ECP fMRI Update
Newly Processed Data
- Joe has processed a lot of subjects
- 175 of the ECP subjects
- Previous folder had 176 subjects
So are the data any better?
Many Comparisons
36P to 36P
Train cross validated SVM classifier to classify Left vs Controls
Use covariance
Previous
New
Guess = 67/120 (55.83)
Classifier = 87/120 (72.5)
Guess = 63/116 (54.31)
Classifier = 81/116 (69.83)
Previous - aCompCor
New - Aroma
Guess = 67/120 (55.83)
Classifier = 90/120 (75)
Guess = 63/116 (54.31)
Classifier = 63/116 (54.31)
How does cortical thickness look for controls?
- 58 control ECP participants
- S1200 thickness
- S1200 thickness
SOE + CAT data