Hosted by the School of Psychological Science

Abstract: Recognition memory can arise from a rapid sense of familiarity whereas recall requires the conjuring of information not currently at hand. We explain recall using a recurrent auto-associative neural network, which recalls information dynamically through iterative pattern completion; memory cues activate an initial pattern, which gradually changes before settling into one of several previously learned attractor states. This filling-in-the-blanks process was termed “redintegration” by early memory researchers. We hypothesized that successful redintegration produces learning that affects the speed of subsequent recall attempts. As predicted by this account, retrieval practice (aka, “the testing effect”) produced faster subsequent recall than restudy practice, even when accuracy was similar following each form of practice. A subsequent study confirmed the prediction that after a failure to recall, participants would more quickly give up during subsequent recall attempts (e.g., a learned tip-of-the-tongue). Finally, we re-examined a classic result previously thought to indicate associative symmetry for memory representations; our reanalysis of the data found that recall was faster when tested in the same direction across recall attempts (e.g., study A-B; first test ?-B; second test ?-B) versus a change of direction (e.g., second test A-?), as predicted by learning within an auto-associative network with asymmetric associations.

Bio: David Huber is a Professor in the Psychology and Neuroscience Department at the University of Colorado Boulder. Previously, he was a Professor at the University of Massachusetts, Amherst, the University of California, San Diego, and the University of Maryland, College Park. After completing a PhD at the Indiana University, working with Richard Shiffrin, he was a postdoctoral research fellow at the University of Colorado Boulder. David’s research focuses on human perception and memory, using computational models to relate behavior to the underlying neural mechanisms.