Generative AI Model Sheds Light on Human Memory and Imagination

UCL Study Reveals How Brain Forms Conceptual Representations and Predicts Future Events

UCL researchers used a generative AI model to simulate how the brain's neural networks learn from and remember events, providing insights into memory, imagination, and planning.
The AI model, representing the hippocampus and neocortex, demonstrated how the brain forms efficient 'conceptual' representations from experiences, enabling the recreation of past events and the generation of new ones.
The study highlights the reconstructive nature of memory, showing how the meaning or 'gist' of an experience is recombined with unique details, resulting in biases in how we remember things.
The research also underscores the predictive role of memory, suggesting that replaying memories helps our brains pick up on patterns from past experiences that can be used to make predictions necessary for survival.
The model explains how the neocortex slowly acquires conceptual knowledge and, in conjunction with the hippocampus, allows us to 're-experience' events by reconstructing them in our minds.