Diffusion Model Analysis Toolbox

DMAT is short for the Diffusion Model Analysis Toolbox. It is a MATLAB™ toolbox for fitting Ratcliff's diffusion model to reaction time and accuracy data. There are currently two releases of DMAT (one for MATLAB R14SP3 until R2007b and one for R2008a and newer versions).

The DMAT was created by Joachim Vandekerckhove and Francis Tuerlinckx of the Research Group of Quantitative Psychology and Individual Differences of the University of Leuven.

References and Manual

Information on how to use DMAT can be found in the included documentation, and an introduction to the toolbox can be found in this article:

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

The methods applied in DMAT are described in:

  • Vandekerckhove, J., & Tuerlinckx, F. (2007). Fitting the Ratcliff diffusion model to experimental data. Psychonomic Bulletin & Review, 14, 1011-1026. PDF

Finally, a Bayesian implementation for population data with small data per participant, where DMAT will sometimes fail:

  • Vandekerckhove, J., Tuerlinckx, F., & Lee, M.D. (2011). Hierarchical diffusion models for two-choice response times. Psychological Methods. 16, 44-62. PDF

Notice

DMAT is an old third-party MATLAB toolbox that is no longer maintained or updated for newer versions of MATLAB. It should be considered deprecated. For recent versions of MATLAB (since 2014a) it is untested and may be broken.

Researchers interested in applying diffusion models should instead look into the hierarchical Bayesian version in Vandekerckhove et al. (2011) using the JAGS plugin available via http://works.cidlab.com/.

disclaimer By clicking on any of these links (including, but not limited to, the downloading of any PDF file from any server), you agree to forfeit and cede on your own behalf and on behalf of any past, current, or future employer, any exclusive claim to intellectual property or copyright relating to any of the materials so obtained.