Pupil dilation indicates the coding of past prediction errors: Evidence for attentional learning theory

Area of Science:

• Cognitive Psychology
• Neuroscience
• Learning Theory

Background:

• The Pearce-Hall attentional learning theory posits increased attention to uncertain cues associated with high prediction error.
• Understanding the neural and physiological correlates of error-driven attention is crucial for explaining learning dynamics.

Purpose of the Study:

• To investigate the relationship between cue-elicited pupil dilation and error-driven attentional processes in associative learning.
• To test predictions of the Pearce-Hall theory using pupillometry as a measure of attention.

Main Methods:

• Three experiments involving associative learning with appetitive, motor, and aversive outcomes.
• Participants learned associations between continuously reinforced (certain) and partially reinforced (uncertain) cues.
• Cue-elicited pupil dilation was measured during outcome anticipation and analyzed across learning phases.

Main Results:

• Initially, consistently reinforced cues elicited greater pupil dilation than uncertain cues.
• This pattern reversed over time, with uncertain cues eliciting greater pupil dilation by the end of learning.
• The observed changes in pupil dilation mirrored predicted shifts in prediction error and attention.

Conclusions:

• Conditioned pupil responses during associative learning reflect dynamic, error-driven attentional allocation.
• The findings support the Pearce-Hall attentional learning theory by demonstrating how attention shifts towards uncertain cues with experience.
• Pupillometry serves as a valuable, non-invasive measure for studying attentional mechanisms in learning.