Uncertainty and predictiveness determine attention to cues during human associative learning

Area of Science:

• Cognitive psychology
• Computational neuroscience
• Human learning

Background:

• Attention is guided by predicting outcomes and exploring uncertainty.
• Previous studies showed either exploitation or exploration, but not both simultaneously.
• Methodological differences hindered pinpointing factors influencing attentional dominance.

Purpose of the Study:

• To systematically investigate attentional exploitation and exploration within a single experimental paradigm.
• To determine the independent and interactive effects of cue predictiveness and outcome uncertainty on attention.
• To provide empirical evidence for a hybrid model of attention.

Main Methods:

• Two experiments were conducted manipulating cue predictiveness and outcome uncertainty.
• Participants' attentional responses to stimuli were measured.
• A hybrid attentional model was used to simulate observed data patterns.

Main Results:

• Both attentional exploration (driven by uncertainty) and exploitation (driven by predictiveness) were observed within the same experiment.
• Predictiveness influenced novel learning in a subsequent stage.
• Uncertainty's effect on learning diminished in the new context, unlike predictiveness.

Conclusions:

• Attentional exploration is more sensitive to contextual changes than attentional exploitation.
• The findings support a hybrid model where both processes operate dynamically.
• This research clarifies the distinct roles of prediction and uncertainty in guiding attention and learning.