An exemplar-based random walk model of speeded classification

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Perception

Background:

• Existing models of categorization and automaticity have limitations in predicting response times.
• There is a need for a unified model that integrates exemplar-based retrieval and decision-making processes.

Purpose of the Study:

• To propose and validate an exemplar-based random walk model for predicting response times in speeded, multidimensional perceptual classification.
• To integrate principles from Nosofsky's generalized context model and Logan's instance-based model.

Main Methods:

• Developed a computational model combining exemplar retrieval and a random walk decision process.
• Simulated classification tasks using the proposed model.
• Compared model predictions with empirical data on response times.

Main Results:

• The model accurately predicts the influence of within- and between-category similarity on response times.
• It accounts for the effects of individual-object familiarity and extended practice.
• The model demonstrates successful integration of categorization and automaticity concepts.

Conclusions:

• The exemplar-based random walk model offers a robust framework for understanding perceptual classification and response time.
• This model bridges the gap between categorization and automaticity research.
• It provides a computational account for how memory retrieval and decision processes interact.