Hi Catherine, As far as I know, what you see from atlasquery is not a proportion of content of voxel (or ROI), but the (estimated) probability of the voxel to be included in a certain class (region or label) considering individual variability since the atlas was made by averaging tract-masks of 28 normal subjects (http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2724595/). You may want to check to probabilistic maps of WM atlas to get a more clear idea. Also note that the classification of WM bundles are not exclusive, so one voxel (124,91,104) has 76% of SLF and 53% of SLF (temporal), which can be included in both regions if we threshold > 0.50. The sum of the given probabilities isn't necessary 1. Best, --- Seung-Goo KIM On 2, Apr, 2013, at 10:26 PM, SUBSCRIBE FSL Nucleos wrote: > Hi FSLers, > > When the atlasquery output states 35% SLF_R (temporal part) and 32% SLF_R, does this mean that 67% is SLF_R with 35% of that specifically temporal part. > > OR > > Does it mean that at that ROI it is 35% SLF_R (temporal part) and 32% of that SLF_R (temporal part) is SLF_R. > > Just looking for clarification. > > Thank you, > Catherine