help
help > RE: too much movement?
Feb 26, 2021 05:02 PM | Alfonso Nieto-Castanon - Boston University
RE: too much movement?
Hi Amy,
In Setup.Covariates (2nd-level) you will find (assuming the data was preprocessed in CONN) the variables QC_InvalidScans (number of outlier scans identified for each subject), QC_MaxMotion (maximum FD for each subject), and QC_MeanMotion (average FD across non-outlier scans for each subject), which you may use to determine whether it may be necessary to remove/disregard some subjects from your group-level analyses (e.g. if you run in CONN the Quality Assurance plot labeled "distribution of subject-level QC measures", that will list potential outlier subjects based on each of the selected measures - I typically recommend using QC_ValidScans and QC_MeanMotion).
Best
Alfonso
Originally posted by Amy Roy:
In Setup.Covariates (2nd-level) you will find (assuming the data was preprocessed in CONN) the variables QC_InvalidScans (number of outlier scans identified for each subject), QC_MaxMotion (maximum FD for each subject), and QC_MeanMotion (average FD across non-outlier scans for each subject), which you may use to determine whether it may be necessary to remove/disregard some subjects from your group-level analyses (e.g. if you run in CONN the Quality Assurance plot labeled "distribution of subject-level QC measures", that will list potential outlier subjects based on each of the selected measures - I typically recommend using QC_ValidScans and QC_MeanMotion).
Best
Alfonso
Originally posted by Amy Roy:
I am new to CONN. Previously when I analyzed
resting state functional connectivity data I would examine the
mean, max and min FD for each participant to determine whether to
use their data. I cannot seem to find outputs in CONN that provide
this information. So, how can I tell if a participant has too much
motion to be included in the analyses? Is it assumed that because
CONN uses scrubbing, that all participants are used regardless of
the amount of movement? I would appreciate some guidance here as I
am willing to remove participants with excessive movement but I am
not sure how to determine this using the preprocessing outputs in
CONN. I have patient groups where group membership may be related
to movement so I need to be sure this is appropriately accounted
for.
Thank you
Amy
Thank you
Amy
Threaded View
| Title | Author | Date |
|---|---|---|
| Amy Roy | Feb 26, 2021 | |
| Alfonso Nieto-Castanon | Feb 26, 2021 | |