Humans use underestimates of auditory spatial and temporal uncertainty for perceptual inference

Area of Science:

• Neuroscience
• Cognitive Science
• Decision Making

Background:

• The brain makes decisions using noisy sensory data, considering signal uncertainty.
• Accurate estimation of sensory uncertainty is assumed for optimal perceptual inference.

Purpose of the Study:

• To investigate if perceptual inferences depend on accurate sensory uncertainty estimates.
• To determine if sensory uncertainty is consistently and accurately estimated across different sensory domains.

Main Methods:

• Participants performed auditory, visual, and audiovisual spatial and temporal tasks.
• Bayesian observer models were fitted to individual participant data for each task.
• Model variants assessed the impact of independent uncertainty estimates on perceptual inference.

Main Results:

• Perceptual decisions consistently relied on underestimates of auditory spatial uncertainty.
• Audiovisual temporal uncertainty was also underestimated in perceptual decisions.
• Model comparisons indicated that actual uncertainty was not accurately reflected in decision-making.

Conclusions:

• Human perceptual decisions do not always optimally account for sensory uncertainty.
• Findings challenge the assumption of accurate sensory uncertainty utilization across all sensory domains.
• The brain may employ heuristics or biases in estimating sensory uncertainty for decision-making.