Simultaneous modeling of choice, confidence, and response time in visual perception.

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Decision Science

Background:

• Simultaneously modeling choice, reaction time, and confidence is a challenge in decision-making research.
• Existing models often struggle to integrate these three key aspects of cognition.
• The drift-diffusion model provides a foundation for understanding evidence accumulation in decision tasks.

Purpose of the Study:

• To introduce and validate the dynamical weighted evidence and visibility (dynWEV) model.
• To account for choice, reaction time, and confidence judgments within a unified framework.
• To extend the capabilities of drift-diffusion models in explaining decision-making processes.

Main Methods:

• Proposed the dynWEV model, extending the drift-diffusion model.
• Incorporated postdecisional evidence accumulation and parallel stimulus reliability assessment.
• Tested model fits against two-stage dynamical signal detection theory and race models in two perceptual experiments (motion and orientation discrimination).

Main Results:

• The dynWEV model provided acceptable fits for choices, confidence, and reaction time.
• Compared to other models, dynWEV demonstrated superior performance in capturing these decision-making aspects.
• Experimental data from motion and orientation discrimination tasks supported the dynWEV model's predictions.

Conclusions:

• Confidence judgments are influenced by both choice evidence and stimulus discriminability.
• Postdecisional accumulation of evidence plays a role in confidence formation.
• The dynWEV model offers a more comprehensive approach to modeling simultaneous choice, reaction time, and confidence in decision-making.