Statistics#

frites.stats:

Statistical methods.

This submodule contains a collection of statistical internal methods divided into two categories :

  1. Random efect estimation : t-test related functions

  2. P-values correction for multiple comparisons : test-wise and cluster based corrections

Most of those stastical functions are using MNE Python

Non-parametric statistics#

permute_mi_vector(y, suj[, mi_type, ...])

Permute regressor variable for performing non-parameteric statistics.

permute_mi_trials(suj[, inference, n_perm, ...])

Generate random partitions for swapping trials.

bootstrap_partitions(n_epochs, *groups[, ...])

Generate partitions for bootstrap.

confidence_interval(data[, axis, cis, ...])

Compute the confidence interval of repeated measurements.

Random-effect (rfx)#

ttest_1samp(x, pop_mean[, axis, ...])

One-sample t-test.

rfx_ttest(mi, mi_p[, center, sigma, ttested])

Perform the t-test across subjects.

Correction for multiple comparisons#

cluster_correction_mcp(x, x_p, th[, tail])

Cluster-based correction for MCP using non-parametric statistics.

testwise_correction_mcp(x, x_p[, tail, mcp])

Test-wise correction for MCP using non-parametric statistics.

cluster_threshold(x, x_p[, alpha, tail, ...])

Threshold detection for cluster-based inferencse.