Error estimation abilities and self-controlled feedback schedules

Abstract

The learning advantages of self-controlled feedback schedules are often attributed to the development of a more sensitive error detection and correction mechanism. That is, during practice participants engage in error estimation processes upon movement completion, which guides the feedback decision. Indeed, past research has shown that self-controlled feedback schedules lead to more accurate error estimation in retention and transfer relative to yoked feedback schedules (e.g., Carter et al. 2014). A limitation of these experiments is that no baseline of one's error detection and correction mechanism was collected. Thus, it is difficult to rule out participants in the self-controlled group naturally having more accurate error estimation abilities. Here, we addressed this limitation by randomly assigning half of the self-controlled participants to error estimate after each trial in a no-feedback pre-test. On Day 1, participants practiced an aiming task involving a rapid 40-deg extension movement in exactly 225 ms. Participants returned 24-hrs later to complete no-feedback retention and transfer tests, and all participants were asked to error estimate after each trial. Participants had less accurate error estimations in pre-test as compared to retention, and retention was more accurate than transfer. Self-controlled error feedback schedules (e.g., -24-deg) lead to more accurate estimations compared to self-controlled graded feedback schedules (e.g., too short). These data, along with previous findings, suggest that self-controlled feedback is effective for training error estimation abilities. However, this benefit may be restricted to feedback that provides both magnitude and direction information rather than only direction information.

Acknowledgments: Funded by NSERC