def test_z_score():
p = np.random.rand(10)
assert_array_almost_equal(norm.sf(z_score(p)), p)
# check the numerical precision
for p in [1.e-250, 1 - 1.e-16]:
assert_array_almost_equal(z_score(p), norm.isf(p))
assert_array_almost_equal(z_score(np.float32(1.e-100)), norm.isf(1.e-300))
def print_images1(files, masker, mask_array, curr_work_folder, coords):
"""Convert NIfTI files to images"""
for file in files:
pdata = nib.load(os.path.join(curr_work_folder, file)).get_data()
pvals = pdata[mask_array > 0]
z_map = masker.inverse_transform(z_score(np.power(10, -np.abs(pvals))))
threshold1 = fdr_threshold(np.power(10, -np.abs(pvals)), 0.05)
thresholded_map2, threshold2 = map_threshold(z_map,
threshold=.05,
height_control='fdr')
print(file, threshold1, threshold2)
# According to internet
plotting.plot_stat_map(thresholded_map2,
threshold=threshold2,
cut_coords=coords,
title='Thresholded z map, expected fdr = .05',
output_file=os.path.join(
curr_work_folder,
os.path.splitext(file)[0] + '_01'))
# Flop
plotting.plot_stat_map(os.path.join(working_folder, file),
threshold=threshold1,
cut_coords=coords,
title='Thresholded z map, expected fdr = .05',
output_file=os.path.join(
curr_work_folder,
os.path.splitext(file)[0] + '_02'))
# Does this make sense
plotting.plot_stat_map(thresholded_map2,
threshold=threshold1,
cut_coords=coords,
title='Thresholded z map, expected fdr = .05',
output_file=os.path.join(
curr_work_folder,
os.path.splitext(file)[0] + '_03'))
#
plotting.plot_stat_map(os.path.join(working_folder, file),
threshold=threshold2,
cut_coords=coords,
title='Thresholded z map, expected fdr = .05',
output_file=os.path.join(
curr_work_folder,
os.path.splitext(file)[0] + '_04'))
memory='nilearn_cache',
memory_level=1) # cache options
cmap_filenames = localizer_dataset.cmaps
fmri_masked = nifti_masker.fit_transform(cmap_filenames)
#########################################################################
# Perform the second level analysis
# ----------------------------------
# perform a one-sample test on these values
from scipy.stats import ttest_1samp
_, p_values = ttest_1samp(fmri_masked, 0)
#########################################################################
# z-transform of p-values
from nistats.utils import z_score
z_map = nifti_masker.inverse_transform(z_score(p_values))
#########################################################################
# Threshold the resulting map:
# false positive rate < .001, cluster size > 10 voxels
from nistats.thresholding import map_threshold
thresholded_map1, threshold1 = map_threshold(z_map,
threshold=.001,
height_control='fpr',
cluster_threshold=10)
#########################################################################
# Now use FDR <.05, no cluster-level threshold
thresholded_map2, threshold2 = map_threshold(z_map,
threshold=.05,
height_control='fdr')
n_subjects=n_samples)
# mask data
nifti_masker = NiftiMasker(
smoothing_fwhm=5,
memory='nilearn_cache', memory_level=1) # cache options
cmap_filenames = localizer_dataset.cmaps
fmri_masked = nifti_masker.fit_transform(cmap_filenames)
# perform a one-sample test on these values
from scipy.stats import ttest_1samp
_, p_values = ttest_1samp(fmri_masked, 0)
# z-transform of p-values
from nistats.utils import z_score
z_map = nifti_masker.inverse_transform(z_score(p_values))
# Threshold the resulting map:
# false positive rate < .001, cluster size > 10 voxels
from nistats.thresholding import map_threshold
thresholded_map1, threshold1 = map_threshold(
z_map, threshold=.001, height_control='fpr', cluster_threshold=10)
# Now use FDR <.05, no cluster-level threshold
thresholded_map2, threshold2 = map_threshold(
z_map, threshold=.05, height_control='fdr')
# Visualization
from nilearn import plotting
display = plotting.plot_stat_map(z_map, title='Raw z map')
plotting.plot_stat_map(
