Practicing one thing at a time: the secret to reward-based learning?
van der Kooij, K.; van Mastrigt, N.; Smeets, J.
Show abstract
Binary reward feedback on movement success is sufficient for learning in some simple reaching tasks, but not in some more complex ones. It is unclear what the critical conditions for learning are. Here, we ask how reward-based sensorimotor learning depends on the number of factors that are task-relevant. In a task that involves two factors, we test whether learning improves by giving feedback on each factor in a separate phase of the learning. Participants learned to perform a 3D trajectory matching task on the basis of binary reward-feedback in three phases. In the first and second phase, the reward could be based on the produced slant, the produced length or the combination of the two. In the third phase, the feedback was always based on the combination of the two factors. The results showed that reward-based learning did not depend on the number of factors that were task-relevant. Consistently, providing feedback on a single factor in the first two phases did not improve motor learning in the third phase.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
- Self-controlled practice and nudging during structural learning of a novel control interface 97%
- Quantifying exploration in reward-based motor learning 96%
- Asymmetry in kinematic generalisation between visual and passive lead-in movements are consistent with a forward model in the sensorimotor system 95%
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.