hello ,

i have 100 nifti image and i have JHU atlas which contains 48 regions.

i compute the average of signals for each region to have X of shape (n_samples,48)

i train a regression model to get a weight of shape (1,48) .

i want to return to voxel space

how can i do this ?

@bthirion thank you very much

Use masker.invers_transform() to transform the 48-dimensional vector to a brain map.

Best,

yes ! thank you very much