Functional connectivity seed analysis of right motor cortex

Hands on Practical for Resting-state Functional Connectivity

For this hands on lab practice, we will be using data from sub-001 that you downloaded and processed for the previous block.

They should be located at ~/fmriLab/flankerData_n4/, and you should be able to navigate to it by doing cd ~/fmriLab/flankerData_n4.

You should also have the "skull-striped T1 image" saved under ~/fmriLab/flankerData_n4/sub-001/anat/

What you will learn from this lab practice

• How to replicate this famous finding from Biswal 1995

• Understand how to preprocess fMRI data for resting-state functional connectivity analysis
• Understand how to extract and input a "seed" timeseries to search for brain regions that show resting-state connectivity with this "seed" region of interest.

Step 1, preprocess resting-state fMRI data

We will use the steps you learned from last block to preprocess the resting-state data.

• Let's use FEAT again to set up preprocessing

  • Go to the subject folder and launch FSL

    • cd ~/fmriLab/flankerData_n4/sub-001/func
    • fsl

  • In FSL GUI, open 'FEAT FMRI analysis'

    • Define scope of 'First-level analysis' to 'Preprocessing' at the top of the GUI
    • Select sub-001-task-rest_bold.nii.gz as input 4D.
    • Set output to ~/fmriLab/flankerData_n4/sub-001/func/rest_rmot.feat
    • We will delete the first 4 volumes in Delete volumes
    • Leave the High pass filter cutoff (s) to 100s

  • Go to the Pre-stats tab. We will do the following preprocessing

    • Motion correction with MCFLIRT
    • Select BET
    • Spatial smoothing of 6 mm
    • select Highpass

  • Go to the Registration tab, do the following:

    • Select Main Structural Image
    • Select the "brain extracted" T1 as the main structural image
    • Change to normal search and 12 DOF
    • For standard space, change to normal search and 12 DOF

• Hit Go!

• We then need to manually do "low-pass" filter. That is because FEAT does not support it (only does highpass). We have to do it via command line in the terminal.

  • Now open the terminal, move to the rest_rmot.feat folder you just made: cd ~/fmriLab/flankerData_n4/sub-001/func/rest_rmot.feat
  • Note, this might not work if you didn't save rest_rmot.feat under sub-001, in that case you have to find our where you saved it.
  • Then run this command: fslmaths filtered_func_data.nii.gz -bptf -1 2.5 filtered_func_data.nii.gz

• Here we are doing lowpass filtering of 0.08 hz.

  • the -bptf option expects a high-pass sigma and a low-pass sigma, which can be caluclated by
  • highpass_sigma = 1 / (2.35 * TR * HP_freq) (we use "-1" because we already highpassed the data)
  • lowpass_sigma = 1 / (2.35 * TR * LP_freq)) (remember TR is 2s)
  • If interested in an explanation for this, see here: https://www.jiscmail.ac.uk/cgi-bin/webadmin?A2=fsl;fc5b33c5.1205

Step 2. Locate a seed region of interest in fsleyes

• Stay in the preprocessed output folder ~/fmriLab/flankerData_n4/sub-001/func/rest_rmot.feat/

• Locate the right motor cortex

  • Use fsleyes to open up the preprocessed structural image
    • file, add from file, goto reg, select highres
  • Also add the filtered_func_data
  • In the Overlay list panel, select filtered_func_data image
  • To navigate to the right motor cortex, enter the following x y z coordinate for scanner anatomical space:
    x=34 (top row), y=-12 (middle row), z=24 (lower row)

Step 3. Create a functional connectivity seed map using fsleyes and save it as a nifti file.

• With your cursor on your seed mask, go to View -> Timeseries
• With your cursor still on your seed mask, go to Tools -> Seed correlation (Pearson) - what do you see?!
• Notice we now have a new image in our Overlay list, and we can explore it's range and distribution to understand it more
• Based on the distribution, what would be a reasonable threshold to try to replicate the seed map at the start of the tutorial?
• Save your seedmap in an image by clicking on the disk to the left of the image name in your Overlay list. Your saved image will end in .nii.gz which is a "nifti file" that you can open again in fsl later.

Step 4. Threshold your functional connectivity seed map

• Below is an example of steps in fsleyes for thresholding your seedmap to keep only the voxels with the strongest correlation to your seed.

Technical Assignment. Run the same seed analysis for the "task state" by using the flanker data

• Replicate all the steps above with sub-001 but use their flanker data as input (sub-001_task-flanker_bold.nii.gz)

• I suggest you name your .feat folder something distinct, like flanker_rmot.feat

• Create a seed map for the right motor cortex, that you threshold at r>.70 like above, and display your timeseries in fsleyes

• Overlay your seedmap from resting state, threshold at r>.70, and display the map in a different color

• Are they more similar or different than you expected and why?

• With your task timeseries and thresholded seedmaps from rest and task in view, save your work by taking a screenshot with these steps:

  • Applications -> Accessories -> Scroll down to Screenshot
  • Select a region -> use your cursor to select your fsleyes display
  • Save your map as: hawkid_rmot_seedmaps.png
  • Upload this .png file to ICON to turn in your technical assignment