Persistent residual errors in motor adaptation tasks: reversion to baseline and exploratory escape.

Area of Science:

• Motor control
• Neuroscience
• Cognitive psychology

Background:

• Humans and animals often show incomplete compensation during motor adaptation tasks, leaving persistent residual errors.
• State-space models attribute these errors to a balance between error-based learning and a natural tendency to revert to baseline (forgetting).
• Prior research indicated that this baseline reversion isn't fixed and can be altered by manipulating feedback during specific trials.

Purpose of the Study:

• To investigate if novel error-clamp trials with non-zero error and variance can elicit different learning mechanisms.
• To determine if these mechanisms can overcome the residual error typically seen in adaptation tasks.
• To test the predictive power of state-space models under varied feedback conditions.

Main Methods:

• Subjects participated in motor adaptation tasks with perturbed movements.
• Novel error-clamp trials were introduced, featuring constrained feedback with non-zero error and variance.
• Behavioral data was analyzed to assess learning policies and compare them with state-space model predictions.

Main Results:

• When error clamps had non-zero error but zero variance, subjects shifted to an exploratory learning policy in response to residual error.
• When error clamp feedback was naturalistic (persistent mean error with variance), state-space models accurately predicted behavior, even without task success.
• Altering the error distribution during clamp trials triggered learning mechanisms not captured by the standard state-space model dynamics.

Conclusions:

• Residual error in motor adaptation stems from an error-dependent learning process with a reversion-to-baseline property.
• This process can inhibit other, potentially more effective, learning mechanisms.
• Manipulating feedback during error clamps can reveal and engage these alternative learning strategies.