Code of my master thesis investigating functional connectivity changes within the fronto-limbic network and default mode network of participants with bipolar disorder or being at-risk of developing it and healthy controls. Code contains resting-state analysis with nilearn, statistical analysis and plots of the results.
The analysis steps are ordered from 1 to 4:
- Creating a mask for all ROIs (containing regions of the FLN and DMN) based on the DiFuMo atlas
- Extracting of the timeseries of every participant and calculation the partialy correlation matrices based on the individual timeseries for each subject + calculating the group partial correlation matrices
- Creating a dataframe that contains all the necessary information (demographical data + partial correlation values of all connections for each participant)
- Calculating an OLS regression for the effect of the groups and IDSC & YMRS scores
Further, the code for every visualisation is made available, containing:
- connectome and circular plots for the groups
- plot of all the regions contained within the created mask
- plot of a single-subject partial correlation matrix