamalle, L., Krainik, A., Kahane, P., & Baciu, M.