EAPSI: Investigating the mechanisms of visual working memory by testing predictions of a computational model

Swan Garrett, Boston MA

Investigators

Abstract

Successfully interacting with the environment depends upon the ability to briefly retain information in visual working memory. This form of memory is characterized as a balancing act, with the maximum resolution of each memory being constrained by the total number of objects stored. How this trade-off is implemented in the brain remains an open question and one that is vital to developing accurate theories of memory. The goal of this project is to better understand this trade-off between quality and quantity by testing memory recall in individuals with brain damage. The effects that the brain damage has on memory precision will provide insights into how visual working memory storage is implemented in the brain. This research will be conducted in collaboration with Dr. David Shum, an expert in memory research at Griffith University in Brisbane, Australia. To better understand the mechanisms underlying the trade-off between the quality and quantity of memories stored in visual working memory, a simulated lesion model approach will be used. Behavioral data collected from individuals with brain damage performing a delayed estimation task, which measures the quality of an individual memory representation, will be compared to simulations from a computational memory model called the Binding Pool with varying degrees of simulated damage. The model predicts that damage to the individuation of memories causes binding errors, but only when multiple items are being stored. However, if there is damage to the visual features that support memory, then memory quality will be degraded regardless of the number of items being stored. The model?s simulations of the effects of brain damage will be compared to the memory recall performance of individuals with brain damage. Successful predictions of the model provide confirmatory support for this model of human memory. This NSF EAPSI award is funded in collaboration with the Australian Academy of Science.