Home

Welcome to the ARC EEG resting-state wiki!

What is this wiki for

This wiki is mainly intended to provide some minimal guidance on how to use the scripts provided in this section of my github. These scripts are to be used for analysis of resting state EEG data that was recorded on a 64-channel BioSemi ActiveTwo system with external electrode 6 (placed on the mastoid) as a reference electrode. It assumes trigger codes are used for eyes-closed and eyes-open segments and presently only uses eyes-closed segments. The currently assumed pre and post-processing steps are as follows:

1. Trials and triggers codes are read and only eyes-closed segments are selected
2. Very basic pre-processing includes: demeaning, detrending, low-pass filter below 60Hz, re-referencing to mastoid reference, bandstop filtering of possible 50Hz line-noise and resampling to 1024Hz.
3. Optional visual inspection after basic pre-processing
4. Optional reconstruction of noisy channels (thought only use this if a channel is absolutely rubbish!)
5. Segmenting continuous recording into 4-second segments to be used for WPLI analyses later on
6. Removal of noise using ICA decomposition and visual inspection of resulting components
7. Resetting the trial timestamp to allow fieldtrip to consider them separate 'trials'.
8. Calculate the weighted phase lag index using wavelet decomposition for each of the 5 frequency bands separately
9. Calculate various graph metrics from the WPLI adjacency matrices

What scripts do what

There are 3 main scripts to get all of this done:

1. The pre-processing script found here takes care of steps 1-7 This script is intended to be run manually for each individual subject. At the start it only requires the filename without an extension. Then each block of the script can be run manually. The reason for having it done manually are two-fold. First it allows the users to inspect the data at at every step so that you can for example check if the triggers are read correctly, if there are specific channels that are noisy, if there are specific 'trials' that are noisy.
2. The WPLI script found here takes care of step 8
3. The WPLI script found here takes care of step 9