fNIRS validation study

This repository contains files for producing the results found in: https://doi.org/10.1016/j.nicl.2024.103637

The data repository for this study can be found in: https://doi.org/10.48723/vscr-eq07. Details on how this dataset was prepared can be found in: https://github.com/alkvi/fnirs_dataset_preparation.

Data analysis

Analysis is performed on the data found in the data repository.

These scripts assume the following directory structure:

fnirs_validation_study
│-- Protocol_1_analysis.m
│-- ...
Data
│-- (fNIRS data etc)
│-- temp_data
│-- mat_files

The temp_data and mat_files folders have to be created. Note that the BIDS dataset comes zipped and needs to be unzipped before analysis.

Demographics and gait variables

Important demographics are created and then gait variables used in the mixed models are extracted from the full dataset.

python summarize_demographics.py
python prepare_gait_params_for_mixed_model.py

Signal quality

To run the quality scripts, run them from the signal quality directory

cd Signal_quality

This calculates SCI per condition and participant, and calculates channels below acceptable threshold (0.7), and bad channels are formatted for analysis.

python signal_quality_complex_walk.py
python prepare_bad_channel_list.py

Peak power is calculated by running power_calculation.m in MATLAB.

1st and 2nd level analysis

These steps are all run in MATLAB.

• Preprocessing and 1st level analysis is performed in Protocol_1_analysis.m
• 2nd level analysis is performed in Protocol_1_group_analysis.m. This is run on each hemoglobin type, specified in the beginning of the script.

Producing figures

• A figure with visualized T statistics for each hypothesis is created with the R markdown file summarize_results.Rmd.
• To prepare 1st level ROI beta variables for visualization, the file prepare_betas_interaction_plot.m is run.
• A figure with 1st level beta variables plotted against gait variables is created with plot_interactions.Rmd.