def test_cluster_permutation_test():
"""Test cluster level permutations tests."""
condition1_1d, condition2_1d, condition1_2d, condition2_2d = \
_get_conditions()
for condition1, condition2 in zip((condition1_1d, condition1_2d),
(condition2_1d, condition2_2d)):
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2],
n_permutations=100,
tail=1,
seed=1,
buffer_size=None)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2],
n_permutations=100,
tail=0,
seed=1,
buffer_size=None)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
# test with 2 jobs and buffer_size enabled
buffer_size = condition1.shape[1] // 10
T_obs, clusters, cluster_p_values_buff, hist =\
permutation_cluster_test([condition1, condition2],
n_permutations=100, tail=0, seed=1,
n_jobs=2, buffer_size=buffer_size)
assert_array_equal(cluster_p_values, cluster_p_values_buff)
def test_cluster_permutation_test(numba_conditional):
"""Test cluster level permutations tests."""
condition1_1d, condition2_1d, condition1_2d, condition2_2d = \
_get_conditions()
for condition1, condition2 in zip((condition1_1d, condition1_2d),
(condition2_1d, condition2_2d)):
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2],
n_permutations=100,
tail=1,
seed=1,
buffer_size=None,
out_type='mask')
p_min = np.min(cluster_p_values)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
assert_allclose(p_min, 0.01, atol=1e-6)
# test with 2 jobs and buffer_size enabled
buffer_size = condition1.shape[1] // 10
T_obs, clusters, cluster_p_values_buff, hist =\
permutation_cluster_test([condition1, condition2],
n_permutations=100, tail=1, seed=1,
n_jobs=2, buffer_size=buffer_size,
out_type='mask')
assert_array_equal(cluster_p_values, cluster_p_values_buff)
def stat_fun(X, Y):
return stats.f_oneway(X, Y)[0]
with pytest.warns(RuntimeWarning, match='is only valid'):
permutation_cluster_test([condition1, condition2],
n_permutations=1,
stat_fun=stat_fun,
out_type='mask')
def test_thresholds(numba_conditional):
"""Test automatic threshold calculations."""
# within subjects
rng = np.random.RandomState(0)
X = rng.randn(10, 1, 1) + 0.08
want_thresh = -stats.t.ppf(0.025, len(X) - 1)
assert 0.03 < stats.ttest_1samp(X[:, 0, 0], 0)[1] < 0.05
my_fun = partial(ttest_1samp_no_p)
with catch_logging() as log:
with pytest.warns(RuntimeWarning, match='threshold is only valid'):
out = permutation_cluster_1samp_test(X,
stat_fun=my_fun,
seed=0,
verbose=True,
out_type='mask')
log = log.getvalue()
assert str(want_thresh)[:6] in log
assert len(out[1]) == 1 # 1 cluster
assert_allclose(out[2], 0.033203, atol=1e-6)
# between subjects
Y = rng.randn(10, 1, 1)
Z = rng.randn(10, 1, 1) - 0.7
X = [X, Y, Z]
want_thresh = stats.f.ppf(1. - 0.05, 2, sum(len(a) for a in X) - len(X))
p = stats.f_oneway(*X)[1]
assert 0.03 < p < 0.05
my_fun = partial(f_oneway) # just to make the check fail
with catch_logging() as log:
with pytest.warns(RuntimeWarning, match='threshold is only valid'):
out = permutation_cluster_test(X,
tail=1,
stat_fun=my_fun,
seed=0,
verbose=True,
out_type='mask')
log = log.getvalue()
assert str(want_thresh)[:6] in log
assert len(out[1]) == 1 # 1 cluster
assert_allclose(out[2], 0.041992, atol=1e-6)
with pytest.warns(RuntimeWarning, match='Ignoring argument "tail"'):
permutation_cluster_test(X, tail=0, out_type='mask')
# nan handling in TFCE
X = np.repeat(X[0], 2, axis=1)
X[:, 1] = 0
with pytest.warns(RuntimeWarning, match='invalid value'): # NumPy
out = permutation_cluster_1samp_test(X,
seed=0,
threshold=dict(start=0, step=0.1),
out_type='mask')
assert (out[2] < 0.05).any()
assert not (out[2] < 0.05).all()
X[:, 0] = 0
with pytest.raises(RuntimeError, match='finite'):
with np.errstate(invalid='ignore'):
permutation_cluster_1samp_test(X,
seed=0,
threshold=dict(start=0, step=0.1),
buffer_size=None,
out_type='mask')
def test_thresholds():
"""Test automatic threshold calculations."""
# within subjects
rng = np.random.RandomState(0)
X = rng.randn(10, 1, 1) + 0.08
want_thresh = -stats.t.ppf(0.025, len(X) - 1)
assert 0.03 < stats.ttest_1samp(X[:, 0, 0], 0)[1] < 0.05
my_fun = partial(ttest_1samp_no_p)
with catch_logging() as log:
with pytest.warns(RuntimeWarning, match='threshold is only valid'):
out = permutation_cluster_1samp_test(X, stat_fun=my_fun,
verbose=True)
log = log.getvalue()
assert str(want_thresh)[:6] in log
assert len(out[1]) == 1 # 1 cluster
assert 0.03 < out[2] < 0.05
# between subjects
Y = rng.randn(10, 1, 1)
Z = rng.randn(10, 1, 1) - 0.7
X = [X, Y, Z]
want_thresh = stats.f.ppf(1. - 0.05, 2, sum(len(a) for a in X) - len(X))
p = stats.f_oneway(*X)[1]
assert 0.03 < p < 0.05
my_fun = partial(f_oneway) # just to make the check fail
with catch_logging() as log:
with pytest.warns(RuntimeWarning, match='threshold is only valid'):
out = permutation_cluster_test(X, tail=1, stat_fun=my_fun,
verbose=True)
log = log.getvalue()
assert str(want_thresh)[:6] in log
assert len(out[1]) == 1 # 1 cluster
assert 0.03 < out[2] < 0.05
with pytest.warns(RuntimeWarning, match='Ignoring argument "tail"'):
permutation_cluster_test(X, tail=0)
def test_cluster_permutation_test():
"""Test cluster level permutations tests
"""
condition1_1d, condition2_1d, condition1_2d, condition2_2d = \
_get_conditions()
for condition1, condition2 in zip((condition1_1d, condition1_2d),
(condition2_1d, condition2_2d)):
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2],
n_permutations=100, tail=1, seed=1,
buffer_size=None)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2],
n_permutations=100, tail=0, seed=1,
buffer_size=None)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
# test with 2 jobs and buffer_size enabled
buffer_size = condition1.shape[1] // 10
T_obs, clusters, cluster_p_values_buff, hist =\
permutation_cluster_test([condition1, condition2],
n_permutations=100, tail=0, seed=1,
n_jobs=2, buffer_size=buffer_size)
assert_array_equal(cluster_p_values, cluster_p_values_buff)
def test_cluster_permutation_test():
"""Test cluster level permutations tests."""
condition1_1d, condition2_1d, condition1_2d, condition2_2d = \
_get_conditions()
for condition1, condition2 in zip((condition1_1d, condition1_2d),
(condition2_1d, condition2_2d)):
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2], n_permutations=100, tail=1, seed=1,
buffer_size=None)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2], n_permutations=100, tail=0, seed=1,
buffer_size=None)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
# test with 2 jobs and buffer_size enabled
buffer_size = condition1.shape[1] // 10
T_obs, clusters, cluster_p_values_buff, hist =\
permutation_cluster_test([condition1, condition2],
n_permutations=100, tail=0, seed=1,
n_jobs=2, buffer_size=buffer_size)
assert_array_equal(cluster_p_values, cluster_p_values_buff)
def stat_fun(X, Y):
return stats.f_oneway(X, Y)[0]
with warnings.catch_warnings(record=True) as w:
permutation_cluster_test([condition1, condition2], n_permutations=1,
stat_fun=stat_fun)
assert_equal(len(w), 1)
assert 'is only valid' in str(w[0].message)
def test_cluster_permutation_test():
"""Test cluster level permutations tests."""
condition1_1d, condition2_1d, condition1_2d, condition2_2d = \
_get_conditions()
for condition1, condition2 in zip((condition1_1d, condition1_2d),
(condition2_1d, condition2_2d)):
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2], n_permutations=100, tail=1, seed=1,
buffer_size=None)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2], n_permutations=100, tail=1, seed=1,
buffer_size=None)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
# test with 2 jobs and buffer_size enabled
buffer_size = condition1.shape[1] // 10
T_obs, clusters, cluster_p_values_buff, hist =\
permutation_cluster_test([condition1, condition2],
n_permutations=100, tail=1, seed=1,
n_jobs=2, buffer_size=buffer_size)
assert_array_equal(cluster_p_values, cluster_p_values_buff)
def stat_fun(X, Y):
return stats.f_oneway(X, Y)[0]
with pytest.warns(RuntimeWarning, match='is only valid'):
permutation_cluster_test([condition1, condition2], n_permutations=1,
stat_fun=stat_fun)
def test_thresholds(numba_conditional):
"""Test automatic threshold calculations."""
# within subjects
rng = np.random.RandomState(0)
X = rng.randn(10, 1, 1) + 0.08
want_thresh = -stats.t.ppf(0.025, len(X) - 1)
assert 0.03 < stats.ttest_1samp(X[:, 0, 0], 0)[1] < 0.05
my_fun = partial(ttest_1samp_no_p)
with catch_logging() as log:
with pytest.warns(RuntimeWarning, match='threshold is only valid'):
out = permutation_cluster_1samp_test(X, stat_fun=my_fun,
seed=0, verbose=True)
log = log.getvalue()
assert str(want_thresh)[:6] in log
assert len(out[1]) == 1 # 1 cluster
assert_allclose(out[2], 0.033203, atol=1e-6)
# between subjects
Y = rng.randn(10, 1, 1)
Z = rng.randn(10, 1, 1) - 0.7
X = [X, Y, Z]
want_thresh = stats.f.ppf(1. - 0.05, 2, sum(len(a) for a in X) - len(X))
p = stats.f_oneway(*X)[1]
assert 0.03 < p < 0.05
my_fun = partial(f_oneway) # just to make the check fail
with catch_logging() as log:
with pytest.warns(RuntimeWarning, match='threshold is only valid'):
out = permutation_cluster_test(X, tail=1, stat_fun=my_fun,
seed=0, verbose=True)
log = log.getvalue()
assert str(want_thresh)[:6] in log
assert len(out[1]) == 1 # 1 cluster
assert_allclose(out[2], 0.041992, atol=1e-6)
with pytest.warns(RuntimeWarning, match='Ignoring argument "tail"'):
permutation_cluster_test(X, tail=0)
def test_cluster_permutation_test():
"""Test cluster level permutations tests."""
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2], n_permutations=500,
tail=1)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2], n_permutations=500,
tail=0)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
def test_cluster_permutation_test():
"""Test cluster level permutations tests."""
for condition1, condition2 in zip((condition1_1d, condition1_2d),
(condition2_1d, condition2_2d)):
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2], n_permutations=500, tail=1, seed=1)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2], n_permutations=500, tail=0, seed=1)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
# test with 2 jobs
T_obs, clusters, cluster_p_values_buff, hist =\
permutation_cluster_test([condition1, condition2],
n_permutations=500, tail=0, seed=1,
n_jobs=2)
assert_array_equal(cluster_p_values, cluster_p_values_buff)
def test_cluster_permutation_test():
"""Test cluster level permutations tests."""
for condition1, condition2 in zip((condition1_1d, condition1_2d),
(condition2_1d, condition2_2d)):
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2],
n_permutations=100, tail=1, seed=1)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
T_obs, clusters, cluster_p_values, hist = permutation_cluster_test(
[condition1, condition2],
n_permutations=100, tail=0, seed=1)
assert_equal(np.sum(cluster_p_values < 0.05), 1)
# test with 2 jobs
T_obs, clusters, cluster_p_values_buff, hist =\
permutation_cluster_test([condition1, condition2],
n_permutations=100, tail=0, seed=1,
n_jobs=2)
assert_array_equal(cluster_p_values, cluster_p_values_buff)
def _conduct_analysis(data, contrast, space, avg_freq, selection,
n_permutations, cluster_p_threshold, confounds, verbose):
'''Conduct analysis.'''
stat_info, adjacency = data['stat_info'], data['adjacency']
if 'vs' in contrast and not confounds:
hi, lo = data['hi'], data['lo']
psd_ndim = hi.ndim
else:
hilo = data['hilo']
bdi = data['bdi']
psd_ndim = hilo.ndim
# put spatial dimension last for cluster-based test
if not avg_freq or psd_ndim == 3:
if 'hi' in data:
hi = hi.transpose((0, 2, 1))
lo = lo.transpose((0, 2, 1))
else:
hilo = hilo.transpose((0, 2, 1))
# statistical analysis
# --------------------
if 'pairs' not in selection:
# cluster-based permutation tests for multiple comparisons
stat_info.update(dict(cluster_p_threshold=cluster_p_threshold))
if 'vs' in contrast and not confounds:
# t test in cluster-based permutation test
from scipy.stats import t
from sarna.stats import ttest_ind_welch_no_p
from mne.stats.cluster_level import permutation_cluster_test
# calculate t test threshold
df = hi.shape[0] + lo.shape[0] - 2
threshold = t.ppf(1 - cluster_p_threshold / 2, df)
# run cluster-based permutation test
args = dict(threshold=threshold,
n_permutations=n_permutations,
stat_fun=ttest_ind_welch_no_p,
verbose=verbose,
out_type='mask')
try:
stat, clusters, pval, _ = permutation_cluster_test(
[hi, lo], **args, connectivity=adjacency)
except TypeError:
stat, clusters, pval, _ = permutation_cluster_test(
[hi, lo], **args, adjacency=adjacency)
else:
# regression in cluster-based permutation test
from borsar.cluster import cluster_based_regression
args = dict(n_permutations=n_permutations,
adjacency=adjacency,
alpha_threshold=cluster_p_threshold,
cluster_pred=-1)
stat, clusters, pval = cluster_based_regression(hilo, bdi, **args)
# construct Clusters with stat_info in description
return _construct_clusters(clusters, pval, stat, space, stat_info,
data['info'], data['src'],
data['subjects_dir'], data['subject'],
data['ch_names'], data['freq'])
else:
# for selected pairs (two channel pairs) we don't correct for
# multiple comparisons:
if 'vs' in contrast and not confounds:
from scipy.stats import t, ttest_ind
stat, pval = ttest_ind(hi, lo, equal_var=False)
else:
from borsar.stats import compute_regression_t
# compute regression and ignore intercept:
stat, pval = compute_regression_t(hilo, bdi, return_p=True)
stat, pval = stat[-1], pval[-1]
stat_info.update(dict(stat=stat, pval=pval, ch_names=ch_names))
return stat_info
