6.3.2. ICA on mouse

Independent components analysis on 5 mice.

6.3.2.1. Retrieve the fMRI data

from sammba import data_fetchers

retest = data_fetchers.fetch_zurich_test_retest(subjects=range(5),
                                                correct_headers=True)

6.3.2.2. Load the template and its brain mask

dorr = data_fetchers.fetch_atlas_dorr_2008(downsample='100')
dorr_masks = data_fetchers.fetch_masks_dorr_2008(downsample='100')
print('Path to template is {} and to the brain mask is {}'.format(dorr.t2,
      dorr_masks.brain))

Out:

Path to template is /home/salma/nilearn_data/dorr_2008/Dorr_2008_average_100um.nii.gz and to the brain mask is /home/salma/nilearn_data/dorr_2008/dorr_2008_brain_mask_100.nii.gz

6.3.2.3. Register to the template

import os
from sammba.registration import TemplateRegistrator

registrator = TemplateRegistrator(brain_volume=400, caching=True,
                                  template=dorr.t2, use_rats_tool=False,
                                  template_brain_mask=dorr_masks.brain,
                                  registration_kind='affine')

registered_funcs = []
for anat, func in zip(retest.anat, retest.func):
    animal_id = os.path.basename(os.path.dirname(anat))
    registrator.output_dir = os.path.join('ica', animal_id)
    registrator.fit_anat(anat)
    registrator.fit_modality(func, 'func', t_r=1., voxel_size=(.3, .3, .3),
                             prior_rigid_body_registration=True)
    registered_funcs.append(registrator.registered_func_)

6.3.2.4. Run ICA

from nilearn.decomposition import CanICA

canica = CanICA(n_components=30, smoothing_fwhm=.3, n_jobs=-1)
canica.fit(registered_funcs)

Retrieve the independent components in brain space.

components_img = canica.masker_.inverse_transform(canica.components_)

6.3.2.5. Visualize the components

We can plot the outline of all components on one figure.

from nilearn import plotting

plotting.plot_prob_atlas(components_img,
                         bg_img=registrator.template_brain_,
                         display_mode='z',
                         title='All ICA components')

Total running time of the script: ( 8 minutes 31.678 seconds)