I am using epireg called through the FEAT gui to perform motion correct, distortion correction and boundary based registration for preprocessing resting state fMRI data. Everything seems fine but I have noticed that in about 10% of my data there appears to be approximately a 1% change in the global signal intensity around the reference volume used for motion correction. The global signal intensity change is not present in the original data. The change is positive at it approaches the reference volume and then switches signs after the reference volume. The signal change then returns back to the normal mean for volumes further away from the reference.
When I run fsl_motion_outliers on this data with the -nomoco option it identifies these central volumes as outliers. If I run the fsl_motion_outliers on the original data without the -nomoco flag it identifies only a few outliers but they are not clustered around the reference volume. I am hoping that someone could explain this behavior.
I have also run the data using the first volume as a reference. This resulted in a different set of outliers being identified. The data sets that I am working with have very little motion (less than 0.5mm). I am also using a non-FSL tool (ART https://www.nitrc.org/projects/artifact_detect/). This tool is also identifying outliers but only based upon the change in global signal intensity around the reference volume.
It seems that these small global signal intensities changes is causing the outlier detection tools to identify a large portion of outliers (20-42) volumes out of 190. I am still trying understand and isolate the problem. Any suggestion on why this is occurring would be greatly appreciated.
Has anyone else encountered this issue? What is the appropriate way to identify outliers? Should I just do it on the raw data and not worry about the changes in global signal intensity. I suspect that these changes won't affect the analysis because they will be regressing them out with CompCorr.
Thanks for your help and advice.
Bob
Original data with motion correction
➜ fsl_motion_outliers -i epi.nii.gz -o fslmo.txt -p dvars.png -s dvars.txt
➜ cat -n fslmo.txt | grep " 1 "
96 0 1 0 0 0 0
97 0 0 1 0 0 0
117 0 0 0 1 0 0
126 0 0 0 0 1 0
138 0 0 0 0 0 1
Original data without motion correction
➜ fsl_motion_outliers -i epi.nii.gz -o fslmo.txt -p dvars.png -s dvars.txt --nomoco
➜ cat -n fslmo.txt | grep " 1 "
96 0 1 0 0 0 0
97 0 0 1 0 0 0
126 0 0 0 1 0 0
138 0 0 0 0 1 0
143 0 0 0 0 0 1
The outliers identified are the same with and without the motion correction flag. After I perform EPIREG and BBR the situation is very different.
Output of bold.nii after epireg. Reference is center volume
➜ fsl_motion_outliers -i filtered_func_data.nii.gz -o fslmo.txt -p dvars.png -s dvars.txt --nomoco
➜ cat -n fslmo.txt | grep " 1 "
83 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
84 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
85 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
91 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
92 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0
93 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0
94 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0
95 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0
96 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0
97 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
98 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0
99 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
100 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0
101 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
102 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
103 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
173 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
|