https://github.com/dominiquemakowski/easyrt
Tools and examples for fitting (Hierarchical) Drift Diffusion Models in R
https://github.com/dominiquemakowski/easyrt
bayesian brms ddm decision diffusion drift error hddm lba linear making r reaction rt time
Last synced: 2 months ago
JSON representation
Tools and examples for fitting (Hierarchical) Drift Diffusion Models in R
- Host: GitHub
- URL: https://github.com/dominiquemakowski/easyrt
- Owner: DominiqueMakowski
- License: mit
- Created: 2023-05-16T17:18:34.000Z (about 2 years ago)
- Default Branch: main
- Last Pushed: 2023-07-11T10:26:00.000Z (almost 2 years ago)
- Last Synced: 2025-03-17T12:21:30.314Z (2 months ago)
- Topics: bayesian, brms, ddm, decision, diffusion, drift, error, hddm, lba, linear, making, r, reaction, rt, time
- Language: TeX
- Homepage: https://dominiquemakowski.github.io/easyRT/
- Size: 5.17 MB
- Stars: 10
- Watchers: 2
- Forks: 0
- Open Issues: 1
-
Metadata Files:
- Readme: README.Rmd
- License: LICENSE
Awesome Lists containing this project
README
---
output: github_document
editor_options:
chunk_output_type: console
---
# easyRT
*Tools and examples for modelling Reaction Times in R*
**[WORK IN PROGRESS]**
```{r setup, echo = FALSE, warning=FALSE, message=FALSE}
options(digits = 3)
library(easyRT)
library(tidyverse)
library(patchwork)
knitr::opts_chunk$set(
collapse = TRUE,
dpi = 450,
fig.path = "man/figures/"
)
```
```{r echo=FALSE, fig.height=5, fig.width=8, message=FALSE, warning=FALSE, dpi=600}
set.seed(3)
sim <- ddm_data(n=2000, drift = c(0, 1.3), bs = 1, bias = c(0.4, 0.6), ndt = c(0.15, 0.2), n_traces = 5, ndt_var=0)
sim$data <- mutate(sim$data, condition = fct_recode(condition, "Speed" = "1", "Accuracy" = "2"))
sim$traces <- mutate(sim$traces, condition = fct_recode(condition, "Speed" = "1", "Accuracy" = "2"))
sim$density <- mutate(sim$density, condition = fct_recode(condition, "Speed" = "1", "Accuracy" = "2"))
(ddm_plot_upper(sim$data, breaks = 150, density = sim$density, xlim = c(0, 1)) +
scale_color_manual(values = c("#3949AB", "#FF5722")) +
scale_fill_manual(values = c("#3949AB", "#FF5722"))) /
(ddm_plot_traces(sim$traces, trace_alpha = 1, trace_linewidth = 0.2, xlim = c(0, 1)) +
scale_color_manual(values = c("#3949AB", "#FF5722"))) /
(ddm_plot_lower(sim$data, breaks = 150, density = sim$density, xlim = c(0, 1)) +
scale_color_manual(values = c("#3949AB", "#FF5722")) +
scale_fill_manual(values = c("#3949AB", "#FF5722")))
```
## Motivation
This repo started as my attempt at understanding and implementing sequential models, starting with (Hierarchical) Drift Diffusion Models (DDMs) for reaction times in R. **Please don't hesitate** to open an issue to discuss and suggest things that could be improved or clarified.
## Content
Reaction time (RTs) have been traditionally modelled using traditional linear models (e.g., ANOVAs). However, it is problematic because RTs are **not** normally distributed. A popular mitigation method is to **transform** the data (e.g., by log-transform), but it is **not a good idea** ([Schramm & Rouder, 2019](https://doi.org/10.31234/osf.io/9ksa6)). Instead, one should use statistical models that **describe** or **generate** RT-like data.
You should start by reading:
- [**Lindelov's overview of RT models**](https://lindeloev.github.io/shiny-rt/): An absolute must-read.
- [**De Boeck & Jeon (2019)**](https://www.frontiersin.org/articles/10.3389/fpsyg.2019.00102/full): A paper providing an overview of RT models.
This repository contain the following vignettes:
- [**Drift Diffusion Model (DDM) in R: A Tutorial**](https://dominiquemakowski.github.io/easyRT/articles/ddm.html)
- [**Ex-Gaussian models in R: A Tutorial**](https://dominiquemakowski.github.io/easyRT/articles/exgaussian.html)
*Note: these are work-in-progress. Please get in touch if you want to contribute and help, it's a good way of learning.*
## Installation
```{r eval=FALSE}
remotes::install_github("DominiqueMakowski/easyRT")
library(easyRT)
```
## What does this package do?
Not much. It is mostly about its vignettes, but it also has some convenience functions to generate and plot drift diffusion models. We'll see how it evolves over time.
```{r warning=FALSE, message=FALSE, fig.width=5, fig.height=8, dpi=450}
sim <- ddm_data(drift = c(0, 1), bs = 1, bias = 0.5, ndt = 0.2)
ddm_plot(sim)
```