A Poisson random walk model of response times

Area of Science:

• Cognitive Psychology
• Computational Neuroscience

Background:

• The theory of visual attention (TVA) explains visual categorization performance.
• Existing models may not fully capture response times for confusable stimuli.

Purpose of the Study:

• To extend TVA with a response time model based on continuous evidence sampling.
• To introduce the Poisson random walk model for visual categorization tasks.

Main Methods:

• Developed a Poisson random walk model extending TVA principles.
• Tested the model with speeded and non-speeded binary and multiple-category response tasks.

Main Results:

• The Poisson random walk model accurately accounts for response time distributions in visual categorization.
• Model parameters are interpretable and mathematically tractable.
• Model predictions align with TVA's perceptual categorizations.

Conclusions:

• The Poisson random walk model offers a unifying framework for attention and response times in visual categorization.
• The model successfully explains performance across various experimental conditions.