Useful misrepresentation: perception as embodied proactive inference.

Area of Science:

• Cognitive Neuroscience
• Computational Psychiatry
• Philosophy of Mind

Background:

• The predictive processing framework posits that perception aims to represent the environment for action, not objective reality.
• This framework suggests that prior information (expectations) shapes sensory inference.
• Understanding these biases is crucial for explaining perceptual phenomena and their relation to mental health.

Purpose of the Study:

• To argue that perception is optimized for embodied action opportunities.
• To demonstrate how biases in expectations facilitate useful inferences.
• To explore the theoretical implications for psychopathology.

Main Methods:

• Review and synthesis of existing literature on perception, predictive processing, and decision-making under uncertainty.
• Analysis of studies where perceptual accuracy conflicts with utility.
• Interpretation of findings within the predictive processing framework.

Main Results:

• Perception systematically deviates from accuracy when utility and accuracy are in conflict.
• Biases in prior predictive information facilitate inferences that support advantageous actions.
• The brain appears to balance prediction errors from sensory input and action-related outcomes.

Conclusions:

• Perceptual systems prioritize actions that avoid physiologically surprising outcomes, even at the cost of objective accuracy.
• This trade-off between accuracy and utility offers a novel perspective on cognitive function.
• The findings provide new theoretical insights into the mechanisms underlying psychopathology.