From Experience to Engram: Implications of Rapid Neocortical Learning for Alzheimer's Disease

Abstract

The ‘complementary learning systems’ theory of memory describes a division of labor between the hippocampus and the neocortex. The hippocampus takes ‘snapshots’ of a given moment, and learns quickly, while the neocortex learns statistical regularities in the world slowly and incrementally. Recent evidence, however, has shown that the neocortex is capable of forming a rapid memory engram, if the learning material is repeatedly presented and can be integrated into preexisting knowledge structures. In this project, we imaged the spatial and temporal dynamics of memory representation for naturalistic stimuli (movie scenes) using a combination of functional MRI (fMRI) and diffusion-weighted MRI (dW-MRI). Through representational similarity analysis, we decoded the memory engram for high-level ‘event-scripts’ in two contexts: scenes set in airports and scenes set in restaurants. We found that each sub-event within the video clips elicited highly consistent and content-specific functional activation patterns, distributed across brain areas. Further, with repeated exposure, the precuneus significantly increased content-specific memory representation, while hippocampal involvement was rapidly attenuated. This evidence supports rapid neocortical learning through repetition and schema integration. Our findings have implications for long-term care approaches to memory disorders such as Alzheimer’s Disease (AD). Using a case study of a ‘dementia-village’ in Canada, I consider the ways that fast neocortical learning can be applied to optimize the memory capabilities of patients with AD.