How agency shapes behavior and performance: the triple impact of control-feedback on stimulus–response learning, motor reinforcement, and motivated action selection

TL;DR

This study shows how immediate feedback from actions improves learning, motor performance, and voluntary choices.

Contribution

The paper introduces a triple impact model of control feedback on learning, motor reinforcement, and action selection.

Findings

Control feedback improves stimulus–response learning and motor performance.

Immediate feedback reinforces motor execution and action choice preferences.

Reinforced pairings influence voluntary actions even without conscious awareness.

Abstract

Agency confirmation via control feedback (e.g., an immediate sensory consequence of one’s action) has been shown to motivate action choice and reinforce motor responses. Recent work also demonstrated that it qualitatively improves motor performance. The present study tested the hypothesis that this improvement arises because control feedback selectively strengthens stimulus–response (S–R) associations, and further examined its reinforcing impact on motor responses and action choice to provide an integrated account of how agency confirmation shapes behavior and performance.

Three experiments employed an acquisition-test paradigm. During acquisition, specific stimulus–response combinations triggered an immediate perceptual effect, while other combinations produced no effect (Experiments 1 and 3) or a delayed effect (Experiment 2). In the test phase, the perceptual effect depended solely on the response (Experiments 1 and 2) or was absent (Experiment 3). Experiment 3 also included a free-choice phase assessing the motivating impact of control feedback on voluntary action selection and explicit knowledge regarding the S–R pairings.

Control feedback enhanced S–R learning, yielding faster and more accurate performance for previously reinforced pairings compared to delayed or no-effect conditions. Immediate response-contingent effect independently facilitated motor execution (Experiments 1 and 2), and reinforced S–R pairings biased action choice preference (Experiment 3) even without explicit awareness of the pairings.

Agency confirmation via control feedback exerts a triple and partially dissociable influence on behavior, enhancing S–R learning, reinforcing motor execution, and motivating voluntary action. The findings inform models of action control and motor skill learning.

Figures