Concatenate time-series across runs

Submitted by NinaGoossens on

Dear experts, 

We performed three separate fMRI runs, during which we applied muscle vibration at 60 Hz and at 20 Hz in blocks of 18s, which were separated by 18s blocks of rest (no vibration). Each block is preceded by a short verbal instruction (3s, 1 scan). By contrasting 60 Hz to 20 Hz vibration, we determine brain activation during proprioceptive processing.

I am interested in performing a task-related connectivity analysis in GraphVar, and have extracted time-series adjusted for the contrast of interest (60Hz>20Hz) in 20 user-defined ROIs with the VOI-tool in SPM12. However, I would like to know how I can concatenate the three runs, to ultimately acquire one 20x20 correlation matrix.

Can anyone suggest a way to concatenate the three runs?

Thank you in advance!


Nina Goossens




Johann Kruschwitz

Thu, 06/21/2018 - 13:30

I think the way to go would be to use the cPPI toolbox ( to properly extract the time signals (i.e., functional correlation matrix) for the respective contrast. As the cPPI refers to your first-level models it automatically deals with several runs. Once, you have done this you can simply use the cPPI output (i.e., a matrix representing functional connectivity differences of a contrast A>B) in GraphVar.
I hope this hepls,


alternatively you could also use the BASCO toolbox ( that will return you a matrix of coorealtions (beta series correlation) for a specific task condition. In your case you could extract a matrix for each of the conditions and contrast them in a within-subject design in GraphVar.


Dear Johann, 

Thank you for your response! I would like to use cPPI to extract the correlation matrix of my desired contrast, concatenated over the three runs. Could you recommend a toolbox that would be able to use this output to calculate graph measures? Can the correlation matrices be used in GraphVar?

Kind regards, 



Dear Johann, 

Thank you! 

However, I struggle a bit with which time courses to use as inputs. I already extracted time-series adjusted to my contrast of interest (60Hz - 20 Hz). Do I use these time courses as input or do I need to input not-adjusted, original time courses? And how can I extract these (using VOI tool?)?

Kind regards, 


    Nina Goossens

     PhD researcher  
     KU Leuven - Department of Rehabilitation Sciences
     Tervuursevest 101 box 1501 
     3001 Leuven
     tel. +32 16 37 90 40


Van: RFMRI.ORG <> namens The R-fMRI Network <>
Verzonden: maandag 16 juli 2018 14:33
Onderwerp: Re: [RFMRI] Concatenate time-series across runs