Dear experts,
We performed 3 block-design fMRI runs with during each run
- 3 blocks of 18s with muscle vibration at 60 Hz (Cond60)
- 3 blocks of 18s with muscle vibration at 20 Hz (Cond20)
- 6 blocks of 18s rest (no vibration) (Rest)
- 12 shorter blocks of 3s (auditory instruction) (Instr)
and want to perform graph theory to investigate functional connectivity within a proprioceptive network (ROIs based on MNI coordinates from previous study showing significant activation during proprioceptive stimulation (i.e. contrast of 60 Hz - 20 Hz))
I want to use CONN to build connectivity matrices and calculate graph theory measures. We are specifically interested in the contrast of Cond60 - Cond20.
During set-up, I now defined onsets and durations of Cond60, Cond20 and Instr in each run (Rest modeled implicitly). Do I have to include Rest as a condition? I thought that, since I am only interested in Cond60 - Cond20, this would not be necessary. However, we might want to look at the contrast of Cond60 - Rest in the future. What are the consequences of adding Rest when looking at Cond60 - Cond20?
During denoising:
- previous posts learned that I should adjust bandpass filter to [0.008 inf]. However, do I need to included "Effects of Cond60", "Effects of Cond20" , "Effects of Instr" (and possibly "Effects of Rest" depending on question above) to the confounds list?
- do I need to add despiking, or only keep the default of detrending linear?
Thank you very much for your help!
Nina
We performed 3 block-design fMRI runs with during each run
- 3 blocks of 18s with muscle vibration at 60 Hz (Cond60)
- 3 blocks of 18s with muscle vibration at 20 Hz (Cond20)
- 6 blocks of 18s rest (no vibration) (Rest)
- 12 shorter blocks of 3s (auditory instruction) (Instr)
and want to perform graph theory to investigate functional connectivity within a proprioceptive network (ROIs based on MNI coordinates from previous study showing significant activation during proprioceptive stimulation (i.e. contrast of 60 Hz - 20 Hz))
I want to use CONN to build connectivity matrices and calculate graph theory measures. We are specifically interested in the contrast of Cond60 - Cond20.
During set-up, I now defined onsets and durations of Cond60, Cond20 and Instr in each run (Rest modeled implicitly). Do I have to include Rest as a condition? I thought that, since I am only interested in Cond60 - Cond20, this would not be necessary. However, we might want to look at the contrast of Cond60 - Rest in the future. What are the consequences of adding Rest when looking at Cond60 - Cond20?
During denoising:
- previous posts learned that I should adjust bandpass filter to [0.008 inf]. However, do I need to included "Effects of Cond60", "Effects of Cond20" , "Effects of Instr" (and possibly "Effects of Rest" depending on question above) to the confounds list?
- do I need to add despiking, or only keep the default of detrending linear?
Thank you very much for your help!
Nina