Significant Progress in Understanding Human Memory and Imagination through Artificial Intelligence
A study conducted by University College London (UCL) and published in the journal Nature Human Behavior has made significant strides in the field of understanding human memory and imagination using artificial intelligence (AI) technologies. The study utilized an AI model, known as a generative neural network, that mimics brain functions.
The researchers focused on exploring how generative models can help explain how individuals utilize memories to perceive the world, recollect past experiences, and construct imaginative scenarios. The AI model included networks that represented the hippocampus and neocortex, two brain regions known for their involvement in memory processes, imagination, and planning.
Eleanor Spence, a graduate student from UCL Institute of Cognitive Neuroscience, remarked, “Generative networks used in AI demonstrate how information can be derived from experiences, enabling us to recall specific events while also flexibly representing new experiences.”
During the experiment, the researchers provided the models with 10,000 images of simple scenes. The network simulating the hippocampus swiftly encoded each scene as it appeared. Subsequently, these scenes were repeatedly reproduced to train the generative neural network in the neocortex. As a result, the neocortex acquired the ability to generate effective “conceptual” scenes, enabling the recreation of old scenes and the generation of entirely new ones.
Professor Neil Burdess, the senior author of the study from UCL Institute of Cognitive Neuroscience and UCL Queen Square Institute of Neurology, explained, “The way in which memories are reconstructed and are not accurate representations of the past demonstrates how the meaning or essence of an experience combines with unique details, eventually leading to biases in our memory of events.”
This groundbreaking