help
help > RE: MNI dimensions not the same
Apr 21, 2014 06:04 AM | Alfonso Nieto-Castanon - Boston University
RE: MNI dimensions not the same
Hi Mary,
The dimensions/coverage and voxel sizes of any of your files (structural, functional, and ROIs) may be different. As long as all of these files are coregistered (in the same space) the analyses should run without a problem (see https://www.nitrc.org/forum/message.php?msg_id=8875 and https://www.nitrc.org/forum/message.php?... for more details).
Regarding the dimensionality of the analyses in conn, that is defined in the 'Setup->Options->Spatial resolution' field. If you use 'same as template' the analyses will be performed on [91 109 91] voxels with 2mm voxel size. Regarding the dimensionality of the volumes resulting from the normalization/segmentation spatial preprocessing steps, that is controlled by SPM's 'defaults.normalise.write.bb' and 'defaults.normalise.write.vox' fields (the former controls the bounding box size, while the latter controls the voxel resolution). You can modify the default values directly (in spm_defaults.m) or dynamically, using for example the following line (before running your spatial preprocessing steps):
Hope this helps
Alfonso
Originally posted by Mary Newsome:
The dimensions/coverage and voxel sizes of any of your files (structural, functional, and ROIs) may be different. As long as all of these files are coregistered (in the same space) the analyses should run without a problem (see https://www.nitrc.org/forum/message.php?msg_id=8875 and https://www.nitrc.org/forum/message.php?... for more details).
Regarding the dimensionality of the analyses in conn, that is defined in the 'Setup->Options->Spatial resolution' field. If you use 'same as template' the analyses will be performed on [91 109 91] voxels with 2mm voxel size. Regarding the dimensionality of the volumes resulting from the normalization/segmentation spatial preprocessing steps, that is controlled by SPM's 'defaults.normalise.write.bb' and 'defaults.normalise.write.vox' fields (the former controls the bounding box size, while the latter controls the voxel resolution). You can modify the default values directly (in spm_defaults.m) or dynamically, using for example the following line (before running your spatial preprocessing steps):
spm_get_defaults('normalise.write.bb',[-90,-126,-72;
90,90,108]);
Hope this helps
Alfonso
Originally posted by Mary Newsome:
Hello Alfonso and everyone,
I noticed that the normalized (and smoothed) images that Conn produces have the dimensions, 79x95x68. In SPM (without Conn), dimensions of normalized (and smoothed) images are 91x109x91. Do the images normalized by Conn have MNI dimensions (i.e., is it ok that the dimensions are different in SPM)? I noticed that some of the segmented normalized files do have 91x109x91 as their dimensions. It seems that it wouldn't be good if the smoothed and segmented files do not have the same dimensions. I may be wrong and would really appreciate anyone's correction or comments.
Thanks!
I noticed that the normalized (and smoothed) images that Conn produces have the dimensions, 79x95x68. In SPM (without Conn), dimensions of normalized (and smoothed) images are 91x109x91. Do the images normalized by Conn have MNI dimensions (i.e., is it ok that the dimensions are different in SPM)? I noticed that some of the segmented normalized files do have 91x109x91 as their dimensions. It seems that it wouldn't be good if the smoothed and segmented files do not have the same dimensions. I may be wrong and would really appreciate anyone's correction or comments.
Thanks!
Threaded View
| Title | Author | Date |
|---|---|---|
| Mary Newsome | Mar 17, 2014 | |
| Alfonso Nieto-Castanon | Apr 21, 2014 | |