Incorrect ICBM-DTI-81 Atlas Orientation and White Matter Labels

Citation

Rohlfing, T. (2013). Incorrect ICBM-DTI-81 atlas orientation and white matter labels. Frontiers in Neuroscience, 7, 4. doi: 10.3389/fnins.2013.00004

Introduction

The ICBM-DTI-81 atlas (Mori et al., 2008Oishi et al., 2008)1 provides perhaps the most widely used reference coordinate system for group analyses of diffusion tensor MR images. The atlas comprises, among other data, structural images (e.g., T2-weighted), scalar maps of diffusion (e.g., fractional anisotropy), and a map of manually labeled major white matter fiber tracts.

Unfortunately, in our work using this atlas, we have encountered several substantial problems in terms of (a) correctness of spatial orientation and (b) correctness of white matter tract labels. These issues raise questions about the validity of published studies using the ICBM-DTI-81 atlas. We also detected substantial changes applied to the atlas files with no documentation or user notification. Because the changed files are distributed under the same name and the earlier versions are no longer available, it is virtually impossible to reproduce published research using the atlas.

We have contacted the atlas maintainers with our concerns and have been in active communication with them since to discuss possible fixes. As correcting the atlas turned out to take considerable time, however, and because it is not feasible to notify all users of the atlas directly, we feel that communicating our findings to the neuroimaging community at large is warranted. The atlas maintainers were apprised of our intent to write and publish this article and did not raise any objections. Thus, researchers using the ICBM-DTI-81 atlas can verify the validity of their analyses. To this end, we provide a discussion of the significance and potential impact of each problem identified, as well as detailed recommendations for confirming whether or not a particular study was affected.


Read more from SRI