Do humans make good decisions?

Area of Science:

• Cognitive Science
• Neuroeconomics
• Computational Neuroscience

Background:

• Human decision-making exhibits context-dependent inconsistencies, particularly in economic choices, deviating from ideal observer models.
• Perceptual classification tasks often show near-ideal performance, contrasting with the irrationality observed in economic behavior.
• Existing models struggle to unify these divergent findings in human decision-making under varying conditions.

Purpose of the Study:

• To propose a unified computational framework explaining context-dependent suboptimal choices in both perceptual and economic domains.
• To investigate the role of efficient coding principles in shaping decision-making strategies.
• To reconcile the apparent paradox of near-ideal perception versus suboptimal economic behavior.

Main Methods:

• Review of recent empirical work on perceptual and economic decision-making in variable and volatile environments.
• Development of a computational framework based on efficient coding principles.
• Analysis of how efficient coding strategies impact information transmission and decision robustness.

Main Results:

• Efficient coding, prioritizing expected inputs, leads to robust but context-dependent decisions that deviate from strict optimality.
• Suboptimal context-dependence observed in economic choices is mirrored in perceptual decisions under similar environmental variability.
• The proposed framework successfully accounts for findings across both perceptual and economic decision-making domains.

Conclusions:

• Suboptimal choices in economic and perceptual tasks can be understood as a consequence of efficient information coding in dynamic environments.
• Efficient coding represents an adaptive strategy for limited-capacity systems operating in unpredictable conditions.
• A unified computational approach offers a promising avenue for understanding the neural and cognitive underpinnings of decision-making.