Hi all!

Recently, I have been conducting a project comparing Functional Connectivity and Effective Connectivity (with DCM). For that reason, I used CONN for the FC analysis and want to re-use the denoised *.nii file from CONN, and put it into SPM for DCM, so that I can control all other variables in preprocessing for a strict comparison between FC and EC.

However, I wasn't able to get the exact same time-series data from SPM (when compared to CONN). In the attached screenshot, for one ROI from one subject/session, you can see that the time series from SPM and CONN are highly correlated (R = 0.987) but not identical. 

For CONN, the time series was extracted from the /results/preprocessing/ROI_SubjectXXX_Condition000.mat file, within the "data" variable; for SPM, the time series was extracted with the batch function "fMRI Model Specification">"Model Estimation">"Volume of Interest". I entered the derivatives/conn/sub-xxx/dssub-xxx_bold.nii file from CONN to SPM as the SCAN files, and here are the parameters I used:

For fMRI Model specification:

• High-pass filter: Inf


• Global normalisation: None


• Masking threshold: 0


• Serial correlations: None

For Volume-of-Interest:

• Adjust data: 0 (adjust for nothing)

I also used the same label atlas file for both SPM and CONN.

Did I miss anything in my setup? I appreciate any help or ideas on how to get an exact time series result with SPM and CONN!