Did We Get Sensorimotor Adaptation Wrong? Implicit Adaptation as Direct Policy Updating Rather than...

Area of Science:

• Motor control and learning
• Computational neuroscience
• Human motor adaptation

Background:

• The prevailing theory suggests motor adaptation relies on updating an internal forward model to predict movement consequences.
• Evidence supports forward models for body state tracking, but their role in movement planning remains debated.
• An alternative, direct policy learning, proposes motor output adjusts directly based on errors, potentially independent of forward models.

Purpose of the Study:

• To dissociate between forward-model-based learning and direct policy learning mechanisms in human motor adaptation.
• To investigate implicit adaptation under mirror-reversed visual feedback, an extreme perturbation.
• To determine which learning mechanism better explains observed adaptation patterns.

Main Methods:

• Three experiments involving human participants (N=47) were conducted.
• Participants adapted to motor tasks under mirror-reversed visual feedback.
• Behavioral data on adaptation patterns over time and across targets were analyzed.

Main Results:

• Implicit adaptation occurred even under mirror-reversed feedback, amplifying errors rather than reducing them.
• The observed adaptation patterns were consistent with direct policy learning.
• Findings contradicted predictions of the forward-model-based learning theory.

Conclusions:

• The study challenges the necessity of forward-model updates for implicit motor adaptation.
• Direct policy learning offers a more plausible explanation for observed adaptation behaviors.
• Re-examination of the forward-model-based theory of motor adaptation is warranted.