JAGS-Wiener (often stylized as jags-wiener) is a free, open-source plugin extension for the Bayesian modeling software JAGS (Just Another Gibbs Sampler). It expands JAGS’s native capabilities by introducing probability distribution functions for the Wiener diffusion model, which is heavily used in cognitive science and psychology to analyze reaction times and decision-making processes. 💡 Core Functionality
The module allows researchers to implement Hierarchical Bayesian Diffusion Models. It introduces the first-passage time density of a drift-diffusion process directly into JAGS as a stochastic node, meaning you can use it in your code just like standard distributions (such as dnorm or dgamma). When loaded, it grants access to specialized syntax:
dwiener(alpha, tau, beta, delta): Represents the primary stochastic distribution node.
dlogwiener(x, alpha, tau, beta, delta): A logical node for dealing with log-densities. 📊 The Four Mathematical Parameters
The module models the classic Ratcliff diffusion model using four core parameters: Drift rate (
): Represents the speed and quality of information processing (cognitive workload/direction). Boundary separation (
): Dictates the response caution or the amount of evidence required before making a decision. Initial bias (
): Reflects starting-point bias toward one of the two decision options. Non-decision time (
): Accounts for the physiological time needed for sensory encoding and motor response output. 🛠️ Origin and Significance
Leave a Reply