Hi Gang
I am having another problem with 3dRegAna. At the moment I am trying to correlate a number of behavioural measures with brain activation in a clinical and a control group. So far, I have been succesful in carrying out correlations with percent correct responses and also, in the clinical group, with measures of symptom severity. However, when I try to correlate brain activation with correct response reaction time I am geeting the following error:
3dRegAna Error: Error in calculating F - lack of fit
I presume that this means that the model I am using is not suitable. My command is as follows:
3dRegAna \
-rows 12 \
-cols 1 \
-xydata 930 sm00066.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 821 sm0090.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 758 sm00101.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 1297 sm00167.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 918 sm00178.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 1048 sm00325.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 971 sm00764.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 657 sm00775.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 896 sm00830.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 841 sm00863.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 975 sm00874.ThreeRuns.dec.VSAT+tlrc.HEAD \
-xydata 1743 sm6406.ThreeRuns.dec.VSAT+tlrc.HEAD \
-fdisp 10 \
-model 1 : 0 \
-flof 0.01 \
-fcoef 0 CorrRT.CON.constant \
-fcoef 1 CorrRT.CON.linear \
-bucket 0 CorrRT.VSAT.CON.corr
Should I change my model and if so to what?
Thanks in advance for your help!