Related Experiment Video
Updated: May 4, 2026
07:19
A Modified Lean and Release Technique to Emphasize Response Inhibition and Action Selection in Reactive Balance
Published on: March 19, 2020
7.6K
Task-specific response strategy selection on the basis of recent training experience.
Jacqueline M Fulvio1, C Shawn Green1, Paul R Schrater2
1Department of Psychology, University of Wisconsin-Madison, Madison, Wisconsin, United States of America.
Plos Computational Biology
|January 7, 2014
Summary
Human learning strategies adapt based on training complexity. Simple training favors direct mapping, while complex training promotes predictive models, influencing how broadly knowledge is applied.
Area of Science:
- Cognitive Psychology
- Machine Learning Theory
- Neuroscience
Background:
- Training aims for generalizable learning, but predicting learning scope is challenging.
- Existing research shows varied learning strategies (narrow vs. broad) from identical tasks.
Purpose of the Study:
- To test if humans choose strategies maximizing performance and reducing uncertainty during training.
- To determine if the chosen strategy predicts the transferability of learned skills.
Main Methods:
- Subjects trained on a moving dot extrapolation task.
- Two learning strategies evaluated: model-free (input-output mapping) and model-based (predictive model).
- Training varied in the number of distinct trajectories presented.
Main Results:
- Subject behavior aligned with a mapping strategy for low training trajectory numbers.
- Subject behavior shifted to a predictive model strategy for high training trajectory numbers.
- Strategy choice correlated with the extent of learning transfer.
Conclusions:
- Human decision-making in training adapts to task demands, favoring strategies that optimize performance and reduce uncertainty.
- The developed framework can interpret and guide the design of training paradigms for specific learning outcomes.
