Hi Sascha,

So I went back and re-calculated the global signal (in case I had overlooked something/did something incorrectly). However, the situation stands as such.

Yes, I left those usual confounding factors as such and included the brainmask.nii file as an ROI, using "Inf" as the number of dimensions (all components).

I have used the REX_Subjectxxx_*.mat file for calculating the global signal (params.ROIdata)

I am attaching a zip file containing the following:

- A MATLAB variable having the REX variable loaded ('params'), GS calculated by Conn ('conn_globalsignal_psc'), GS calculated by my function ('global_signal_psc'), and the GS (PSC) calculated using your scripts ('global_signal_sascha_psc')
- An Excel sheet having those three variables as columns which highlights the differences between them (note that the difference between conn calculated GS and the other two GS is larger than between the other two)
- A simple plot of the three global signal (you can't see the global signal calculated by my script because the other signal almost exactly matches)


Best
Pravesh


Originally posted by Sascha Froelich: