Speeded multielement decision-making as diffusion in a hypersphere: Theory and application to double-target detection

Area of Science:

• Cognitive psychology
• Computational neuroscience
• Psychophysics

Background:

• Existing models of decision-making often simplify visual stimuli.
• Multielement visual displays present complex challenges for cognitive models.

Purpose of the Study:

• To generalize a 2D diffusion model to a hyperspherical model for multielement visual displays.
• To model evidence accumulation as a vector-valued diffusion process.
• To predict response time and accuracy in speeded decision-making tasks.

Main Methods:

• Generalization of a 2D circular diffusion model to higher dimensions.
• Application of a 4D hyperspherical diffusion model to the double-target deficit paradigm.
• Analysis of response time and accuracy distributions as a function of task difficulty.

Main Results:

• The 4D hyperspherical diffusion model accurately predicted performance in single- and double-target conditions.
• Model parameters suggest that decision templates encode configural stimulus properties related to target count.
• The model successfully generalized to predict performance in higher-dimensional decision spaces.

Conclusions:

• The hyperspherical diffusion model offers a robust framework for understanding speeded decision-making with multielement displays.
• The model provides insights into the mental representation of decision alternatives, termed 'decision templates'.
• This approach has broad potential for studying multiattribute decisions in various cognitive domains.