Neural network dynamics during the generation and reinstatement of mnemonic representations
Abstract
The capacity to generate, reinstate, and mentally simulate mnemonic representations is a fundamental aspect of the human mind. It allows us to internally experience other places and moments of time, abstracting away from the present moment into past experiences or hypothetical future states of the world. Theoretical models posit that this capacity is afforded by a neural network distributed across the brain that codes features from our environment and experiences into neural representations that can be reinstated or flexibly combined in a goal oriented manner. Central to this network is the hippocampus, a region of the medial temporal lobes that putatively indexes both the spatial composition of a mental scene and the pattern of hippocampal-cortical interactions that represent feature details. Despite decades of research on hippocampal function during memory processes, our understanding of how this neural network operates dynamically remains limited. This thesis aim to assist in resolving this by investigating patterns of network reconfiguration that occur as a mnemonic representation of a virtual city is generated and reinstated to guide mental simulations of movement through the city. Chapter 2 provides evidence for a general encoding mechanism where the brain transitions from a state of information integration to localized processing based on encoding demands. These results are extended by showing that the hippocampus demonstrates flexibility in how it interacts with other brain regions to actively reinstate and bind features into a holistic representation that is used for mentally simulating movement. Chapter 3 investigates regional effects associated with the task as a validity check. Chapter 4 uses network reconfiguration processes to show that the default mode network, a putative task-negative system, also demonstrates flexibility by altering the functional interactions between its components and regions of the mental simulation network to facilitate feature integration during mnemonic reinstatement. Collectively, these results provide a schematic for extending existing theoretical models on memory function into a dynamic perspective based on the adaptability of neural networks and the flexibility of network components to alter patterns of functional interactions across the brain to process information in a contextual, goal oriented manner.
Description
Keywords
Neuroscience, Psychology--Cognitive
Citation
Arnold, A. (2017). Neural network dynamics during the generation and reinstatement of mnemonic representations (Doctoral thesis, University of Calgary, Calgary, Canada). Retrieved from https://prism.ucalgary.ca. doi:10.11575/PRISM/27080