Normalizing DARTEL Templates to MNI Space

University of Wisconsin – Madison

August 24, 2009


The Diffeomorphic Anatomical Registration Through Exponentiated Lie Algebra (DARTEL, Ashburner 2007) Toolbox is a high dimensional warping process that increases the registration between individuals, which results in improved localization and increased sensitivity in analyses.

The DARTEL Toolbox in SPM5 outputs normalized data in its own space, which is typically smaller than MNI space. Thus, a second normalization step is necessary to bring the DARTEL processed images into MNI space. Importantly, the second normalization should not resample the data as that decreases the SNR by smoothing the data between adjacent voxels.

To avoid resampling twice, we have created normalize_DARTEL.m, a small matlab script that convolves the transformations between subjects and MNI space while using DARTEL.

What normalize_DARTEL.m does

This script will apply an affine transformation to warped DARTEL images. The output files are not resampled, but instead have modified headers. This is done to avoid the effect of smoothing during the resampling process.

The default option is not to modulate the result. If you wish to apply modulation, then use normalize_DARTEL(1).

Output space must be MNI152 or 112RM-SL. Use 112RM-SL for the rhesus macaque space. Additional space options require that the code be changed at line 40. MNI152 is the default option. If you want to change the space, you must also specify the modulation option of 0 or 1.

How to use our script

  • Estimate the normalization parameters between the Template 6 and the MNI grey matter priors using the “Normalize” and turn off the non-linear options to perform an affine registration.
  • Use the resulting *_sn.mat file and the normalize_DARTEL.m script to normalize the DARTEL data to MNI space without resampling.

For general DARTEL processing details, see the DARTEL guide

Terms of Use

This script is intended for free use by the neuroimaging community. However, we do ask that you please cite the following reference in any publications which make use of this script:

McLaren, DG, Kostmatka, KJ, Kastman, EK, Bendlin, BB, and Johnson, SC. Rhesus Macaque Brain Morphometry: A Methodological Comparison of Voxel-Wise Approaches. Methods 2009 (in press).

Canu, E, McLaren, DG, Fitzgerald, ME, Bendlin, BB, Zoccatelli, G, Alessandrini, F, Pizzini, FB, Ricciardi, GK, Beltramello, A, Johnson, SC, and Frisoni, GB. Microstructural diffusion changes are independent of macrostructural volume loss in moderate to severe Alzheimer’s disease. Journal of Alzheimer’s Disease (in press).

Use of this script or any derivative scripts for monetary gain, without prior written consent, is prohibited.


Download the matlab script: normalize_DARTEL.m

Last Updated: 12-14-2009