PhD thesis - analysis guidelines

Audio analysis

1. Run annotate_export_split_SFS.bat to annotate, extract durations of annotated utterances to a csv file, and split the annotated continuous file into short wav files based on delimeters of annotations.
2. Split wav files will be used in PraatR to extract pitch values: praatrPitchLoop.R
3. Run praatr_analyse.R to summarise and plot acquired values.

Filter raw bdf files:

1. Run gogo_loop.m – use different baseline for self-production (-49 1) than other conditions (50 101). Epoch -49 350 is fine.
2. Matlab script (gogo_loop.m) has to be located in the EEGLab folder - C:\eeglab13_0_1b – EEGLab folders have to be in the Matlab path
3. The files are saved in the L0n folders in the current Matlab folder.

Create Grand Averages:

In R:

1. Use ffr_averages.R

In Matlab:

1. Create new folders with the relevant txt files - use copy_all_type.bat
2. Run makeGA.m (modify the source folder) - type makeGA on the command line

Run FFT

1. Basic version: doFFT.m
2. Loop: doFFT_eryk_loop.m
3. Extracts 10% at the beginning, middle, and end of each file: doFFT_eryk_loop_percentage_bme_onAVG.m

Transform .mat output to .txt files

1. Use mat2txt.m

Plot Grand Averages:

1. Plot the averages from .mat files using load_and_plot_GAs.m

Create plots for individual FFRs:

1. Plot individual files using .avg files (created with Tim’s script) using ffr_plot_ind_loop.R

Create .avg files from .txt (Response in Brainstem Toolbox):

1. Use bt_txt2avg.m in BT2013

bt_txt2avg('filename.txt', 6855, -15.8, 58.89)

where filename.txt is the exported brainstem data file, 6855 is the sampling rate, -15.8 corresponds to the start time (ms) of the response epoch and 58.89 corresponds to the time at the end the response epoch. (In this example, the recording window is 74.69 ms) e.g.

bt_txt2avg('perc-fall.txt', 16384, -49, 350)

Use the script: txt2avg_script.m

Create .avg files from .wav (Stimulus in Brainstem Toolbox):

1. BT2013 must be the current Matlab folder
2. Use the command

wav2avg('xxxxxxx_self_01.wav', 16384)

The sampling rate of stimulus and response avg file has to be the same.

Use Butterworth band pass filter on the wav files and save the filtered data to an avg file:

1. Run the script wavButter.m

Create .wav files from .txt (FFRs):

1. In R:

library(seewave)

savewav(FFR_vector, f = 16384)

Prepare txt files to be loaded into Praat:

1. Find the output of gogo_loop.m
2. Use txt2praat.R

Brainstem Toolbox:

Download BT

1. Start the GUI by typing bt_ptgui in Matlab
2. Load the response (avg), stimulus (avg), and adjust the settings.
3. Excel output is saved in /output_files

Cortical analysis

1. Make ERPs (without ICA) using makeERPs.m
2. Run ICA on filtered/epoched data using ICA_loop_catch.m
3. Average ERPs and append them
4. Channel order for plotting ERPs: 1 5 6 2 7 8 3 4 9

fNIRS analysis

1. Select software OpenGL rendering in Matlab (to enable plotting in Homer2):

matlab -softwareopengl