Hi FSL experts,
I am learning using Aroma to denoise motion-related artefacts in data from one stroke patient. I understand that Aroma registers IC components to MNI 2mm space to identify spatial patterns, without registering time series to the standard space. My question is, will lesions in the patient's brain affect this registration? if they do, should I mask the lesion before preprocessing?
Also, in the final stats analysis, we only register the functional image to the patient's structural, without registering to a standard space. In this case, I assume the processing steps would be:
Step 1: preprocessing before Aroma: motion correction, slice timing correction, spatial smoothing, turning off highpass temporal filtering, 6DOF func2highres, 12DOF highres2standard.
Step 2: run Aroma, using the .feat directory from Step 1.
Step 3: redo preprocessing, this time turning on highpass filtering and turning off "standard space" in registration. Then do stats as normal.
I appreciate if someone can confirm if these steps are correct.
Thank you,
Yuqi
|