Skip to content

E. Statistical Differences in FC across groups

Jump to bottom
isabelgephart edited this page Jun 7, 2024 · 3 revisions

The second question we try to answer in this work is whether systematic differences in in-scanner thought patterns can yield significant differneces in FC. To do this, we rely on Network-based Statistics as implemented in the NBS toolbox.

The notebooks involved in this part of the analyses are:

  1. S13_NBSprepape_BasicAndSbjAware: this notebook prepares all inputs necessary to run NBS based on the separation of scans that we have obtained during the initial analysis of the experiential data. It prepares data for two different NBS runs. One, in which we do not take into account that there are multiple scans per subject. This was used in some original analyses, but it is now outdated, as ignoring subject identity is not necessarily the best way to go. A second one in which do take into account subject identity. This second NBS analysis is the one we report on the manuscript.

Once you run this notebook you need to:

1.1. Run NBS for the different contrasts, as instructed in the notebook. Do not forget to save the models.

1.2. Run the matlab/NBS2BrainViewer.m script to transform the models from .mat to .txt so that the following notebooks can read them. This might change in the future, as we could also load the models back into python using the loadmat functions from scipy.

  1. S14_PostNBS_PlotFCdiffs: we use this notebook to explore the NBS models. This notebook can generate Circos plots and the network-level summary matrices presented in the figures.

  2. CONN_NBS_IndividualContrast_onBrain: this MATLAB script used CONN to generate Glass Brain views of the connections surviving the NBS contrast.

  3. S15_NBS_Literature_Search_Results: this notebook creates the boxplots that we use to compare our NBS results to those from prior literature.

Clone this wiki locally