Diffusion model analysis with MATLAB: A DMAT primer

Vandekerckhove and Tuerlinckx

There are several ‘fitting models’ available for the drift diffusion model (DDM), including EZ, RWiener, JAGS and DMAT. The article by Vanedekerckhove and Tuerlinckx (2008) is focused on DMAT, a fitting model that the authors propose for the DDM. DMAT stands for Diffusion Model Analysis Toolbox and is based specifically on Ratcliff’s diffusion model (RDM), one version in the DDM evolution. Vanedekerckhove and Tuerlinckx (2008) highlight the fact that the RDM, as a source for their fitting model, is prohibitively “difficult to apply in practice” and that this fact has generated a host of articles to assist researchers in using this model. In the article the authors briefly discuss the RDM, which is a more complex form of the typical DDMs provided by other authors.

The rest of the article is focused on taking the reader through the DMAT toolbox. The article comprises a series of sections on software requirements, installation procedures, interfaces, data set requirements, and use of the toolbox. The authors use two data sets (a simple data set and a more complicated data set) to walk the reader through the input and output of the DMAT program. This discussion is supported by several appendices to support the user through the process.

As the DMAT is highlighted in the literature as one of the fitting modules for the DDM, this article provides a useful and detailed overview of how to use DMAT. From this perspective the article is very useful. However, bearing in mind that the model was proposed in 2008 and that the DMAT model is now viewed as a “deprecated” – see https://ppw.kuleuven.be/okp/software/dmat, the value of this module and its complexity in dealing with the DDM must be question. Especially when considering more recent models for fitting data to the DDM, such as EZ (see Van Ravenswaaij, Donken and Vandekerckhove [2017]).

 

Reference:

Vandekerckhove, J. and Tuerlinckx, F. Diffusion model analysis with MATLAB: A DMAT primer. Behavior Research Methods, 2008, 40 (1), pp. 61-72.