Print

---

hi - am replying to my own question raised a few weeks ago,
namely how to relate resting state data 
to task-data I got from the same Ps 
From what I have gathered, one nice and simple 
approach in FSL would be this. Could you please
comment if this is sensible?

 I have a posterior parietal network
engaged during my experimental task, so one hypothesis I want to test
is that parietal responses at rest may correlate
with task-based responses.  To do this

1. Use concat ICA in my resting data to derive networks

2. Select the components with posterior parietal involvement

3. Use dual regression to get individual maps

4. Derive masks based on the task data around the peaks of posterior parietal clusters (PPCs).

5. Use fslmeants to derive an integration measure of these PPCs masks in the resting state data
within the selected ICs  involving parietal cortex

this would be something like  for each subject

fslmeants -i dr_stage2_[IC_InvolvingParietal] -o integrat_score.txt -m PPC_mask_from task data

This would give me for each P, the strenght of the correlation between the PPC voxels with the rest of the network

6. Extract the percent signal change from the PPC response in the task data and correlate it with the integration scores
as derived by (5)

Significant correlation would indicate that the degree of the integration of PPC at rest (e.g, within the right frontoparietal IC)
is 'predictive' of the PPC response during my task paradigm.


does this make some sense? thanks in advance of any advise

cheers, david