Identifying novel memory traces that improve action precision

Harvard University, Cambridge MA

Investigators

Abstract

How do we remember a phone number long enough to dial it or remember two numbers long enough to add them together in our head? This type of memory, called short-term working memory, is a cognitive store that allows a few items to be recalled and acted upon within a short span of time. Early studies of working memory suggested that we can typically remember about 5 to 9 unrelated items. More recent work suggests a true capacity of just 3-4 distinct items. A briefer but far more vivid form of short-term memory, called sensory memory, has been identified in the visual, auditory, and tactile systems. Visual sensory memory, for example, is remarkable in that it can provide incredibly detailed information about 64 or more items in recent visual scenes, a far greater capacity than shown for working memory. The present work will study a previously unidentified proprioceptive sensory memory (proprioception refers to information about the position and movement of the body). Like visual sensory memory, proprioceptive sensory memory provides a high-precision but short-lasting store for sensory information and for information about recent motor actions. The hypothesis is that both type of memories, when available, provide high-precision input into the sensorimotor neural circuitry involved in action planning, allowing for extremely high levels of motor precision. The goals of this proposal are to develop an understanding of the relationship between the availability of novel high-precision proprioceptive and motor command memories and the spatiotemporal properties of improvements in motor precision. The team will begin by identifying the existence of a high-precision sensory memory for proprioception by determining the link between the availability of this novel memory and improved action precision. They will then characterize the extent and time course of the rapid reduction in time scale variability that this memory can provide. Finally, the research will parcel out two novel high-precision hyper-transient memories, proprioceptive sensory memory and motor command memory, based on both spatial and temporal properties, using geometric characterization and direct experimental manipulation. The planned work will develop a framework for understanding how recent sensory and motor memories can work both separately and in combination to improve motor precision during voluntary movement. This project provides a fertile research training opportunity to apply computational and engineering principles and tools to the study of learning and memory in neural circuits for trainees ranging from the undergraduate to the postdoctoral level, and introduces community middle school and high school students to how learning and memory shape the ability to precisely control actions in humans. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.