This repository is for the EEGLAB sessions of the practical MEEG 2022 workshop. There are 5 sessions:
- Preprocessing
- Single sensor analysis (ERP/ERF)
- Single and distributed sources
- Time-frequency domain
- Group-level analysis
For each session, we have prepared a script detailed below.
We will use data from the multimodal face recognition dat. BIDS dataset containing a pruned version of the OpenNeuro dataset ds000117. It is available here.
The dataset above only contains one subject. For group level analysis, please use the following BIDS repository here.
The scripts using the single subject data assume the datafiles are located in the folder (Data/sub-01) located in the parent folder of this repository in your file system. See below the code used in the scripts to locate the file:
RootFolder = fileparts(pwd); % Getting root folder
path2data = fullfile(RootFolder,'Data', 'sub-01'); % Path to data
For Session 5, copy the data folder (please rename to 'ds002718') containing the ds002718 in the same 'Data' folder. These files will be distributed later on.
For this presentation, we will first import the data with the PracticalMEEG_Import_Data_Session_1.m script. This script has 11 steps.
- Step 1: Importing MEG data files with FileIO
- Step 2: Adding fiducials and rotating montage
- Step 3: Recomputing head center (for display only)
- Step 4: Re-import events from STI101 channel (the original ones are incorect)
- Step 5: Selecting EEG or MEG data
- Step 6: Cleaning artefactual events (keep only valid event codes)
- Step 7: Fix button press info
- Step 8: Renaming button press events
- Step 9: Correcting event latencies (events have a shift of 34 ms as per the authors)
- Step 10: Replacing original imported channels
- Step 11: Creating folder to save data if does not exist yet
After importing the data, it is preprocessed using the PracticalMEEG_Preprocess_Data_Session_1.m script. This script itself has several steps.
- Re-Reference the data
- Resampling the data (for speed)
- Filter the data
- Automatic rejection of bad channels
- Re-Reference again
- Repair bursts and reject bad portions of data
- run ICA to detect brain and artifactual components
- automatically classify Independent Components using IC Label
- Save dataset
For this presentation, we will use different vizualization techniques using the PracticalMEEG_ERP_Analysis_Session_2.m script. The script first further process the data as follow.
- Extract data epochs for the famous, scrambled, and unfamiliar face stimuli
- Remove the baseline from -1000 ms to 0 pre-stimulus
- Apply a threshold methods to remove spurious epochs
- Resave the data
Then it plots the data using the following methods:
- Plot ERP butterfly plot and scalp distribution at different latencies
- Plot ICA component contribution to the ERP
- Remove ICA artifactual components and replot
- Plot series of scalp topography at different latencies
- Plot conditions overlaid on each other
- Plot ERPimages
For this presentation, we will the script PracticalMEEG_Source_Reconstruction_Session_4.m. It performs the following steps.
- Definition of head model and source model
- Localization of ICA components
- Plotting of ICA components overlaid on 3-D template MRI
For this presentation, we will the script PracticalMEEG_Time_Frequency_Analysis_Session_3.m. It performs the following steps.
- Spectral analysis for each of the conditions
- Time-frequency analysis for each of the conditions
The script PracticalMEEG_ERP_Analysis_GroupAnalysis_support.m perform group analysis on a group of subjects.
- Removing components flagged for rejection using ICLabel
- Plotting grand average ERPs