Implicit and explicit learning of Bayesian priors differently impacts bias during perceptual decision-making

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Visual Perception

Background:

• Implicit knowledge influences behavior without conscious awareness.
• The role of implicit statistical learning in visual perceptual decision-making remains unclear.
• Distinguishing between implicit and explicit knowledge processing is crucial for understanding cognition.

Purpose of the Study:

• To investigate whether implicit knowledge of statistical structure informs visual perceptual decisions.
• To determine if explicit knowledge of statistical probabilities is necessary for visual decision-making.
• To explore the mechanisms by which implicit and explicit priors are implemented in perceptual tasks.

Main Methods:

• A novel visual decision-making task involving translational Glass patterns of different colors.
• Each color was associated with distinct orientation probabilities, allowing assessment of general and specific priors.
• The drift diffusion model framework was used to analyze decision-making parameters.

Main Results:

• Both implicit and explicit learners utilized a general orientation bias by adjusting the starting point of evidence accumulation.
• Explicit learners also modified the drift rate offset for general priors.
• Stimulus-specific biases were primarily implemented by adjusting the drift rate offset.

Conclusions:

• Humans can implicitly learn statistical priors for perceptual decision-making.
• The implementation of learned priors differs based on the level of awareness (implicit vs. explicit).
• Awareness dictates the specific neural or cognitive mechanisms employed to utilize statistical information in perception.