Reza Shadmehr, along with a team of colleagues from four other countries, has been awarded a grant from the Human Frontier Science Program. The grant in the amount of $450,000 per year for three years will support the team’s research on the topic “The multiple timescales of motor memory.”
Human Frontier awardees assemble an international team consisting of scientists from different disciplines tackling a single problem from multiple directions, combining their expertise to approach questions that could not be answered by individual laboratories. Shadmehr’s proposed work involves labs in 5 countries (US, Canada, Japan, UK, and France). The collaborators are colleagues who are world leaders in the field of motor control.
Shadmehr provides a short description of the project:
“Textbooks present motor memory as monolithic: once acquired, never forgotten. They dissociate it from declarative memory by suggesting that motor memory does not have a short-term form. However, recent theoretical and experimental results have overturned this view, suggesting that motor memory is supported by processes that have multiple timescales: fast processes that learn quickly but decay rapidly with time, and slow processes that are less sensitive to movement errors but show better retention. Indeed, it appears that formation of motor memory may be a gradual transformation from the fast to the slow processes, and this transformation may depend on time and the statistical properties of the learner’s performance errors. We propose to use this idea to shed new light on one of the fundamental problems in neuroscience: the distinct functions of the cortical motor structures vs. the cerebellum in learning and retention of skilled movements.
“Our basic hypothesis is that the faster timescales of memory are dependent on the cerebellum, effectively learning to predict the consequences of motor commands and correcting the motor commands via internal feedback, and the slower timescales are dependent on the cerebral cortex, effectively learning to produce the motor commands appropriate for the specific conditions of the task.
“Our project combines computational, psychophysical, neuropsychological, and neurophysiological approaches to investigate time and statistics of performance. We will use common eye and arm movement paradigms in behavioral and neurophysiological experiments to explore how controlled changes in sensory and motor noise effect motor control and learning. This will enable us to directly relate high level behavior to neural mechanisms.”