 |
 |
First of all, I would suggest checking the voxel sizes and origins of your header files (via the display button). The origin field should be the same in all the headers of the time series and should reflect the co-ordinates of a voxel that is somewhere near the AC. It is also possible that the ".mat" files of the images could have got messed up. If these files exist, then SPM ignores what's in the origin and voxel size fields (in the realignment, coregistration and spatial normalization), and just uses the information in the ".mat" files. I would suggest deleting these and starting afresh if you have any doubts about their validity. Good luck, -John > I am quite new to SPM, and have just started running some data. The > first few runs went well (running SPM under Matlab 4.2 on a Sun Ultra 10 > running Solaris). However, I just tried running new data and got an > error message when running Realign (including with just coregister > only). The errors messages, which repeat multiple times per volume in > the Matlab window, read as follows: > > "Warning: Rank deficient, rank = 4 tol = 2.1207e-07" > > The data consist of 48 volumes of 12 128x128 slices each acquired from a > Siemens Vision magnet. The 576 dicom images were imported into Analyze, > which flipped the z access to SPM orientation and wrote out one large > (576 slice) .img/.hdr volume. I then changed the header file to read 12 > slices (height=128, width=128, depth=12, volume=1) and ran a short > program provided on this listserver to generate the 48 separate volumes > of 12 slices each. I used this method successfully on a previous data > set, but to no avail with the most recent set (acquired in the same > format, same size, and from the same magnet). > > Each of the 48 files (xxx.img and xxx.hdr) can be read and displayed in > SPM OK, and the header info for each shows the correct 128 x 128 x 12 > matrix. However, the error messages are generated when the files are > coregistered. It sounds like I'm getting a matrix of insufficient rank > somewhere in the data; what are ranks to which SPM refers? Are they > based on the X x Y x Z header info? Any ideas? %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%