def test_anisotropic_sphere_extraction():
data = np.random.RandomState(42).random_sample((3, 3, 3, 5))
affine = np.eye(4)
affine[0, 0] = 2
affine[2, 2] = 2
img = nibabel.Nifti1Image(data, affine)
masker = NiftiSpheresMasker([(2, 1, 2)], radius=1)
# Test the fit
masker.fit()
# Test the transform
s = masker.transform(img)
mask = np.zeros((3, 3, 3), dtype=np.bool)
mask[1, :, 1] = True
assert_array_equal(s[:, 0], np.mean(data[mask], axis=0))
# Now with a mask
mask_img = np.zeros((3, 2, 3))
mask_img[1, 0, 1] = 1
affine_2 = affine.copy()
affine_2[0, 0] = 4
mask_img = nibabel.Nifti1Image(mask_img, affine=affine_2)
masker = NiftiSpheresMasker([(2, 1, 2)], radius=1, mask_img=mask_img)
masker.fit()
s = masker.transform(img)
assert_array_equal(s[:, 0], data[1, 0, 1])
def test_small_radius():
affine = np.eye(4)
shape = (3, 3, 3)
data = np.random.random(shape)
mask = np.zeros(shape)
mask[1, 1, 1] = 1
mask[2, 2, 2] = 1
affine = np.eye(4) * 1.2
seed = (1.4, 1.4, 1.4)
masker = NiftiSpheresMasker([seed],
radius=0.1,
mask_img=nibabel.Nifti1Image(mask, affine))
masker.fit_transform(nibabel.Nifti1Image(data, affine))
# Test if masking is taken into account
mask[1, 1, 1] = 0
mask[1, 1, 0] = 1
masker = NiftiSpheresMasker([seed],
radius=0.1,
mask_img=nibabel.Nifti1Image(mask, affine))
assert_raises_regex(ValueError,
'Sphere around seed #0 is empty', masker.fit_transform,
nibabel.Nifti1Image(data, affine))
masker = NiftiSpheresMasker([seed],
radius=1.6,
mask_img=nibabel.Nifti1Image(mask, affine))
masker.fit_transform(nibabel.Nifti1Image(data, affine))
def test_gaussian_coord_smoothing():
coords = [(0.0, 0.0, 0.0), (10.0, -10.0, 30.0)]
computed_img = img_utils.gaussian_coord_smoothing(coords)
masker = NiftiSpheresMasker(coords + [(-10.0, 10.0, -30)]).fit()
values = masker.transform(computed_img)[0]
assert (values[:2] > computed_img.get_data().max() / 2.0).all()
assert values[-1] == pytest.approx(0.0)
def test_seed_extraction():
data = np.random.random((3, 3, 3, 5))
img = nibabel.Nifti1Image(data, np.eye(4))
masker = NiftiSpheresMasker([(1, 1, 1)])
# Test the fit
masker.fit()
# Test the transform
s = masker.transform(img)
assert_array_equal(s[:, 0], data[1, 1, 1])
def test_seed_extraction():
data = np.random.random((3, 3, 3, 5))
img = nibabel.Nifti1Image(data, np.eye(4))
masker = NiftiSpheresMasker([(1, 1, 1)])
# Test the fit
masker.fit()
# Test the transform
s = masker.transform(img)
assert_array_equal(s[:, 0], data[1, 1, 1])
def extract_timeseries_coords(filename, raw_coords, confounds=None):
"""Because the power parcellation is given in coordinates and not labels,
we dedicate an exclusive function to deal with it.
"""
coords = np.vstack(
(raw_coords.rois['x'], raw_coords.rois['y'], raw_coords.rois['z'])).T
spheres_masker = NiftiSpheresMasker(seeds=coords,
radius=5.,
standardize=True)
time_series = spheres_masker.fit_transform(filename, confounds=confounds)
return time_series
def signal_extract(data,atlas,t_r=2.2,masker_type='Spheres',saveas='file'):
"""
Extracts BOLD time-series from regions of interest
Parameters
----------
data: Filenames of subjects.
atlas: regions or coordinates to extract signals from.
masker_type : Type of masker used to extract BOLD signals . types are : 'Spheres','Maps','Labels'
saveas : Destination to save and load output (.npz)
Returns
---------
subject_ts : array-like , 2-D (n_subjects,n_regions)
Array of BOLD time-series
"""
subjects_ts=[]
if os.path.exists(saveas):
subjects_ts=np.load(saveas)['arr_0']
else:
if masker_type== 'Spheres':
masker = NiftiSpheresMasker(
seeds=atlas, smoothing_fwhm=6, radius=4 ,mask_img=brainmask,
detrend=False, standardize=True, low_pass=0.1, high_pass=0.01, t_r=t_r)
elif masker_type == 'Maps':
masker = NiftiMapsMasker(maps_img=atlas,mask_img=brainmask,standardize=True,
high_pass=0.01,low_pass=0.1,detrend=False,t_r=t_r,
memory_level=2,smoothing_fwhm=5,resampling_target='data',
memory=mem,verbose=5)
elif masker_type == 'Labels':
masker = NiftiLabelsMasker(labels_img=atlas,mask_img=brainmask,standardize=True,
high_pass=0.01,low_pass=0.1,detrend=False,t_r=t_r,
memory_level=2,smoothing_fwhm=5,resampling_target='data',
memory=mem,verbose=5)
else:
raise ValueError("Please provide masker type")
for func_file in data:
time_series = masker.fit_transform(func_file)
subjects_ts.append(time_series)
np.savez(saveas,subjects_ts)
return subjects_ts
def fit(self):
# Create mask
print("Fit the SphereMasker...")
self.n_seeds = int(self.seeds_img.get_data().sum())
# Pass our xform_fn for a callback on each seed.
self.sphere_masker = NiftiSpheresMasker(
seeds=self.seeds_img,
mask_img=self.seeds_img,
radius=self.radius,
xform_fn=self.rsa_on_ball_axis_1,
standardize=False) # no mem
self.sphere_masker.fit()
def _fmri_roi_extract_image(data, atlas_path, atlas_type, radius, overlap_ok,mask = None):
if 'label' in atlas_type:
logging.debug('Labels Extract')
label_masker = NiftiLabelsMasker(atlas_path, mask_img=mask)
timeseries = label_masker.fit_transform(data)
if 'sphere' in atlas_type:
atlas_path = np.loadtxt(atlas_path)
logging.debug('Sphere Extract')
spheres_masker = NiftiSpheresMasker(atlas_path, float(radius),mask_img=mask, allow_overlap = overlap_ok)
timeseries = spheres_masker.fit_transform(data)
if 'maps' in atlas_type:
logging.debug('Maps Extract')
maps_masker = NiftiMapsMasker(atlas_path,mask_img=mask, allow_overlap = overlap_ok)
timeseries = maps_masker.fit_transform(data)
timeseries[timeseries == 0.0] = np.nan
return timeseries
def test_small_radius():
affine = np.eye(4)
shape = (3, 3, 3)
data = np.random.random(shape)
mask = np.zeros(shape)
mask[1, 1, 1] = 1
mask[2, 2, 2] = 1
affine = np.eye(4) * 1.2
seed = (1.4, 1.4, 1.4)
masker = NiftiSpheresMasker([seed], radius=0.1,
mask_img=nibabel.Nifti1Image(mask, affine))
masker.fit_transform(nibabel.Nifti1Image(data, affine))
# Test if masking is taken into account
mask[1, 1, 1] = 0
mask[1, 1, 0] = 1
masker = NiftiSpheresMasker([seed], radius=0.1,
mask_img=nibabel.Nifti1Image(mask, affine))
assert_raises_regex(ValueError, 'Sphere around seed #0 is empty',
masker.fit_transform,
nibabel.Nifti1Image(data, affine))
masker = NiftiSpheresMasker([seed], radius=1.6,
mask_img=nibabel.Nifti1Image(mask, affine))
masker.fit_transform(nibabel.Nifti1Image(data, affine))
def extract_parcellation_time_series(in_data, parcellation_name,
parcellations_dict, bp_freqs, tr):
'''
Depending on parcellation['is_probabilistic'] this function chooses either NiftiLabelsMasker or NiftiMapsMasker
to extract the time series of each parcel
if bp_freq: data is band passfiltered at (hp, lp), if (None,None): no filter, if (None, .1) only lp...
tr in ms (e.g. from freesurfer ImageInfo())
returns np.array with parcellation time series and saves this array also to parcellation_time_series_file, and
path to pickled masker object
'''
from nilearn.input_data import NiftiLabelsMasker, NiftiMapsMasker, NiftiSpheresMasker
import os, pickle
import numpy as np
if parcellations_dict[parcellation_name][
'is_probabilistic'] == True: # use probab. nilearn
masker = NiftiMapsMasker(
maps_img=parcellations_dict[parcellation_name]['nii_path'],
standardize=True)
elif parcellations_dict[parcellation_name]['is_probabilistic'] == 'sphere':
atlas = pickle.load(
open(parcellations_dict[parcellation_name]['nii_path']))
coords = atlas.rois
masker = NiftiSpheresMasker(coords,
radius=5,
allow_overlap=True,
standardize=True)
else: # 0/1 labels
masker = NiftiLabelsMasker(
labels_img=parcellations_dict[parcellation_name]['nii_path'],
standardize=True)
# add bandpass filter (only executes if freq not None
hp, lp = bp_freqs
masker.low_pass = lp
masker.high_pass = hp
if tr is not None:
masker.t_r = tr
else:
masker.t_r = None
masker.standardize = True
masker_file = os.path.join(os.getcwd(), 'masker.pkl')
with open(masker_file, 'w') as f:
pickle.dump(masker, f)
parcellation_time_series = masker.fit_transform(in_data)
parcellation_time_series_file = os.path.join(
os.getcwd(), 'parcellation_time_series.npy')
np.save(parcellation_time_series_file, parcellation_time_series)
return parcellation_time_series, parcellation_time_series_file, masker_file
def build_series(brain4d):
coords = [
(30, -45, -10),
(0, -85, 10),
(0, 47, 30),
(-35, -20, 50),
]
masker = NiftiSpheresMasker(
coords,
radius=5,
detrend=True,
standardize=True,
low_pass=0.1,
high_pass=0.01,
t_r=2,
)
series = masker.fit_transform(brain4d)
return series
def report_flm_adhd_dmn(): # pragma: no cover
t_r = 2.
slice_time_ref = 0.
n_scans = 176
pcc_coords = (0, -53, 26)
adhd_dataset = nilearn.datasets.fetch_adhd(n_subjects=1)
seed_masker = NiftiSpheresMasker([pcc_coords],
radius=10,
detrend=True,
standardize=True,
low_pass=0.1,
high_pass=0.01,
t_r=2.,
memory='nilearn_cache',
memory_level=1,
verbose=0)
seed_time_series = seed_masker.fit_transform(adhd_dataset.func[0])
frametimes = np.linspace(0, (n_scans - 1) * t_r, n_scans)
design_matrix = make_first_level_design_matrix(frametimes,
hrf_model='spm',
add_regs=seed_time_series,
add_reg_names=["pcc_seed"])
dmn_contrast = np.array([1] + [0] * (design_matrix.shape[1] - 1))
contrasts = {'seed_based_glm': dmn_contrast}
first_level_model = FirstLevelModel(t_r=t_r, slice_time_ref=slice_time_ref)
first_level_model = first_level_model.fit(run_imgs=adhd_dataset.func[0],
design_matrices=design_matrix)
report = make_glm_report(
first_level_model,
contrasts=contrasts,
title='ADHD DMN Report',
cluster_threshold=15,
height_control='bonferroni',
min_distance=8.,
plot_type='glass',
report_dims=(1200, 'a'),
)
output_filename = 'generated_report_flm_adhd_dmn.html'
output_filepath = os.path.join(REPORTS_DIR, output_filename)
report.save_as_html(output_filepath)
report.get_iframe()
def _set_volume_masker(roi_file, as_voxels=False, **kwargs):
"""Check and see if multiple ROIs exist in atlas file"""
if not isinstance(roi_file, str):
raise ValueError('roi_file must be a file name string')
if roi_file.endswith('.csv') or roi_file.endswith('.tsv'):
roi = _read_coords(roi_file)
n_rois = len(roi)
print(' {} region(s) detected from coordinates'.format(n_rois))
if kwargs.get('radius') is None:
warnings.warn('No radius specified for coordinates; setting '
'to nilearn.input_data.NiftiSphereMasker default '
'of extracting from a single voxel')
masker = NiftiSpheresMasker(roi, **kwargs)
elif roi_file.endswith('.nii.gz'):
# remove args for NiftiSpheresMasker
if 'radius' in kwargs:
kwargs.pop('radius')
if 'allow_overlap' in kwargs:
kwargs.pop('allow_overlap')
roi_img = image.load_img(roi_file)
if len(roi_img.shape) == 4:
n_rois = roi_img.shape[-1]
print(' {} region(s) detected from {}'.format(n_rois,
roi_img.get_filename()))
masker = NiftiMapsMasker(roi_img, allow_overlap=True,**kwargs)
else:
n_rois = len(np.unique(roi_img.get_fdata())) - 1
print(' {} region(s) detected from {}'.format(n_rois,
roi_img.get_filename()))
if n_rois > 1:
masker = NiftiLabelsMasker(roi_img, **kwargs)
elif n_rois == 1:
# binary mask for single ROI
if as_voxels:
if 'mask_img' in kwargs:
kwargs.pop('mask_img')
masker = NiftiMasker(roi_img, **kwargs)
else:
# more computationally efficient if only wanting the mean
masker = NiftiLabelsMasker(roi_img, **kwargs)
else:
raise ValueError('No ROI detected; check ROI file')
else:
raise ValueError('Invalid file type for roi_file. Must be one of: '
'.nii.gz, .csv, .tsv')
return masker, n_rois
def test_is_nifti_spheres_masker_give_nans():
affine = np.eye(4)
data_with_nans = np.zeros((10, 10, 10), dtype=np.float32)
data_with_nans[:, :, :] = np.nan
data_without_nans = np.random.random((9, 9, 9))
indices = np.nonzero(data_without_nans)
# Leaving nans outside of some data
data_with_nans[indices] = data_without_nans[indices]
img = nibabel.Nifti1Image(data_with_nans, affine)
seed = [(7, 7, 7)]
# Interaction of seed with nans
masker = NiftiSpheresMasker(seeds=seed, radius=2.)
assert_false(np.isnan(np.sum(masker.fit_transform(img))))
mask = np.ones((9, 9, 9))
mask_img = nibabel.Nifti1Image(mask, affine)
# When mask_img is provided, the seed interacts within the brain, so no nan
masker = NiftiSpheresMasker(seeds=seed, radius=2., mask_img=mask_img)
assert_false(np.isnan(np.sum(masker.fit_transform(img))))
def test_sphere_extraction():
data = np.random.random((3, 3, 3, 5))
img = nibabel.Nifti1Image(data, np.eye(4))
masker = NiftiSpheresMasker([(1, 1, 1)], radius=1)
# Test the fit
masker.fit()
# Test the transform
s = masker.transform(img)
mask = np.zeros((3, 3, 3), dtype=np.bool)
mask[:, 1, 1] = True
mask[1, :, 1] = True
mask[1, 1, :] = True
assert_array_equal(s[:, 0], np.mean(data[mask], axis=0))
# Now with a mask
mask_img = np.zeros((3, 3, 3))
mask_img[1, :, :] = 1
mask_img = nibabel.Nifti1Image(mask_img, np.eye(4))
masker = NiftiSpheresMasker([(1, 1, 1)], radius=1, mask_img=mask_img)
masker.fit()
s = masker.transform(img)
assert_array_equal(
s[:, 0],
np.mean(data[np.logical_and(mask, mask_img.get_data())], axis=0))
def test_nifti_spheres_masker_overlap():
# Test resampling in NiftiMapsMasker
affine = np.eye(4)
shape = (5, 5, 5)
data = np.random.random(shape + (5,))
fmri_img = nibabel.Nifti1Image(data, affine)
seeds = [(0, 0, 0), (2, 2, 2)]
overlapping_masker = NiftiSpheresMasker(seeds, radius=1,
allow_overlap=True)
overlapping_masker.fit_transform(fmri_img)
overlapping_masker = NiftiSpheresMasker(seeds, radius=2,
allow_overlap=True)
overlapping_masker.fit_transform(fmri_img)
noverlapping_masker = NiftiSpheresMasker(seeds, radius=1,
allow_overlap=False)
noverlapping_masker.fit_transform(fmri_img)
noverlapping_masker = NiftiSpheresMasker(seeds, radius=2,
allow_overlap=False)
assert_raises_regex(ValueError, 'Overlap detected',
noverlapping_masker.fit_transform, fmri_img)
def _set_masker(roi_file, as_voxels=False, **kwargs):
"""Check and see if multiple ROIs exist in atlas file"""
if isinstance(roi_file, str) and roi_file.endswith('.tsv'):
roi = _read_coords(roi_file)
n_rois = len(roi)
is_coords = True
print(' {} region(s) detected from coordinates'.format(n_rois))
else:
roi = load_img(roi_file)
n_rois = len(np.unique(roi.get_data())) - 1
is_coords = False
print(' {} region(s) detected from {}'.format(n_rois,
roi.get_filename()))
if is_coords:
if kwargs.get('radius') is None:
warnings.warn('No radius specified for coordinates; setting '
'to nilearn.input_data.NiftiSphereMasker default '
'of extracting from a single voxel')
masker = NiftiSpheresMasker(roi, **kwargs)
else:
if 'radius' in kwargs:
kwargs.pop('radius')
if 'allow_overlap' in kwargs:
kwargs.pop('allow_overlap')
if n_rois > 1:
masker = NiftiLabelsMasker(roi, **kwargs)
elif n_rois == 1:
# single binary ROI mask
if as_voxels:
if 'mask_img' in kwargs:
kwargs.pop('mask_img')
masker = NiftiMasker(roi, **kwargs)
else:
# more computationally efficient if only wanting the mean of ROI
masker = NiftiLabelsMasker(roi, **kwargs)
else:
raise ValueError('No ROI detected; check ROI file')
return masker
def test_sphere_extraction():
data = np.random.random((3, 3, 3, 5))
img = nibabel.Nifti1Image(data, np.eye(4))
masker = NiftiSpheresMasker([(1, 1, 1)], radius=1)
# Test the fit
masker.fit()
# Test the transform
s = masker.transform(img)
mask = np.zeros((3, 3, 3), dtype=np.bool)
mask[:, 1, 1] = True
mask[1, :, 1] = True
mask[1, 1, :] = True
assert_array_equal(s[:, 0], np.mean(data[mask], axis=0))
# Now with a mask
mask_img = np.zeros((3, 3, 3))
mask_img[1, :, :] = 1
mask_img = nibabel.Nifti1Image(mask_img, np.eye(4))
masker = NiftiSpheresMasker([(1, 1, 1)], radius=1, mask_img=mask_img)
masker.fit()
s = masker.transform(img)
assert_array_equal(s[:, 0],
np.mean(data[np.logical_and(mask, mask_img.get_data())],
axis=0))
def test_anisotropic_sphere_extraction():
data = np.random.random((3, 3, 3, 5))
affine = np.eye(4)
affine[0, 0] = 2
affine[2, 2] = 2
img = nibabel.Nifti1Image(data, affine)
masker = NiftiSpheresMasker([(2, 1, 2)], radius=1)
# Test the fit
masker.fit()
# Test the transform
s = masker.transform(img)
mask = np.zeros((3, 3, 3), dtype=np.bool)
mask[1, :, 1] = True
assert_array_equal(s[:, 0], np.mean(data[mask], axis=0))
# Now with a mask
mask_img = np.zeros((3, 2, 3))
mask_img[1, 0, 1] = 1
affine_2 = affine.copy()
affine_2[0, 0] = 4
mask_img = nibabel.Nifti1Image(mask_img, affine=affine_2)
masker = NiftiSpheresMasker([(2, 1, 2)], radius=1, mask_img=mask_img)
masker.fit()
s = masker.transform(img)
assert_array_equal(s[:, 0], data[1, 0, 1])
def test_standardization():
data = np.random.RandomState(42).random_sample((3, 3, 3, 5))
img = nibabel.Nifti1Image(data, np.eye(4))
# test zscore
masker = NiftiSpheresMasker([(1, 1, 1)], standardize='zscore')
# Test the fit
s = masker.fit_transform(img)
np.testing.assert_almost_equal(s.mean(), 0)
np.testing.assert_almost_equal(s.std(), 1)
# test psc
masker = NiftiSpheresMasker([(1, 1, 1)], standardize='psc')
# Test the fit
s = masker.fit_transform(img)
np.testing.assert_almost_equal(s.mean(), 0)
np.testing.assert_almost_equal(
s.ravel(),
data[1, 1, 1] / data[1, 1, 1].mean() * 100 - 100,
)
def test_is_nifti_spheres_masker_give_nans():
affine = np.eye(4)
data_with_nans = np.zeros((10, 10, 10), dtype=np.float32)
data_with_nans[:, :, :] = np.nan
data_without_nans = np.random.RandomState(42).random_sample((9, 9, 9))
indices = np.nonzero(data_without_nans)
# Leaving nans outside of some data
data_with_nans[indices] = data_without_nans[indices]
img = nibabel.Nifti1Image(data_with_nans, affine)
seed = [(7, 7, 7)]
# Interaction of seed with nans
masker = NiftiSpheresMasker(seeds=seed, radius=2.)
assert not np.isnan(np.sum(masker.fit_transform(img)))
mask = np.ones((9, 9, 9))
mask_img = nibabel.Nifti1Image(mask, affine)
# When mask_img is provided, the seed interacts within the brain, so no nan
masker = NiftiSpheresMasker(seeds=seed, radius=2., mask_img=mask_img)
assert not np.isnan(np.sum(masker.fit_transform(img)))
class RsaSearchlight(object):
def __init__(self,
mask_img,
seeds_img,
radius=10.,
distance_method='correlation',
memory_params=None):
# Defs
self.memory_params = memory_params or dict()
self.seeds_img = seeds_img
self.mask_img = mask_img
self.radius = radius
self.distance_method = distance_method
def rsa_on_ball_axis_1(self, sphere_data):
"""
Data: axis=1: [nvoxels, nslices]
"""
# sphere_data could be a single voxel; in this case, we'll get
# nan
similarity_comparisons = pdist(sphere_data.T, self.distance_method)
self.similarity_comparisons[self.si, :] = similarity_comparisons
self.n_voxels[self.si] = sphere_data.shape[0]
self.si += 1
if self.memory_params.get('verbose', 0) > 1 and self.si % 100 == 99:
print 'Processed %s of %s...' % (self.si + 1, self.n_seeds)
return similarity_comparisons.std() # output value for all slices
def fit(self):
# Create mask
print("Fit the SphereMasker...")
self.n_seeds = int(self.seeds_img.get_data().sum())
# Pass our xform_fn for a callback on each seed.
self.sphere_masker = NiftiSpheresMasker(
seeds=self.seeds_img,
mask_img=self.seeds_img,
radius=self.radius,
xform_fn=self.rsa_on_ball_axis_1,
standardize=False) # no mem
self.sphere_masker.fit()
def transform(self, func_img):
print("Transforming the image...")
n_images = func_img.shape[3]
n_compares = n_images * (n_images - 1) / 2
# These are computed within the callback.
self.si = 0
self.n_voxels = np.empty((self.n_seeds))
self.similarity_comparisons = np.empty((self.n_seeds, n_compares))
similarity_std = self.sphere_masker.transform(func_img)
# Pull the values off of self, set locally.
n_voxels = self.n_voxels
similarity_comparisons = self.similarity_comparisons
delattr(self, 'si')
delattr(self, 'n_voxels')
delattr(self, 'similarity_comparisons')
# Clean up
good_seeds = np.logical_not(
np.isnan(similarity_comparisons.mean(axis=1)))
n_voxels = n_voxels[good_seeds]
similarity_comparisons = similarity_comparisons[good_seeds]
similarity_std = similarity_std[:, good_seeds] # slices x seeds
return similarity_comparisons, similarity_std, n_voxels
def visualize(self,
similarity_comparisons,
similarity_std=None,
anat_img=None,
labels=None):
print("Plotting the results...")
self.visualize_seeds(anat_img=anat_img)
self.visualize_mask(anat_img=anat_img)
self.visualize_comparisons(
similarity_comparisons=similarity_comparisons,
labels=labels,
anat_img=anat_img)
self.visualize_comparisons_std(similarity_std=similarity_std,
anat_img=anat_img)
def visualize_seeds(self, anat_img=None):
plot_roi(self.sphere_masker.seeds_img_,
bg_img=anat_img,
title='seed img')
def visualize_mask(self, anat_img=None):
plot_roi(self.sphere_masker.mask_img_,
bg_img=anat_img,
title='mask img')
def visualize_comparisons(self,
similarity_comparisons,
labels=None,
anat_img=None):
# Plot (up to) twenty comparisons.
plotted_similarity = similarity_comparisons[:, 0]
plotted_img = self.sphere_masker.inverse_transform(
plotted_similarity.T)
plot_stat_map(plotted_img,
bg_img=anat_img,
title='RSA comparison %s vs. %s' % tuple(labels[:2]))
# Plot mosaic of up to 20
# Choose the comparisons
idx = np.linspace(0, similarity_comparisons.shape[1] - 1, 20)
idx = np.unique(np.round(idx).astype(int)) # if there's less than 20
# Make (and filter) titles
if labels is None:
titles = None
else:
titles = []
for ai, label1 in enumerate(labels):
for bi, label2 in enumerate(labels[(ai + 1):]):
titles.append('%s vs. %s' % (label1, label2))
titles = np.asarray(titles)[idx]
# Create the image
plotted_similarity = similarity_comparisons[:, idx]
plotted_img = self.sphere_masker.inverse_transform(
plotted_similarity.T)
fh = plt.figure(figsize=(18, 10))
plot_mosaic_stat_map(plotted_img,
colorbar=False,
display_mode='z',
bg_img=anat_img,
cut_coords=1,
figure=fh,
title=titles)
def visualize_comparisons_std(self, similarity_std, anat_img=None):
if similarity_std is not None:
RSA_std_img = self.sphere_masker.inverse_transform(
similarity_std[0])
plot_stat_map(RSA_std_img, bg_img=anat_img, title='RSA std')
def make_parcellation(data_path, parcellation, parc_type=None, parc_params=None):
"""
Performs a parcellation which reduces voxel space to regions of interest (brain data).
Parameters
----------
data_path : str
Path to .nii image.
parcellation : str
Specify which parcellation that you would like to use. For MNI: 'gordon2014_333', 'power2012_264', For TAL: 'shen2013_278'.
It is possible to add the OH subcotical atlas on top of a cortical atlas (e.g. gordon) by adding:
'+OH' (for oxford harvard subcortical atlas) and '+SUIT' for SUIT cerebellar atlas.
e.g.: gordon2014_333+OH+SUIT'
parc_type : str
Can be 'sphere' or 'region'. If nothing is specified, the default for that parcellation will be used.
parc_params : dict
**kwargs for nilearn functions
Returns
-------
data : array
Data after the parcellation.
NOTE
----
These functions make use of nilearn. Please cite nilearn if used in a publicaiton.
"""
if isinstance(parcellation, str):
parcin = ''
if '+' in parcellation:
parcin = parcellation
parcellation = parcellation.split('+')[0]
if '+OH' in parcin:
subcortical = True
else:
subcortical = None
if '+SUIT' in parcin:
cerebellar = True
else:
cerebellar = None
if not parc_type or not parc_params:
path = tenetopath[0] + '/data/parcellation_defaults/defaults.json'
with open(path) as data_file:
defaults = json.load(data_file)
if not parc_type:
parc_type = defaults[parcellation]['type']
print('Using default parcellation type')
if not parc_params:
parc_params = defaults[parcellation]['params']
print('Using default parameters')
if parc_type == 'sphere':
parcellation = load_parcellation_coords(parcellation)
seed = NiftiSpheresMasker(np.array(parcellation), **parc_params)
data = seed.fit_transform(data_path)
elif parc_type == 'region':
path = tenetopath[0] + '/data/parcellation/' + parcellation + '.nii.gz'
region = NiftiLabelsMasker(path, **parc_params)
data = region.fit_transform(data_path)
else:
raise ValueError('Unknown parc_type specified')
if subcortical:
subatlas = fetch_atlas_harvard_oxford('sub-maxprob-thr0-2mm')['maps']
region = NiftiLabelsMasker(subatlas, **parc_params)
data_sub = region.fit_transform(data_path)
data = np.hstack([data, data_sub])
if cerebellar:
path = tenetopath[0] + '/data/parcellation/Cerebellum-SUIT_space-MNI152NLin2009cAsym.nii.gz'
region = NiftiLabelsMasker(path, **parc_params)
data_cerebellar = region.fit_transform(data_path)
data = np.hstack([data, data_cerebellar])
return data
def test_nifti_spheres_masker_overlap():
# Test resampling in NiftiMapsMasker
affine = np.eye(4)
shape = (5, 5, 5)
data = np.random.RandomState(42).random_sample(shape + (5, ))
fmri_img = nibabel.Nifti1Image(data, affine)
seeds = [(0, 0, 0), (2, 2, 2)]
overlapping_masker = NiftiSpheresMasker(seeds,
radius=1,
allow_overlap=True)
overlapping_masker.fit_transform(fmri_img)
overlapping_masker = NiftiSpheresMasker(seeds,
radius=2,
allow_overlap=True)
overlapping_masker.fit_transform(fmri_img)
noverlapping_masker = NiftiSpheresMasker(seeds,
radius=1,
allow_overlap=False)
noverlapping_masker.fit_transform(fmri_img)
noverlapping_masker = NiftiSpheresMasker(seeds,
radius=2,
allow_overlap=False)
with pytest.raises(ValueError, match='Overlap detected'):
noverlapping_masker.fit_transform(fmri_img)
# Find the coordinates of the peak
from nibabel.affines import apply_affine
values = z_map.get_data()
coord_peaks = np.dstack(
np.unravel_index(np.argsort(values.ravel()), values.shape))[0, 0, :]
coord_mm = apply_affine(z_map.affine, coord_peaks)
###############################################################################
# We create a masker for the voxel (allowing us to detrend the signal)
# and extract the time course
from nilearn.input_data import NiftiSpheresMasker
mask = NiftiSpheresMasker([coord_mm],
radius=3,
detrend=True,
standardize=True,
high_pass=None,
low_pass=None,
t_r=7.)
sig = mask.fit_transform(fmri_img)
##########################################################
# Let's plot the signal and the theoretical response
plt.plot(frame_times, sig, label='voxel %d %d %d' % tuple(coord_mm))
plt.plot(design_matrix['active'], color='red', label='model')
plt.xlabel('scan')
plt.legend()
plt.show()
def make_parcellation(data_path,
parcellation,
parc_type=None,
parc_params=None):
"""
Performs a parcellation which reduces voxel space to regions of interest (brain data).
Parameters
----------
data_path : str
Path to .nii image.
parcellation : str
Specify which parcellation that you would like to use. For MNI: 'gordon2014_333', 'power2012_264', For TAL: 'shen2013_278'.
It is possible to add the OH subcotical atlas on top of a cortical atlas (e.g. gordon) by adding:
'+sub-maxprob-thr0-1mm', '+sub-maxprob-thr0-2mm', 'sub-maxprob-thr25-1mm', 'sub-maxprob-thr25-2mm',
'+sub-maxprob-thr50-1mm', '+sub-maxprob-thr50-2mm'.
e.g.: gordon2014_333+submaxprob-thr0-2mm'
parc_type : str
Can be 'sphere' or 'region'. If nothing is specified, the default for that parcellation will be used.
parc_params : dict
**kwargs for nilearn functions
Returns
-------
data : array
Data after the parcellation.
NOTE
----
These functions make use of nilearn. Please cite nilearn if used in a publicaiton.
"""
if isinstance(parcellation, str):
if '+' in parcellation:
parcin = parcellation.split('+')
parcellation = parcin[0]
subcortical = parcin[1]
else:
subcortical = None
if not parc_type or not parc_params:
path = teneto.__path__[
0] + '/data/parcellation_defaults/defaults.json'
with open(path) as data_file:
defaults = json.load(data_file)
if not parc_type:
parc_type = defaults[parcellation]['type']
print('Using default parcellation type')
if not parc_params:
parc_params = defaults[parcellation]['params']
print('Using default parameters')
if parc_type == 'sphere':
parcellation = teneto.utils.load_parcellation_coords(parcellation)
seed = NiftiSpheresMasker(np.array(parcellation), **parc_params)
data = seed.fit_transform(data_path)
elif parc_type == 'region':
path = teneto.__path__[
0] + '/data/parcellation/' + parcellation + '.nii'
region = NiftiLabelsMasker(path, **parc_params)
data = region.fit_transform(data_path)
else:
raise ValueError('Unknown parc_type specified')
if subcortical:
subatlas = fetch_atlas_harvard_oxford('sub-maxprob-thr0-2mm')['maps']
region = NiftiLabelsMasker(subatlas, **parc_params)
data_sub = region.fit_transform(data_path)
data = np.hstack([data, data_sub])
return data
def calc_sc_sphere(in_file, coords, coords_labels, MNI_brain_mask, radius,
FWHM):
from nilearn.input_data import NiftiMasker, NiftiSpheresMasker
import numpy as np
import os
# MNI mask
MNI_brain_mask = MNI_brain_mask
# probabilistic seed regions, 4d nifti file & corresponding labels
coords = coords
coords_labels = coords_labels
# extract time series from coords
seed_masker = NiftiSpheresMasker(coords,
radius=radius,
standardize=True,
memory='nilearn_cache',
memory_level=5,
verbose=5)
seed_time_series_array = seed_masker.fit_transform(in_file)
# extract time series brain-wide
brain_masker = NiftiMasker(mask_img=MNI_brain_mask,
smoothing_fwhm=FWHM,
standardize=True,
memory='nilearn_cache',
memory_level=5,
verbose=2)
brain_time_series = brain_masker.fit_transform(in_file)
# check if length of coords_labels is equal to number of seed time series
# (dependend on number of coordinate sets given with "coords")
# break if not equal
if len(coords_labels) == seed_time_series_array.shape[1]:
icoord = 0
corr_maps_dict = dict.fromkeys(coords_labels)
for seed in coords_labels:
print("##################################")
print(seed)
print("##################################")
# assign extracted seed time series from seed_time_series_array
# at column icoord to seed_time_series & transpose it (because otherwise they have not the right dimensions?)
seed_time_series = np.matrix(seed_time_series_array[:, icoord]).T
# compute correlation of all voxels with seed regions
seed_based_correlations = np.dot(brain_time_series.T, seed_time_series) / \
seed_time_series.shape[0]
# increase imask to iterate through columns of seed_time_series_array
icoord = icoord + 1
# Fisher-z transform the data to achieve a normal distribution
seed_based_correlations_fisher_z = np.arctanh(
seed_based_correlations)
# transform the 2 dim matrix with the value of each voxel back to a 3 dim image
seed_based_correlation_img = brain_masker.inverse_transform(
seed_based_correlations_fisher_z.T)
print("##################################")
# initialize an empty file & "fill" it with the calculated img, necessary becaus nipype needs file types or so... aehm hmm
out_file = os.path.abspath('corr_map_' + seed + '_rad5.nii.gz')
seed_based_correlation_img.to_filename(out_file)
corr_maps_dict[seed] = out_file
return corr_maps_dict
else:
print(
"#####################################################################################"
)
print(
"#####################################################################################"
)
print(
"Number of labels in prob_masks_labels and volumes in prob_masks does not match!!!!!!!"
)
print(
"#####################################################################################"
)
print(
"#####################################################################################"
)
# Prepare timing
t_r = 2.
slice_time_ref = 0.
n_scans = 176
# Prepare seed
pcc_coords = (0, -53, 26)
#########################################################################
# Estimate contrasts
# ------------------
# Specify the contrasts
seed_masker = NiftiSpheresMasker([pcc_coords], radius=10, detrend=True,
standardize=True, low_pass=0.1,
high_pass=0.01, t_r=2.,
memory='nilearn_cache',
memory_level=1, verbose=0)
seed_time_series = seed_masker.fit_transform(adhd_dataset.func[0])
frametimes = np.linspace(0, (n_scans - 1) * t_r, n_scans)
design_matrix = make_first_level_design_matrix(frametimes, hrf_model='spm',
add_regs=seed_time_series,
add_reg_names=["pcc_seed"])
dmn_contrast = np.array([1] + [0]*(design_matrix.shape[1]-1))
contrasts = {'seed_based_glm': dmn_contrast}
#########################################################################
# Perform first level analysis
# ----------------------------
# Setup and fit GLM
first_level_model = FirstLevelModel(t_r=t_r, slice_time_ref=slice_time_ref)
def test_errors():
masker = NiftiSpheresMasker(([1, 2]), radius=.2)
with pytest.raises(ValueError, match='Seeds must be a list .+'):
masker.fit()
def atlas_masker_spheres(coords, radius):
masker = NiftiSpheresMasker(coords, radius)
assert isinstance(masker, object)
return masker
# Prepare timing
t_r = 2.
slice_time_ref = 0.
n_scans = 176
# Prepare seed
pcc_coords = (0, -53, 26)
#########################################################################
# Estimate contrasts
# ------------------
# Specify the contrasts
seed_masker = NiftiSpheresMasker([pcc_coords], radius=10, detrend=True,
standardize=True, low_pass=0.1,
high_pass=0.01, t_r=2.,
memory='nilearn_cache',
memory_level=1, verbose=0)
seed_time_series = seed_masker.fit_transform(adhd_dataset.func[0])
frametimes = np.linspace(0, (n_scans - 1) * t_r, n_scans)
design_matrix = make_first_level_design_matrix(frametimes, hrf_model='spm',
add_regs=seed_time_series,
add_reg_names=["pcc_seed"])
dmn_contrast = np.array([1] + [0]*(design_matrix.shape[1]-1))
contrasts = {'seed_based_glm': dmn_contrast}
#########################################################################
# Perform first level analysis
# ----------------------------
# Setup and fit GLM
first_level_model = FirstLevelModel(t_r=t_r, slice_time_ref=slice_time_ref)
def test_errors():
masker = NiftiSpheresMasker(([1, 2]), radius=.2)
assert_raises_regex(ValueError, 'Seeds must be a list .+', masker.fit)
def test_nifti_spheres_masker_inverse_overlap():
rng = np.random.RandomState(42)
# Test overlapping data in inverse_transform
affine = np.eye(4)
shape = (5, 5, 5)
data = rng.random_sample(shape + (5, ))
fmri_img = nibabel.Nifti1Image(data, affine)
# Apply mask image - to allow inversion
mask_img = new_img_like(fmri_img, np.ones(shape))
seeds = [(0, 0, 0), (2, 2, 2)]
# Inverse data
inv_data = rng.random_sample(len(seeds))
overlapping_masker = NiftiSpheresMasker(seeds,
radius=1,
allow_overlap=True,
mask_img=mask_img).fit()
overlapping_masker.inverse_transform(inv_data)
overlapping_masker = NiftiSpheresMasker(seeds,
radius=2,
allow_overlap=True,
mask_img=mask_img).fit()
overlap = overlapping_masker.inverse_transform(inv_data)
# Test whether overlapping data is averaged
assert_array_almost_equal(get_data(overlap)[1, 1, 1], np.mean(inv_data))
noverlapping_masker = NiftiSpheresMasker(seeds,
radius=1,
allow_overlap=False,
mask_img=mask_img).fit()
noverlapping_masker.inverse_transform(inv_data)
noverlapping_masker = NiftiSpheresMasker(seeds,
radius=2,
allow_overlap=False,
mask_img=mask_img).fit()
with pytest.raises(ValueError, match='Overlap detected'):
noverlapping_masker.inverse_transform(inv_data)
def test_nifti_spheres_masker_overlap():
# Test resampling in NiftiMapsMasker
affine = np.eye(4)
shape = (5, 5, 5)
data = np.random.random(shape + (5, ))
fmri_img = nibabel.Nifti1Image(data, affine)
seeds = [(0, 0, 0), (2, 2, 2)]
overlapping_masker = NiftiSpheresMasker(seeds,
radius=1,
allow_overlap=True)
overlapping_masker.fit_transform(fmri_img)
overlapping_masker = NiftiSpheresMasker(seeds,
radius=2,
allow_overlap=True)
overlapping_masker.fit_transform(fmri_img)
noverlapping_masker = NiftiSpheresMasker(seeds,
radius=1,
allow_overlap=False)
noverlapping_masker.fit_transform(fmri_img)
noverlapping_masker = NiftiSpheresMasker(seeds,
radius=2,
allow_overlap=False)
assert_raises_regex(ValueError, 'Overlap detected',
noverlapping_masker.fit_transform, fmri_img)
def test_small_radius_inverse():
affine = np.eye(4)
shape = (3, 3, 3)
data = np.random.RandomState(42).random_sample(shape)
mask = np.zeros(shape)
mask[1, 1, 1] = 1
mask[2, 2, 2] = 1
affine = np.eye(4) * 1.2
seed = (1.4, 1.4, 1.4)
masker = NiftiSpheresMasker([seed],
radius=0.1,
mask_img=nibabel.Nifti1Image(mask, affine))
spheres_data = masker.fit_transform(nibabel.Nifti1Image(data, affine))
masker.inverse_transform(spheres_data)
# Test if masking is taken into account
mask[1, 1, 1] = 0
mask[1, 1, 0] = 1
masker = NiftiSpheresMasker([seed],
radius=0.1,
mask_img=nibabel.Nifti1Image(mask, affine))
masker.fit(nibabel.Nifti1Image(data, affine))
with pytest.raises(ValueError, match='These spheres are empty'):
masker.inverse_transform(spheres_data)
masker = NiftiSpheresMasker([seed],
radius=1.6,
mask_img=nibabel.Nifti1Image(mask, affine))
masker.fit(nibabel.Nifti1Image(data, affine))
masker.inverse_transform(spheres_data)
def test_nifti_spheres_masker_inverse_transform():
# Applying the sphere_extraction example from above backwards
data = np.random.RandomState(42).random_sample((3, 3, 3, 5))
img = nibabel.Nifti1Image(data, np.eye(4))
masker = NiftiSpheresMasker([(1, 1, 1)], radius=1)
# Test the fit
masker.fit()
# Transform data
with pytest.raises(ValueError, match='Please provide mask_img'):
masker.inverse_transform(data[0, 0, 0, :])
# Mask describes the extend of the masker's sphere
mask = np.zeros((3, 3, 3), dtype=np.bool)
mask[:, 1, 1] = True
mask[1, :, 1] = True
mask[1, 1, :] = True
# Now with a mask
mask_img = np.zeros((3, 3, 3))
mask_img[1, :, :] = 1
mask_img = nibabel.Nifti1Image(mask_img, np.eye(4))
masker = NiftiSpheresMasker([(1, 1, 1)], radius=1, mask_img=mask_img)
masker.fit()
s = masker.transform(img)
# Create an array mask
array_mask = np.logical_and(mask, get_data(mask_img))
inverse_map = masker.inverse_transform(s)
# Testing whether mask is applied to inverse transform
assert_array_equal(
np.mean(get_data(inverse_map), axis=-1) != 0, array_mask)
# Test whether values are preserved
assert_array_equal(get_data(inverse_map)[array_mask].mean(0), s[:, 0])
# Test whether the mask's shape is applied
assert_array_equal(inverse_map.shape[:3], mask_img.shape)
def signal_extract(func_data=None,confounds=None,atlas_img=None,masker_type='Spheres',smoothing_fwhm=6,high_pass=0.01,low_pass=0.1,t_r=2.2,detrend=False,saveas='file'):
"""
Extracts BOLD time-series from regions of interest
Parameters
----------
func_data: functional images ( Default= None )
confounds: Confounds file used to clean signals ( Default= None )
atlas_img: regions or coordinates to extract signals from ( Default= None )
masker_type : Type of masker used to extract BOLD signals . types are : 'Spheres','Maps','Labels'
smoothing_fwhm : Smoothing width applied to signals in mm ( Default= 6 mm )
high_pass, low_pass: Bandpass-Filtering ( Default= 0.01-0.1 Hz )
detrend: Detrending signals ( Default= False )
saveas : Destination to save and load output (.npz)
Returns
---------
subject_ts : array-like , 2-D (n_subjects,n_regions)
Array of BOLD time-series
"""
subjects_ts=[]
if os.path.exists(saveas):
subjects_ts=np.load(saveas)['arr_0']
else:
if
if masker_type== 'Spheres':
masker = NiftiSpheresMasker(
seeds=atlas_img, smoothing_fwhm=smoothing_fwhm, radius=4 ,mask_img=brainmask,
detrend=False, standardize=True, low_pass=low_pass, high_pass=high_pass, t_r=t_r
)
elif masker_type == 'Maps':
masker = NiftiMapsMasker(
maps_img=atlas_img,mask_img=brainmask,standardize=True,
low_pass=low_pass, high_pass=high_pass, t_r=t_r,
memory_level=2,smoothing_fwhm=smoothing_fwhm,resampling_target='data',
memory=mem,verbose=5
)
elif masker_type == 'Labels':
masker = NiftiLabelsMasker(
labels_img=atlas_img,mask_img=brainmask,standardize=True,
high_pass=high_pass,low_pass=low_pass,detrend=False,t_r=t_r,
memory_level=2,smoothing_fwhm=smoothing_fwhm,resampling_target='data',
memory=mem,verbose=5
)
else:
raise ValueError("Please provide masker type")
if confounds is not None:
for func_file, confound_file in zip(func_data,confounds):
time_series = masker.fit_transform(func_file,confounds=confound_file)
subjects_ts.append(time_series)
np.savez(saveas,subjects_ts)
else:
for func_file in data:
time_series = masker.fit_transform(func_file)
subjects_ts.append(time_series)
np.savez(saveas,subjects_ts)
return subjects_ts
