def test_high_level_glm_with_paths():
with InTemporaryDirectory():
shapes = ((7, 8, 9, 1),)
mask, FUNCFILE, _ = write_fake_fmri_data_and_design(shapes)
FUNCFILE = FUNCFILE[0]
func_img = load(FUNCFILE)
# Ordinary Least Squares case
model = SecondLevelModel(mask_img=mask)
# asking for contrast before model fit gives error
with pytest.raises(ValueError):
model.compute_contrast([])
# fit model
Y = [func_img] * 4
X = pd.DataFrame([[1]] * 4, columns=['intercept'])
model = model.fit(Y, design_matrix=X)
c1 = np.eye(len(model.design_matrix_.columns))[0]
z_image = model.compute_contrast(c1, output_type='z_score')
assert isinstance(z_image, Nifti1Image)
assert_array_equal(z_image.affine, load(mask).affine)
# try with target_shape
target_shape = (10, 10, 10)
target_affine = np.eye(4)
target_affine[0, 3] = 1
model = SecondLevelModel(mask_img=mask, target_shape=target_shape,
target_affine=target_affine)
z_image = model.fit(Y, design_matrix=X).compute_contrast(c1)
assert_array_equal(z_image.shape, target_shape)
assert_array_equal(z_image.affine, target_affine)
# Delete objects attached to files to avoid WindowsError when deleting
# temporary directory (in Windows)
del Y, FUNCFILE, func_img, model
def test_second_level_model_contrast_computation_with_memory_caching():
with InTemporaryDirectory():
shapes = ((7, 8, 9, 1),)
mask, FUNCFILE, _ = write_fake_fmri_data_and_design(shapes)
FUNCFILE = FUNCFILE[0]
func_img = load(FUNCFILE)
# Ordinary Least Squares case
model = SecondLevelModel(mask_img=mask, memory='nilearn_cache')
# fit model
Y = [func_img] * 4
X = pd.DataFrame([[1]] * 4, columns=['intercept'])
model = model.fit(Y, design_matrix=X)
ncol = len(model.design_matrix_.columns)
c1 = np.eye(ncol)[0, :]
# test memory caching for compute_contrast
model.compute_contrast(c1, output_type='z_score')
# or simply pass nothing
model.compute_contrast()
# Delete objects attached to files to avoid WindowsError when deleting
# temporary directory (in Windows)
del func_img, FUNCFILE, model, X, Y
def test_second_level_model_glm_computation():
with InTemporaryDirectory():
shapes = ((7, 8, 9, 1),)
mask, FUNCFILE, _ = write_fake_fmri_data_and_design(shapes)
FUNCFILE = FUNCFILE[0]
func_img = load(FUNCFILE)
# Ordinary Least Squares case
model = SecondLevelModel(mask_img=mask)
Y = [func_img] * 4
X = pd.DataFrame([[1]] * 4, columns=['intercept'])
model = model.fit(Y, design_matrix=X)
model.compute_contrast()
labels1 = model.labels_
results1 = model.results_
labels2, results2 = run_glm(
model.masker_.transform(Y), X.values, 'ols')
assert_almost_equal(labels1, labels2, decimal=1)
assert len(results1) == len(results2)
# Delete objects attached to files to avoid WindowsError when deleting
# temporary directory (in Windows)
del func_img, FUNCFILE, model, X, Y
def test_second_level_model_contrast_computation():
with InTemporaryDirectory():
shapes = ((7, 8, 9, 1),)
mask, FUNCFILE, _ = write_fake_fmri_data_and_design(shapes)
FUNCFILE = FUNCFILE[0]
func_img = load(FUNCFILE)
# Ordinary Least Squares case
model = SecondLevelModel(mask_img=mask)
# asking for contrast before model fit gives error
with pytest.raises(ValueError):
model.compute_contrast('intercept')
# fit model
Y = [func_img] * 4
X = pd.DataFrame([[1]] * 4, columns=['intercept'])
model = model.fit(Y, design_matrix=X)
ncol = len(model.design_matrix_.columns)
c1, cnull = np.eye(ncol)[0, :], np.zeros(ncol)
# smoke test for different contrasts in fixed effects
model.compute_contrast(c1)
z_image = model.compute_contrast(c1, output_type='z_score')
stat_image = model.compute_contrast(c1, output_type='stat')
p_image = model.compute_contrast(c1, output_type='p_value')
effect_image = model.compute_contrast(c1, output_type='effect_size')
variance_image = \
model.compute_contrast(c1, output_type='effect_variance')
# Test output_type='all', and verify images are equivalent
all_images = model.compute_contrast(c1, output_type='all')
assert_array_equal(get_data(all_images['z_score']),
get_data(z_image))
assert_array_equal(get_data(all_images['stat']),
get_data(stat_image))
assert_array_equal(get_data(all_images['p_value']),
get_data(p_image))
assert_array_equal(get_data(all_images['effect_size']),
get_data(effect_image))
assert_array_equal(get_data(all_images['effect_variance']),
get_data(variance_image))
# formula should work (passing variable name directly)
model.compute_contrast('intercept')
# or simply pass nothing
model.compute_contrast()
# passing null contrast should give back a value error
with pytest.raises(ValueError):
model.compute_contrast(cnull)
# passing wrong parameters
with pytest.raises(ValueError):
model.compute_contrast([])
with pytest.raises(ValueError):
model.compute_contrast(c1, None, '')
with pytest.raises(ValueError):
model.compute_contrast(c1, None, [])
with pytest.raises(ValueError):
model.compute_contrast(c1, None, None, '')
# check that passing no explicit contrast when the design
# matrix has more than one columns raises an error
X = pd.DataFrame(np.random.rand(4, 2), columns=['r1', 'r2'])
model = model.fit(Y, design_matrix=X)
with pytest.raises(ValueError):
model.compute_contrast(None)
# Delete objects attached to files to avoid WindowsError when deleting
# temporary directory (in Windows)
del func_img, FUNCFILE, model, X, Y
ax = plot_design_matrix(design_matrix)
ax.set_title('Second level design matrix', fontsize=12)
ax.set_ylabel('maps')
##########################################################################
# Next, we specify and fit the second-level model when loading the data and also
# smooth a little bit to improve statistical behavior.
from nilearn.stats.second_level_model import SecondLevelModel
second_level_model = SecondLevelModel(smoothing_fwhm=2.0, mask_img=mask_img)
second_level_model.fit(gray_matter_map_filenames, design_matrix=design_matrix)
##########################################################################
# Estimating the contrast is very simple. We can just provide the column
# name of the design matrix.
z_map = second_level_model.compute_contrast(second_level_contrast=[1, 0, 0],
output_type='z_score')
###########################################################################
# We threshold the second level contrast at uncorrected p < 0.001 and plot it.
from nilearn import plotting
from nilearn.stats import map_threshold
_, threshold = map_threshold(z_map, alpha=.05, height_control='fdr')
print('The FDR=.05-corrected threshold is: %.3g' % threshold)
display = plotting.plot_stat_map(
z_map,
threshold=threshold,
colorbar=True,
display_mode='z',
cut_coords=[-4, 26],
title='age effect on grey matter density (FDR = .05)')
# We just have to provide the list of fitted FirstLevelModel objects
# to the SecondLevelModel object for estimation. We can do this because
# all subjects share a similar design matrix (same variables reflected in
# column names).
from nilearn.stats.second_level_model import SecondLevelModel
second_level_input = models
#########################################################################
# Note that we apply a smoothing of 8mm.
second_level_model = SecondLevelModel(smoothing_fwhm=8.0)
second_level_model = second_level_model.fit(second_level_input)
#########################################################################
# Computing contrasts at the second level is as simple as at the first level.
# Since we are not providing confounders we are performing a one-sample test
# at the second level with the images determined by the specified first level
# contrast.
zmap = second_level_model.compute_contrast(
first_level_contrast='language-string')
#########################################################################
# The group level contrast reveals a left lateralized fronto-temporal
# language network.
plotting.plot_glass_brain(zmap,
colorbar=True,
threshold=p001_unc,
title='Group language network (unc p<0.001)',
plot_abs=False,
display_mode='x')
plotting.show()
import pandas as pd
second_level_input = data['cmaps']
design_matrix = pd.DataFrame([1] * len(second_level_input),
columns=['intercept'])
############################################################################
# Model specification and fit
from nilearn.stats.second_level_model import SecondLevelModel
second_level_model = SecondLevelModel(smoothing_fwhm=8.0)
second_level_model = second_level_model.fit(second_level_input,
design_matrix=design_matrix)
##########################################################################
# To estimate the contrast is very simple. We can just provide the column
# name of the design matrix.
z_map = second_level_model.compute_contrast(output_type='z_score')
###########################################################################
# We threshold the second level contrast at uncorrected p < 0.001 and plot
from scipy.stats import norm
p_val = 0.001
p001_uncorrected = norm.isf(p_val)
from nilearn.stats import cluster_level_inference
proportion_true_discoveries_img = cluster_level_inference(z_map,
threshold=[3, 4, 5],
alpha=.05)
plotting.plot_stat_map(
proportion_true_discoveries_img,
threshold=0.,
############################################################################
# plot the design_matrix.
from nilearn.reporting import plot_design_matrix
plot_design_matrix(design_matrix)
############################################################################
# We formally specify the analysis model and fit it.
from nilearn.stats.second_level_model import SecondLevelModel
second_level_model = SecondLevelModel().fit(second_level_input,
design_matrix=design_matrix)
##########################################################################
# Estimating the contrast is very simple. We can just provide the column
# name of the design matrix.
z_map = second_level_model.compute_contrast('vertical vs horizontal',
output_type='z_score')
###########################################################################
# We threshold the second level contrast and plot it.
threshold = 3.1 # correponds to p < .001, uncorrected
display = plotting.plot_glass_brain(
z_map,
threshold=threshold,
colorbar=True,
plot_abs=False,
title='vertical vs horizontal checkerboard (unc p<0.001')
plotting.show()
###########################################################################
# Unsurprisingly, we see activity in the primary visual cortex, both positive
# We define the input maps and the design matrix for the second level model
# and fit it.
import pandas as pd
design_matrix = pd.DataFrame(np.hstack((tested_var, np.ones_like(tested_var))),
columns=['fluency', 'intercept'])
###########################################################################
# Fit of the second-level model
from nilearn.stats.second_level_model import SecondLevelModel
model = SecondLevelModel(smoothing_fwhm=5.0)
model.fit(contrast_map_filenames, design_matrix=design_matrix)
##########################################################################
# To estimate the contrast is very simple. We can just provide the column
# name of the design matrix.
z_map = model.compute_contrast('fluency', output_type='z_score')
###########################################################################
# We compute the fdr-corrected p = 0.05 threshold for these data
from nilearn.stats import map_threshold
_, threshold = map_threshold(z_map, alpha=.05, height_control='fdr')
###########################################################################
# Let us plot the second level contrast at the computed thresholds
from nilearn import plotting
plotting.plot_stat_map(
z_map,
threshold=threshold,
colorbar=True,
title='Group-level association between motor activity \n'
'and reading fluency (fdr=0.05)')