Print

Print


Hi there:

I was very happily running fsl_motion_outliers to calculate dvars and refrms for all my datasets/studies using a little shell (bash) script that called fsl_motion_outliers (without motion correction) and output the usual output plus files of the metrics so I can concatenate them into one study array and use this in multivariate analysis as confound regressors at the second level.

I have just run into an issue that for one set of scans (for the normalized/warped scans), the output of all the fsl_motion_outliers (graphs, first level matrices and the metric matrices) is zero for every timepoint in every subject.

This dataset luckily has the following structure, so it allows comparison within subjects and across subjects:

Collection Site 1: Session 1 and Session 2
Collection Site 2: Session 1 and Session 2

The problem is only in Collection Site 1 and Session 2. The fsl code is running fine -- I ran with verbose output and I can see its performing the operations. Pre-processing was performed exactly the same for all subjects across sites and sessions, using SPM matlabbatch scripts.

The timepoints do not appear to be duplicated so I wondered if this was wonky NaN behavior.

I converted the nifti to MATLAB arrays using spm_vols and spm_read_vols. I then did a spot check inside a few subjects and found no NaNs in the non-problematic scans, but many in the problematic scans.

I am wondering if the fsl_motion_outliers function is exhibiting this behavior because it doesn't ignore NaNs and is zeroing out all the output? Or otherwise if this is a known behavior and there is a fix? Or if perhaps there is another explanation/suggestion?

Thanks very much,

Nina de Lacy

This message and any attached files might contain confidential information protected by federal and state law. The information is intended only for the use of the individual(s) or entities originally named as addressees. The improper disclosure of such information may be subject to civil or criminal penalties. If this message reached you in error, please contact the sender and destroy this message. Disclosing, copying, forwarding, or distributing the information by unauthorized individuals or entities is strictly prohibited by law.