Hi Xinyuan,

The issue is that the model is not being estimated properly. I suspect that when you ran the SPM=spm_estimate_PPI command a second time, it was using the gPPI SPM.mat and you were in that directory.

If you still have the SPM.mat file generated by gPPI, that did not estimate, I can take a look at it. Alternatively, if you send me the original SPM.mat file, I can take a look at what the issue is in the data.

Did you have P.WB=1 or P.WB=0? That might be the first thing to check.

Best,
Donald