Trial-to-trial motor behavior during a reinforcement learning task in children ages 6 to 12

During practice, learners use available feedback from one trial to develop and implement motor commands for the next trial. Unsuccessful trials (i.e., "misses") should be followed by different motor behavior (e.g., goal-directed changes and/or exploration of movement parameters), while successful trials (i.e., "hits") should maintain the same behavior (e.g., minimize variance and recapitulate the same motor plan to the best of one's ability). Measuring the trial-to-trial changes in motor behavior can provide insights into how the motor system uses feedback and regulates movement variability while trying to improve performance. There have been no reports on the trial-to-trial motor behavior of typically developing children despite the profound motor development that occurs in this period and its relevance to long-term functional outcomes.

We recruited 72 typically developing children from ages 6 to 12 to perform a reinforcement learning beanbag toss to a target. Their target errors were used to examine their motor exploration and autocorrelation.

Comparing variability at different trial-to-trial intervals showed that children exhibit motor exploration above and beyond the effect of sampling bias. Mean autocorrelations of different lags were near zero suggesting that successive trials were largely unrelated.

We found evidence that children utilize motor exploration in the target space of a target throwing task. After failed trials they exhibited increased variability to search for more optimal motor solutions. After successes, they minimized variability to create the same successful performance.