Hi FSL experts
Can I ask a couple of questions:
I have a group of patients and healthy controls
I want to see whether network activity differs between the two groups at my 6 month outcome scan mainly in DMN but also in other selected networks
This is what I have done so far:
1. First level Melodic with automatic dimensionality estimation selected - full pre-processing
2. Gone through each subject/control using Melview and identified Signal or Noise
3. Run fix for each subject using text file output from Melview identifying signal/noise
4. Run a group concat with all patients/controls and automatic dimensionality estimation selected
5. Run a dual_regression
My questions are:
1. What is the best strategy for the group concat? Should I let Melodic decide the number of components or restrict them? If the latter, what would be an "acceptable" threshold in the imaging community? 15, 20, 25?
2. When I run my dual regression what is the most acceptable option:
a) Run a group concat on my healthy controls and use the output as my "model" networks to input into DR
b) Use a set of template networks e.g. Smith RSN 10 etc in the DR
I can identify DMN and selected other networks in my healthy control concat but some of the networks seem to have been split
Thanks!
Monty
|