Reaction Time "Mismatch Costs" Change with the Likelihood of Stimulus-Response Compatibility.

Area of Science:

• Cognitive Neuroscience
• Computational Neuroscience
• Human Motor Control

Background:

• Dyadic interactions necessitate dynamic movement correspondence.
• Behavioral responses range from imitation to counter-imitation, adapting to context.
• Predicting others' actions is crucial for optimal joint motor planning.

Purpose of the Study:

• To investigate the mechanisms of statistical learning in stimulus-response compatibility during dyadic interactions.
• To compare computational models (Rescorla-Wagner vs. hierarchical Gaussian filter) for explaining behavioral adaptation.
• To examine how environmental predictability influences imitation and counter-imitation.

Main Methods:

• 28 healthy adults performed imitation and counter-imitation tasks with varying stimulus-response congruence likelihoods.
• Computational modeling compared a fixed learning-rate model (RW) with an adaptive learning-rate model (HGF).
• Analysis focused on reaction times and model evidence to understand behavioral prediction.

Main Results:

• The hierarchical Gaussian filter (HGF) model better explained participant behavior than the Rescorla-Wagner (RW) model.
• Motor response priming (faster reaction times) for matching actions occurred only in highly volatile (unpredictable) contexts.
• In predictable contexts, incongruent (mismatched) responses were faster, reversing the typical stimulus-response compatibility effect.

Conclusions:

• Hierarchical Bayesian learning of environmental statistics may underpin response priming in dyadic interactions.
• Behavioral adaptation in joint actions relies on inferring and adapting to environmental volatility.
• The findings challenge purely imitative views, highlighting adaptive prediction in human motor coordination.