Response Time Distributions [R package rtdists version 0.11-2]

Singmann, Henrik; Brown, Scott; Gretton, Matthew; Heathcote, Andrew · 2020 · Unknown

archive: archived pipeline: cataloged verified

Get this paper ↗ (full text — opens at the source; we link to it, we don't host it)

Summary

The `rtdists` R package provides computational tools for modeling response time (RT) distributions in two-choice discrimination tasks. It addresses the need for accessible, efficient implementations of two prominent cognitive models: the Ratcliff diffusion model and the Linear Ballistic Accumulator (LBA). These models are widely used in experimental psychology to analyze decision-making processes, but their mathematical complexity often requires specialized software. The package aims to facilitate the simulation, likelihood calculation, and parameter estimation for these models within the R statistical environment. The package implements core functions for both models, including probability density functions (PDF), cumulative distribution functions (CDF), quantile functions, and random number generation. For the diffusion model, the implementation is based on C code by Voss and Voss, supporting parameters such as threshold separation ($a$), drift rate ($v$), non-decision time ($t_0$), and various inter-trial variabilities ($s_v$, $s_z$, $st_0$). The LBA implementation supports multiple underlying distributions for the drift rate, including normal, gamma, Frechet, and log-normal. Both implementations are fully vectorized, allowing for trial-wise parameter variations and efficient computation. The package also includes helper functions for likelihood optimization, demonstrating how to recover model parameters from simulated data using numerical minimization techniques. Key findings from the package documentation and examples demonstrate that the functions accurately reproduce theoretical predictions. For instance, simulations using the diffusion model replicate results from Wagenmakers et al. (2007), showing expected mean response times and variances for different parameter settings. The package handles the bivariate nature of these models, where individual accumulators produce defective CDFs that do not reach 1; the sum of CDFs across all response boundaries equals 1. The quantile functions are designed to return predicted RTs up to the maximal probability of the accumulator’s CDF, with options to automatically scale probabilities for convenience. The examples confirm that parameters can be successfully recovered from simulated data using maximum likelihood estimation, validating the numerical accuracy of the density and distribution functions. The significance of `rtdists` lies in its provision of a unified, open-source framework for fitting complex cognitive models to empirical data. By offering interchangeable interfaces for both diffusion and LBA models, it allows researchers to compare these theoretical frameworks directly. The vectorized design and support for trial-wise parameters enhance flexibility in experimental design analysis. The package serves as a critical resource for cognitive psychologists and neuroscientists, enabling rigorous statistical inference on response time data without requiring custom coding for complex mathematical integrations.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed. Discovered via author_sweep_intake on 2026-05-28.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-10; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• perception reaction time