I need your HELP

[JunXingzao-xht]

Have you ever tried projecting onto the fsaverage surface? I'm trying to do this, but I noticed that the wm.mgz file is missing from my /fsaverage/mri folder.

[AbabababaSmart]

Hi, I'm sorry I haven't tried projecting data directly onto the fsaverage surface. But to answer your question, I did try it with random data, and it is indeed possible to visualize a flatmap using the files provided in the db/fsaverage folder, without needing something like wm.mgz in the fsaverage/mri files.

Here's an example:

import cortex
import numpy as np
np.random.seed(1234)
import matplotlib.pyplot as plt

subject = 'fsaverage'
xfm = 'atlas'  # The available xfm options are ["atlas", "atlas_2mm", "atlas336", "atlas_3mm"] (provided in this repo)

# Creating a random dataset that matches the shape for this transform with one
# entry for each voxel
test_data = np.random.randn(182, 218, 182)

# This creates a Volume object for our test dataset for the given subject
# and transform
vol_data = cortex.Volume(test_data, subject, xfm, vmin=-2, vmax=2)
cortex.quickshow(vol_data)
plt.show()

However, there are shape requirements for the data when performing the projection. For example, if you choose the xfm as atlas, the projection data needs to have a shape of (182, 218, 182). I believe this is closely related to the resolution of the template, and you might refer to the following link for more information on resolution conversion: Upsample data from a lower resolution fsaverage template to fsaverage for visualization.
.
On the other hand, I believe using a standard template might introduce some distortion in the final flatmap visualization. As mentioned on the pycortex official website:

In order to plot data, you will need to create surfaces for your particular subject. General surfaces such as fsaverage are NOT recommended, since individual subject anatomy can be highly variable. Averaging across subjects will destroy your signal specificity! (Segmentation Tutorial)

So perhaps using surfaces particular to each subject might be a better choice.