Statistical Model for SDI - first level and second level

[venkateshness]

Set up within- and between-subjects statistical tests, alongside the correction for multiple comparison

Detailed Description

First, check the consistency of the effects at the trials/events (intra-subject level); second, the group-level (inter-subject) model followed by the permutation-based correction for the multiple comparison to identify the strength of the effects observed in the empirical SDI comparing against surrogate SDIs across and within subjects

Context / Motivation

Trials/events-level SDI needs to be looked into for its consistency before the group-level; this hierarchical stat model handles that

Possible Implementation

Set up the code for;

1. Make a distribution from the Empirical and Surrogate SDIs using Signed Rank
2. Perform one-sample ttest to test the SDI in the trials/events against null
3. Perform massive univariate tests among subjects, and subsequently correct for multiple comparisons using permutations

• Reimplement the code for Scipy's signed rank test for natively including it in NiGSP
• Run tests and reproduce the original results; checks out okay ?
• Reimplement the code for MNE's ttest_1samp_no_p for native inclusion
• Run tests and reproduce the original results; checks out okay ?
• Figure out if reimplementing MNE-python's permutation_t_test is a viable option; EDIT: many local imports; making it all native make the code convoluted & hard to read and follow, better to keep mne as optional requirement for stats
• Final tests pass ?
• Moderate and quick refactoring the code
• Polishing the docstring
• Run and test locally
• Automated tests