Hi,
I have an odd problem with the FSLEYES (FSL 5.0.10) display of some merged images. An artifact appears when displayed that is not present in the data. I've attached two sample nifti volumes for fun, subjects 267 and subjects 328, and the merged file that displays the artifact. Here's what happened.
1. The nifti files are registered copes from primates acquired at 9.4T in a pharma experiment. Conversion from 94T to nifti was done with in-house programs, and subject-level analysis performed.
2. Registration to standard was done using AFNI based JIP tools, a process that gives us standard space copes, varcopes, etc for placement in subject-wise reg_standard/ directories for use in group analysis.
3. If we merge the registered copes as is done in group analysis, the resulting file (filtered_func_data in the group analysis) shows some strange properties when displayed in FSLEYES.
4. If you load the data, volume 0 looks ok. Change to volume 1 and periodically spaced transverse slice discontinuities appear like horizontal lines in the sagittal and coronal views. If you look at such a slice in the transverse pane if looks like it was offset and chopped up. This is the artifact.
Some additional properties:
5. If you switch back to volume 0, lines might appear there that were not there before. If you then switch again to volume 1, more lines might appear - its a dynamic situation.
6. This does not happen in fslview_deprecated.
7. "filtering" images through fslmaths, fslview-save, and flirt does not help.
8. Merging the original pre-JIP images does not produce the artifact.
9. Merging the two sample files in swapped order does produce the artifact. Merging a file with itself does not produce the artifact.
Happy to send source single volume source images if anyone is interested
Mike Rohan
|