PSYC 410 fMRI Project

You are all no longer Padawan learners, but have become young fMRI Jedis (that's right, I'm using Star Wars references. I apologize for nothing!!). In this last project you will show off your newly acquired mastery of the brain imaging force (okay okay, I'll stop now) and perform a complete fMRI analysis.

This assignment will be very similar to the lab reports you've done throughout the semester. The primary difference is that I will not furnish you with a step-by-step guide for carrying out the analyses.

For the fMRI analysis you will act as if you're newly minted fMRI researchers, not students. However, there is no scientist who knows and remembers everything and can work in total isolation without the aid of colleagues, references, and the internet. So you are allowed to use any resource you'd like, including our lab wikis, to assist you in your analyses. This also means you are permitted to ask each other for assistance. However, you are not permitted to work together. That is, asking for a little help here and there when you get really stuck with something is fine, but I expect you to primarily work independently.

• fMRI Report: Turn in your report via Moodle
• fMRI Analysis: see below
  • Create a Google Drive folder and share it with me.
    • Name the folder yourlastname_410_fmri
    • Create three sub-folders:
      • first-level
      • second-level
      • third-level
  • First-level analysis: Put your design file (.fsf file) for each run in the first-level folder.
  • Second-level analysis: Put your design file (.fsf file) in the second-level folder.
  • Third-level analysis: Put your design file (.fsf file) and a zipped version of the cope3.gfeat directory in the third-level folder

Throughout the wiki, replace:

• SUBJ with the ID of your assigned subject
• TASK with the task name (motion)
• RUN with whatever run you are analyzing (e.g., run02)

- Create a subject directory on your computer.

mkdir -p ~/Desktop/yourLastName_fmri_proj/SUBJ

- Create a third-level output directory on your computer.

mkdir -p ~/Desktop/yourLastName_fmri_proj/third_level

It has long been known that specific parts of the visual system are particularly sensitive to the perception of motion. More recently it has been discovered that the brain also has regions that are specifically sensitive to biological motion (see example below). This is analogous to face perception in that we know the visual system has some areas that prefer objects to non-objects, but that face “objects” have their own dedicated region.

Point-light display movies are used to identify these biological motion sensitive areas. The movies are simply lights on a black background. Nonetheless, when these lights are placed at particular locations on a body in motion, we easily perceive it as a body and not just randomly moving lights.

So why use point-light movies instead of just a video of a person walking? Good question. The reason point-light movies are great is that it is very easy to create an excellent control condition. The only stimulation reaching your eyes are discs of light in motion. To create a great control condition we simply need to scramble around the location of those discs (see example below). Doing so totally disrupts the perception of biological motion, but preserves all other facets of the display (total luminance, number of items, speed and direction of the individual lights, etc.)

Check out examples of what these movies look like:

• Right click on each of the movies below and select Loop (Use Google Chrome)
• Play both movies
• The movie on the left shows biological motion. It is clearly a man walking from right to left.
• The movie on the right shows non-biological motion. It has all the same elements of the 'walker' video, but these elements are scrambled, thus disrupting the perception of biological motion.

Biological motion is known to activate a region of the lateral temporal cortex (particularly in the right hemisphere) that is very near two related areas; the posterior superior temporal sulcus, which activates to faces, and area MT/V5, which is the motion sensitive area of the visual system. This paper is the first to show this using neuroimaging (this was in 1996, so they used Positron Emission Tomography, not fMRI).

In this image (Engell and McCarthy, 2013, NeuroImage), you can see regions of the brain that have the highest probability of being activated by biological motion more strongly than by non-biological motion. You can see that this set of areas is very similar to the one activated preferentially by faces. Activation in the right hemisphere includes the fusiform gyrus, the posterior superior temporal sulcus, area MT/V5 (which is selective for motion in general), and the “extrastriate body area” (which is preferentially activated by the perception of body parts).

• Seventeen participants
• All right-handed
• 10 female, 7 male
• Mean age = 20.4 years

Subjects in the this study were instructed to view short movies depicting point-light displays of biological or non-biological (random) motion. Between blocks there were rest periods in which the subjects viewed a small fixation cross on a grey background. Occasionally a “movie” would appear in which nothing was moving. Subjects were instructed to count how many of these static movies appeared and report this number at the end of each run. This task is not relevant for our investigation. Rather, it was used solely to ensure the participant continued to pay attention to the movies.

You are assigned a single subject (see below), who has three runs of data. Run 1st level analyses on all three runs of data.

Data (anatomical and functional) for your assigned subject is located in /Users/hnl/Desktop/class/input/fmri/loc/data/nifti

• SUBJ_highres.nii.gz
• SUBJ_coplanar.nii.gz
• SUBJ_motion_run1.nii.gz
• SUBJ_motion_run2.nii.gz
• SUBJ_motion_run3.nii.gz

You will carry out first-, second-, and third-level analyses on data from this experiment.

• The first- and second-level analyses will be performed on a single subject (see your assigned subject above).
• The third-level analysis will be performed across seventeen subjects.

Data should be:

• temporally highpass-filtered with a 100 second cutoff.
• motion corrected.
• corrected for slice scan-time.
• spatially smoothed with a 5 mm kernel.
• registered to standard space.
• The hemodynamic response should be modeled with a Gamma-variate convolution model
  • Phase=0
  • Stddev=3
  • Mean lag=6
• Include contrasts for:
  • the main effect of biomotion
  • the main effect of non-biomotion
  • biomoition > non-biomotion
  • non-biomotion > biomotion
  • Do not include temporal derivatives in your model.

Perform a fixed-effects second-level analysis for your assigned participant. Results should be cluster corrected with a voxel z threshold of 2.3 and a cluster threshold of p < .05.

In addition to your assigned subject, there are sixteen additional subjects. You will run a third-level analysis across all seventeen subjects. However, as with your previous lab, you do not have to do the first or second level analysis first as it has been done for you.

Run a mixed-effects (FLAME1) third-level analysis. Results should be cluster corrected with a voxel z threshold of 3.1 and a cluster threshold of p < .05.

Data are located in ~/Desktop/class/input/fsl/loc/analysis/fsl in the following subdirectories:

2545 2552 2553 2554 2585 2731 2766 2767 2814 3850 3851 3855 3866 5743 5744 5769 5770