Evoked fNIRS Signals for Isolated Pain Event Classification in Thermal Quantitative Sensory Testing

1. Preprocessing Contains the data and code used for preprocessing. The raw data is stored in the following structure in "nirsc.mat". - (16,1) cell : 16 subject └ (1,4) struct : 4 task (Cold Threshold Test, Heat Threshold Test, Cold Tolerance Test, Heat Tolerance Test) └ headers, time, task, mark, count, oxyHb, deoxyHb, totalHb, chanlabs, srate nirsc_filtered.mat Data with wavelet-mdl applied using matlab's nirs-spm and a 0.01-0.09 bandpass filter applied using the matlab function. py_MakePythonData.m Generate h5 files by window size by segmentation from mat file. 2. Train Model Contains files for training the model. The files described below are the code you run to train, while the other files are included for dependencies. Needs output files of 'py_MakePythonData.m'. main_Train.py This code runs k-fold and LOSO CV training on the HbO, HbR, and HbO&HbR datasets. wandb-ORC-T-KFOLD_trialwise_valid-Concentration.py This code runs k-fold CV training on the HbO† datasets. wandb-ORC-T-LOSO16_valid-Concentration.py This code runs LOSO CV training on the HbO† datasets. main_Saliency.py This code calculates saliency for trained model parameters. Need the output file of 'main_Train.py', 'wandb-ORC-T-KFOLD_trialwise_valid-Concentration.py', or 'wandb-ORC-T-LOSO16_valid-Concentration.py'. 3. Figure This is the code that creates the figure and supplementary figure. Need the output file of '2. Train Model'.