Motor adaptation as a greedy optimization of error and effort

Area of Science:

• Neuroscience
• Motor Control
• Robotics

Background:

• Motor adaptation is crucial for navigating new dynamic environments.
• Current models focus on anticipating forces to reduce kinematic error.

Purpose of the Study:

• To propose a more general model of motor adaptation.
• To demonstrate that motor adaptation optimizes a cost function of kinematic error and effort.

Main Methods:

• Developed a theoretical model of motor adaptation as cost function minimization.
• Derived learning dynamics as a linear, auto-regressive equation.
• Validated the model with experimental data from force-field adaptation during walking.

Main Results:

• The proposed optimization model predicts previously observed learning dynamics.
• Model coefficients fall within a validated range.
• The model accurately captures error patterns across various force-field conditions.

Conclusions:

• Motor adaptation can be approximated as a greedy minimization of a weighted sum of error and effort.
• The single-state, auto-regressive equation adequately describes motor adaptation to force fields.
• This framework provides a unified view of motor learning under different experimental conditions.