Sensory and environmental uncertainty in perceptual decision-making

Area of Science:

• Cognitive Neuroscience
• Perceptual Decision-Making
• Computational Psychiatry

Background:

• Perceptual decision-making requires integrating noisy sensory input with prior knowledge about the environment.
• Understanding how different sources of uncertainty (sensory vs. environmental) interact is crucial for explaining perception.
• Previous research has not fully elucidated the interplay between sensory uncertainty and learned environmental regularities.

Purpose of the Study:

• To investigate the relationship between sensory uncertainty and environmental uncertainty in perceptual decision-making.
• To determine how learned beliefs about audiovisual associations are weighted against sensory information under varying uncertainty levels.
• To explore the role of individual differences in belief updating on the integration of sensory evidence and learned beliefs.

Main Methods:

• A combined motion discrimination and audiovisual reversal learning task was employed.
• Bayesian modeling was used to analyze the interplay of sensory and environmental uncertainties.
• Participant behavior was analyzed in relation to individual differences in belief updating tendencies.

Main Results:

• The influence of learned audiovisual associations on perceptual decisions increased with higher sensory uncertainty.
• This modulatory effect of learned beliefs was significantly larger under high environmental uncertainty compared to low environmental uncertainty.
• Individual differences in the tendency to change beliefs correlated with the reliance on learned beliefs during decision-making.

Conclusions:

• The weighting of sensory information versus learned beliefs is dynamically modulated by their respective uncertainties.
• Belief learning itself does not appear to depend on sensory uncertainty, but its application in decision-making does.
• These findings offer insights into perceptual inference and may inform our understanding of conditions like schizophrenia.