Home

Gist/Flow of the Code

This is the high level flow of the code.
Please refer to the docstrings in functions and the sections of the notebook for better understanding.

1. Load data from Benson et. al. PRF analysis, which contains
   1. PRF analysis results
   2. Related atlases
2. Create Pandas DataFrame from the 4D Numpy Array -> df
   1. Add additional columns for parcellation and hemisphere using an atlas
3. Download and read GIFTI files (coordinate and mask/atlas) from HCP to get coordinates
   1. using the get_coors function (use appropriate surface and volume spaces)
   2. Add columns to the DataFrame for coordinates (x, y, and z)
4. Filter the DataFrame to remove points
   1. with less confidence (based on R2) -> df_
   2. not of interest (based on ROI array) -> dfv
5. Create separate DataFrames for either hemispheres -> dfv_l and dfv_r
6. Group DataFrame by subject ID
   1. For each group, use group function to further
      1. group by
         1. parcellation
         2. bins of eccentricity and angular coordinates
      2. aggregate for each combination of parcellation, eccentricity, and angular bin based on provided aggregation function (mean)
   2. Use l2 function to find the difference between two parcellations for each bin and calculate the L2 distance
   3. Create a pivot table of eccentricity and angular bins with distances as values -> dfv_d
7. Plot the final mean pivot table across participants