Title: Drawings reveal changes in object memory, but not spatial memory, across time
Emma Megla, doctoral student in the Bainbridge Lab, Department of Psychology, University of Chicago
Abstract: Time has a powerful sway over our memories. We better remember an experience that lingered than one we barely glimpsed, and better remember what occurred minutes ago than days ago. Despite centuries of mapping the relationship of time on memory, including Ebbinghaus’ famous ‘forgetting curve’, little research has pinpointed the changes in memory content that drive memory performance across time. What features (e.g., spatial accuracy) of a memory are actually changing while an experience is encoded and retained in memory? Here, we leveraged 869 drawings of scenes made from memory after variable encoding (Experiment 1) and retention of that memory (Experiment 2). Through crowdsourced scoring of these drawings by thousands of participants—scoring the objects recalled, the presence of false objects, and spatial accuracy—we were able to quantify how the content of memory changes across time on the feature level. We find that whereas the number of objects recalled from a scene, including the number of false objects, is highly dependent on time, spatial memory is largely precise after just 100 msec of encoding or after one week of retaining the memory. Additionally, we also find that the location and meaning of an object predicts when it will be recalled across encoding, but the saliency of an object when it will be forgotten.