Code to compute partial least square (PLS) correlation results and figure for the publication : Thomasson, M., et al. (2023) Markers of limbic system damage following SARS-CoV-2 infection

The code includes two parts: the first part runs in Matlab (folder MATLAB) and is taken from the myPLS toolbox (available here: https://github.com/MIPLabCH/myPLS); the second part (generation of figures) is implemented in Python.

The following softwares need to be installed:

• For the matlab code you (obviously) need MATLAB (originally created on version R2017a).

• For python code, you will need python

  • Recommended to follow the steps from Conda to install python.
  • You may install the environment from the provided file as described here.

Instructions: You may specifiy your parameters in the myPLS_inputs_CSTM.m file, then run the main script by typing myPLS_main in the Matlab command window or launching the myPLS_main.m file.

Expected output: all PLS computations described in the PLS part of the paper are performed and stored in the outputs/PLS_results/Emo-Rerun-wcov3-wmemo1_GRP folder.

Expected run time: around 2 minutes.

Instructions: open and run the script PLSC_figures.ipynb (in the python folder) in the Jupyter Notebook.

Expected output: The notebook loads the PLS results and produces the A and B subplots of figure 4.

Expected run time: around 1 minutes.