A tbss_sym between group analysis (n= 18; 9 controls, 9 experts) was completed (-n 5000) with the basic contrasts: C1 controls > experts [ 1 -1 ] , C2 experts > controls [-1 1]. One significantly different cluster was found, and L>R in C2 for the experts. After making a mask of corrp_tstat2 in fslmaths (fslmaths LgR5000nocv_tfce_corrp_tstat2.nii.gz -thr 0.90 -bin peakfa_tstat2_th.nii.gz.txt), and using it to extract FA values thresholded at 0.90 (p < .10) in fslmeants (fslmeants -i all_FA_skeletonised_left_minus_right.nii.gz -m peakfa_tstat2_th.nii.gz -o mean_sign_vox.txt), the controls fslmeants output values were negative (expect for one control), and the experts were all positive except for one. The fslmeants C2 output is below. C1 (controls > experts, 1 -1) had values of zero. While this indicates experts have higher FA at this cluster than controls, are the negative values actually positive relative to the contrast, here C2? If so, I assume the all values would be made positive for further analyis in matlab or R?
con -0.066070757
con -0.093227464
con -0.052008955
con -0.073162818
con -0.091825447
con -0.142750694
con -0.180685184
con 0.088013128
con -0.043995853
exp 0.099649527
exp 0.018109312
exp 0.128316277
exp 0.147246512
exp 0.024393221
exp 0.114848495
exp 0.028018308
exp -0.004780046
exp 0.07381856
|