Generalized Linear Regression with DDM.
RegDDM
Build Regression models over Drift Diffusion Model parameters using MCMC!
Installation
You can install latest version of RegDDM using Github. The package will later be available on CRAN.
remotes::install_github("biorabbit/RegDDM")
Example
First, load the package and the example dataset.
library(RegDDM)
data(regddm_tutorial)
data1 is the subject-level dataset:
head(regddm_tutorial$data1)
#> id y c1 c2
#> 1 1 1.9690519 0.08461457 a
#> 2 2 2.6410850 1.82427245 <NA>
#> 3 3 5.1843542 1.23414213 b
#> 4 4 -1.1623707 NA c
#> 5 5 0.9845534 1.77316247 a
#> 6 6 2.0520609 1.37139039 b
data2 is the subject-level dataset:
head(regddm_tutorial$data2)
#> id x1 x2 rt response
#> 1 1 0.4038328 a 0.7533853 1
#> 2 1 -0.8707744 b 0.7314780 1
#> 3 1 1.5737835 c 0.8965344 1
#> 4 1 1.5112327 a 0.9395178 1
#> 5 1 -0.8122571 b 0.6522295 1
#> 6 1 1.1721147 c 0.6013884 0
Specify the model using a list. In this example, the drift rate is influenced by x1. The subject’s outcome y is predicted by baseline drift rate v_0 (drift rate when x1 is 0), the influence of x1 on drift rate v_x1 and covariate c1:
model = list(
v ~ x1,
y ~ v_0 + v_x1 + c1
)
Use the main function of RegDDM to automatically generate the RStan model and summary the results. This could take ~20 minutes to run. The rows starting with ‘beta_’ are the posterior distributions of regression parameters:
fit = regddm(
regddm_tutorial$data1,
regddm_tutorial$data2,
model
)
print(fit)
#> RegDDM Model Summary
#> Number of subjects: 30
#> Number of trials: 3000
#> Model:
#> v ~ x1
#> y ~ v_0 + v_x1 + c1
#> Family: gaussian
#> Sampling: 4 chains, 500 warmups and 1000 iterations were used. Longest elipsed time is 639 s.
#>
#> Regression coefficients:
#> variable mean sd 2.5% 97.5% n_eff Rhat
#> 1 beta_0 1.551 0.992 -0.3261 3.551 1413 0.998
#> 2 beta_v_0 -0.851 0.551 -1.9479 0.253 1669 0.999
#> 3 beta_v_x1 0.917 0.202 0.5067 1.309 3290 0.998
#> 4 beta_c1 0.918 0.389 0.0827 1.661 1791 0.998
#> 5 sigma 1.131 0.180 0.8394 1.538 2068 1.000
#> Maximum R-hat: 1.005
In this example, the outcome is positively correlated with v_x1 and c1, but not v_0. The higher the influence of x1 on drift rate and the higher the covariate, the higher the outcome y.
Using your own data!
If you want to fit the model on your own data, you need to specify data1, data2 and model.
data1 is subject-level data table. It should contain the following: * id: unique indexing column for each subject. * other subject-level variables that we want to include in the regression. Missing value is supported
data2 is trial-level data table. It should contain the following: * id: the subject of each trial using the same index in data1. * rt: response time of the trial in seconds. * `response``: response the trial. must be either 0 or 1. * trial-level variables. These are the variables that differ by trial, such as difficulty of the task or different numbers on the screen. We assume that subjects’ behavior changes according to these variables. These variables cannot contain missing values.
model is the proposed dependency between these parameters. Default is an empty list. It must be a list of 0 - 5 formulas. The outcome of these formulas can be either: * one of the four DDM parameters a, t, z, v, modeling the relationship between DDM parameters and trial-level variables. * one formula for GLM regression, modeling the relationship between estimated DDM parameters and other subject-level variables.
family is the family of distribution of GLM. It can be either "gaussian", "bernoulli" or "poisson". Default is "gaussian".
init is how we initialize the MCMC algorithm. The "default" initialization should work in most conditions
prior determines whether to use the default prior for DDM parameters or not. Default is TRUE
stan_filename is the file loaction for the automatically generated stan model. If an empty string ’’ is provided, a temporary file will be created and deleted after the model is fit. Default is "stan_model.stan"
gen_model determines whether to generate the model or not. Default is TRUE.
fit_model determines whether to fit the model or not. Default is TRUE.
...: additional parameters used by rstan, including warmup,iter,chains,cores etc.
Citation
to be added.