====== Correct Order for Dimensions in the Volume Anatomies ====== In general, if you want to check whether your volumes have the right orientation, try to load it into [[:mrTools:mlrImage#mlrVol|mlrVol]]. The instructions here are a bit out-of-date and specific to volume anatomies from Center for Brain Imaging at NYU and can be ignored by other users. [Please Note: this should not be relevant for most users. It is only relevant for those who want to align across subjects, using the Talairach transformation. The script that calculates the transformation requires a particular order for the dimensions.]. ===== Introduction ===== There are 3 dimensions to any volume anatomy : L-R, P-A, I-S (left to right; posterior to anterior; inferior to superior). Our standard is for them to be saved in that order. This order is called 'LPI'. To be specific, the data matrix (M(x,y,z)) that contains the values that make up the volume anatomy, should: 1) Have dimensionality 176 x 256 x 256 2) The first dimension (x) should move through the anatomy from left to right 3) The second dimension (y) should move through the anatomy from back to front 4) The third dimension (z) should move through the anatomy from bottom to top ===== How to check your volume anatomy ===== To check your volume anatomy, you will load it using cbiReadNifti, and then view one slice at a time from each dimension. === 1: Load it === First, load it into matlab: [d, h] = cbiReadNifti('Your3DAnatomy.hdr'); Check the dimension sizes of d. d should be 176 x 256 x 256. I usually do this by using the command whos === 2: Visually examine it === You want to look at a slice through each dimension in order: figure(1), imagesc(squeeze(45,:,:))); figure(2), imagesc(squeeze(:,70,:))); figure(3), imagesc(squeeze(:,:,70))); Figure 1 should be sagittal, with the top of the head on the right, nose down; Figure 2 should be coronal, with the top of the head on the right, moving from L to R; Figure 3 should be axial, with the nose to the right, moving from bottom to top For example, they should look something like this: Figure 1