Extracting colorbar from `nilearn.plotting.plot_stat_map` figure

shawnrhoads · January 5, 2023, 4:52pm

Hi all,

Does anyone know of a way to extract the colorbar from a display output object generated using the nilearn.plotting.plot_stat_map() function? Or even just recreate the same colorbar from a given stat image?

Thanks!
Shawn

pbellec · January 5, 2023, 10:02pm

This snippet is used to save the colormap in an image file:

github.com

nilearn/nilearn/blob/main/nilearn/plotting/html_stat_map.py#L173-L180

      
        
            def _save_cm(output_cmap, cmap, format='png', n_colors=256):
                """ Save the colormap of an image as an image file.
                """
            
            
    # save the colormap
                data = np.arange(0., n_colors) / (n_colors - 1.)
                data = data.reshape([1, n_colors])
                imsave(output_cmap, data, cmap=cmap, format=format)

It will reproduce exactly what’s shown with nilearn.plotting.view_img and I suspect it should also work for nilearn.plotting.plot_stat_map although you may want to test it.

shawnrhoads · January 5, 2023, 11:36pm

Amazing, thanks!

This pointed me in the right direction. I ended up using the functions in nilearn.plotting.html_stat_map to recreate the colorbar of stat map like this:

import nibabel as nib
from nilearn.plotting.html_stat_map import _mask_stat_map, colorscale
from nilearn.plotting import cm

im_file = 'stats.nii.gz'
nib_img = nib.load(im_file)

mask_img, stat_map_img, data, threshold = _mask_stat_map(nib_img, 1e-6)
        
colors = colorscale(cm.cold_hot, data.ravel(), threshold=threshold,
                        symmetric_cmap=True, vmax=None,vmin=None)

And now I can add the colorbar anywhere in a figure using:

from matplotlib import cm as mpl_cm
fig.colorbar(mpl_cm.ScalarMappable(norm=colors['norm'], cmap=colors['cmap']), **)

My workaround is probably not the most elegant approach, but my dilemma is solved for now!
(Also open to any other tips!)