Comparing head motion between subject groups

[Negar Memarian]

Hi,

I'd like to compare rs-fMRI head motion between controls and patients and want to know if what I have done is correct:

For every subject, in order to calculate a single motion score that reflects the head motion across the entire scan session, I've averaged the values in:

MNINonlinear\Results\rest_acq-ap_run-01\Movement_RelativeRMS.txt 
and
MNINonlinear\Results\rest_acq-pa_run-02\Movement_RelativeRMS.txt

and then ran a two-sample t-test to compare the mean motion of the patient group against the control group. Does this sound right?

Thank you,

Negar

[mawaffak ing]

Dear Negar,

Your observation is very interesting.

Regarding the globally lower connectivity values in the Parkinson's group, I would first be cautious before interpreting this as a biological effect. A global shift in FC can sometimes arise from residual motion effects, differences in signal quality, preprocessing choices, global signal handling, or other confounding factors.

One useful way to investigate this is to move beyond simple edge-wise averages and examine the mathematical structure of the connectivity matrices themselves. For example:

• Eigenvalue spectrum of the FC matrices.
• Graph Laplacian spectrum.
• Network density and degree distributions.
• Global versus local connectivity measures.
• Matrix similarity measures between subjects and groups.

These analyses can sometimes distinguish a true network reorganization from a global scaling or preprocessing artifact.

If you are interested, I would be happy to take a look at a small anonymized sample of the FC matrices (for example a few subjects from each group) and perform some exploratory spectral and network analyses. In many cases, these analyses can reveal whether the observed group difference is driven by a genuine change in network structure or by a more global effect.

Best regards

--
You received this message because you are subscribed to the Google Groups "HCP-Users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hcp-users+...@humanconnectome.org.
To view this discussion visit https://groups.google.com/a/humanconnectome.org/d/msgid/hcp-users/CAC3GDx2Q__%2BQLirWkWpCtu%2B3q2SR%3D-2P%3DVP%3DsZa8Gs%3DLqo%3DrOg%40mail.gmail.com.