osl_dynamics.analysis.statistics#
Functions to perform statistical significance testing.
Note
This module is used in the following tutorials:
Module Contents#
Functions#
|
Check shapes and remove subjects from GLM data which contain nans. |
|
Statistical significant testing for evoked responses. |
|
Statistical significant testing for the difference between two groups. |
|
Statistical significant testing for paired differences. |
Attributes#
- osl_dynamics.analysis.statistics._check_glm_data(data, covariates, assignments=None)[source]#
Check shapes and remove subjects from GLM data which contain nans.
- osl_dynamics.analysis.statistics.evoked_response_max_stat_perm(data, n_perm, covariates=None, metric='copes', n_jobs=1)[source]#
Statistical significant testing for evoked responses.
This function fits a General Linear Model (GLM) with ordinary least squares and performs a sign flip permutations test with the maximum statistic to determine a p-value for evoked responses.
- Parameters:
data (np.ndarray) – Baseline corrected evoked responses. This will be the target data for the GLM. Must be shape (n_subjects, n_samples, …).
n_perm (int) – Number of permutations.
covariates (dict, optional) – Covariates (extra regressors) to add to the GLM fit. These will be z-transformed. Must be of shape (n_subjects,).
metric (str, optional) – Metric to use to build the null distribution. Can be
'tstats'or'copes'.n_jobs (int, optional) – Number of processes to run in parallel.
- Returns:
pvalues – P-values for the evoked response. Shape is (n_subjects, n_samples, …).
- Return type:
np.ndarray
- osl_dynamics.analysis.statistics.group_diff_max_stat_perm(data, assignments, n_perm, covariates=None, metric='tstats', n_jobs=1)[source]#
Statistical significant testing for the difference between two groups.
This function fits a General Linear Model (GLM) with ordinary least squares and performs a row shuffle permutations test with the maximum statistic to determine a p-value for differences between two groups.
- Parameters:
data (np.ndarray) – Subject-specific quantities to compare. This will be the target data for the GLM. Must be shape (n_subjects, features1, features2, …).
assignments (np.ndarray) – 1D numpy array containing group assignments. A value of 1 indicates Group1 and a value of 2 indicates Group2. Note, we test the contrast
abs(Group1 - Group2) > 0.n_perm (int) – Number of permutations.
covariates (dict, optional) – Covariates (extra regressors) to add to the GLM fit. These will be z-transformed. Must be of shape (n_subjects,).
metric (str, optional) – Metric to use to build the null distribution. Can be
'tstats'or'copes'.n_jobs (int, optional) – Number of processes to run in parallel.
- Returns:
group_diff (np.ndarray) – Group difference: Group1 - Group2. Shape is (features1, features2, …).
pvalues (np.ndarray) – P-values for the features. Shape is (features1, features2, …).
- osl_dynamics.analysis.statistics.paired_diff_max_stat_perm(data, n_perm, metric='tstats', n_jobs=1)[source]#
Statistical significant testing for paired differences.
This function fits a General Linear Model (GLM) with ordinary least squares and performs a sign flip permutations test with the maximum statistic to determine a p-value for paired differences.
- Parameters:
data (np.ndarray) – Paired differences to compare. This will be the target data for the GLM. Must be shape (n_subjects, features1, features2, …).
n_perm (int) – Number of permutations.
metric (str, optional) – Metric to use to build the null distribution. Can be
'tstats'or'copes'.n_jobs (int, optional) – Number of processes to run in parallel.
- Returns:
paired_diff (np.ndarray) – Paired differences. Shape is (features1, features2, …).
pvalues (np.ndarray) – P-values for the features. Shape is (features1, features2, …).