Environmental dynamics shape perceptual decision bias.

Area of Science:

• Cognitive Neuroscience
• Computational Neuroscience
• Decision Science

Background:

• The brain integrates ambiguous sensory data with prior experience for interpretation.
• Environmental contexts can shift unpredictably, creating uncertainty about the current situation.

Purpose of the Study:

• To determine how context-specific prior knowledge optimally guides sensory interpretation in changing environments.
• To investigate if human decision-making strategies match these optimal Bayesian predictions.

Main Methods:

• Subjects performed a task involving ambiguous visual stimuli from dynamically switching distributions (contexts).
• Predictions were derived for an ideal Bayesian observer incorporating knowledge of environmental dynamics.
• Human choice data was analyzed to compare against model predictions.

Main Results:

• An ideal Bayesian observer's decisions are biased by the dynamically changing task context.
• Decision bias magnitude correlates with context reliability, environmental stability, and time since context switch.
• Human behavior validated all three predictions derived from the Bayesian model.

Conclusions:

• The brain effectively leverages knowledge of environmental change statistics for interpreting ambiguous sensory signals.
• Human decision-making strategies in dynamic environments closely resemble optimal Bayesian inference.
• This suggests sophisticated neural mechanisms for adapting to and predicting environmental shifts.