The suboptimality of perceptual decision making with multiple alternatives

Area of Science:

• Cognitive Psychology
• Neuroscience
• Computational Modeling

Background:

• Human perceptual decision-making is known to be suboptimal.
• Previous studies often used simplified two-alternative tasks, limiting understanding of naturalistic decisions.
• The precise nature and origins of this suboptimality are not well understood.

Purpose of the Study:

• To investigate the suboptimality of human perceptual decision-making in tasks with multiple alternatives.
• To identify the decision-level representations underlying these choices.
• To quantify information loss during sensory-to-decision transformations.

Main Methods:

• Conducted four experiments using tasks with multiple stimulus alternatives.
• Employed computational modeling to analyze decision-level representations.
• Quantified information loss between sensory and decision representations.

Main Results:

• All experiments demonstrated robust suboptimality in decision-making.
• A significant amount of information from sensory representations is lost.
• This loss occurs during the transformation to a decision-level representation.

Conclusions:

• Human decision-making with multiple alternatives exhibits severe limitations in representation quality.
• Substantial information loss compromises perceptual decision accuracy in naturalistic settings.
• Findings highlight critical constraints on how the brain processes complex choices.