4. 🧠 Source‐Based Analysis Reference - jasonnan2/Automated-Analysis-of-EEG GitHub Wiki
This section outlines the standard analysis workflow using the SourceObject class to process, analyze, and visualize source localized EEG data. Reference script is runSource.m
Example
# load in required variables
% Initialize SourceObject
sourceObject = SourceObject(project.sourceData, project.info, baselineTime, timeRange, cfg);
% Clean missing subject data
sourceObject = sourceObject.cleanDatasets();
% Preprocess: outlier removal, baseline correction
sourceObject = sourceObject.standardProcessing();
% Calculate average source activity across ROIs/freqs/time
sourceObject = sourceObject.calRoiData();
% Compile all significant ROIs into a summary table
sourceObject = sourceObject.calSigTbl();
% Compute average signal per network using predefined netwrk struct
sourceObject = sourceObject.calNetData(netwrk);
% Compute ROI-to-ROI and network-to-network connectivity
sourceObject = sourceObject.calConnectivity(netwrk);
% Run group comparisons with default cfg
sourceObject = sourceObject.analyzeRoi();
% Plot average network-level activity
sourceObject.plotNetwork();
% Plot network connectivity matrices overriding default cfg and set to 0.
sourceObject.plotNetConnectivity('FDRflag', 0);
📚 Initial Set Up and Processing
obj = SourceObject(project.sourceData, project.info, baselineTime, timeRange,cfg)
Initializes a new source analysis object
| Parameter | Description |
|---|---|
sourceData |
Struct with source-level data for each group |
info |
Metadata from project.info |
baselineTime |
1×2 vector (in ms) of baseline correction |
timeRange |
Struct defining labeled time windows |
cfg (optional) |
Configuration struct for analysis/plotting |
obj.cleanDatasets()
Removes subjects missing across any group or variable.
obj.standardProcessing(...)
Performs outlier removal and optional baseline correction.
| Parameter | Description | Default |
|---|---|---|
'outlier' |
'5SD' or '3MAD' method for subject-level outlier removal |
'5SD' |
'calBaseline' |
Whether to apply baseline correction | 1 |
obj.calRoiData()
Computes average roi-level activity across frequencies, time windows, and groups and stores in obj.sourceResults.roiData
| Parameter | Description |
|---|---|
| None | No Parameters |
Defining netwrk for network based analysis
Each network should be defined as a struct with the following fields:
| Field | Description |
|---|---|
name |
String label of the network (e.g., 'FPN', 'DMN') |
roi |
Vector of integers representing ROI indices assigned to that network |
fpn = [5 6 55 66 59 60];
netwrk(1).name = 'FPN';
netwrk(1).roi = fpn;
con = [3 4 19 20 37 38 39 40 41 42 57 58 67 68];
netwrk(2).name = 'CON';
netwrk(2).roi = con;
obj.calNetData(netwrk)
Computes network-level averages and stores in sourceResults.netData.
| Parameter | Description |
|---|---|
netwrk |
Struct array defining brain networks |
obj.calConnectivity(netwrk, ...)
Computes ROI-to-ROI and network-to-network connectivity matrices.
-
'roi'(default):
Computes ROI-to-ROI connectivity, applies Fisher Z-transform, averages within each network, then converts back to connectivity. -
'network':
Averages ROI time series within each network first, then computes network-to-network connectivity directly.
Intra-network connectivity is always computed using the ROI-level method.
| Parameter | Description |
|---|---|
netwrk |
Struct of networks as above |
method (Optional) |
'net' or 'roi'. 'net' is default. |
Main Analysis Section
⚠️ If no name-value pairs are specified, values in cfg are used by default. Name-value pairs override
cfg.
💻 obj.analyzeRoi(...)
Performs group comparisons and identifies significant ROIs.
| Parameter | Description |
|---|---|
'hmFile' |
Path to head model file (default used if empty) |
'rois2plot' |
Numeric array of ROI indices to include |
'vars2plot' |
Cell array of variable names to include |
'freq2plot' |
Frequencies to analyze |
'times2plot' |
Time windows to compare (e.g., 'time1') |
'combinations' |
Nx2 array of group indices to compare (e.g., [1 2]) |
'FDRflag' |
1 = apply FDR correction; 0 = uncorrected |
'toPlot' |
1 = generate brain Maps |
'isnormal' |
'auto', 1 (normal), or 0 (non-parametric) |
Populates
sourceResults.sigROIsandsigROIsP.
obj.calSigTbl(...)
Extracts significantly different neural measures and saves them into a table for each variable and time window. Must be run after obj.analyzeRoi().
| Parameter | Description |
|---|---|
chanTypes (optional) |
A string specifying which ROIs to include in the table: • 'sig' (default) extracts only significant ROIs• 'all' includes all ROIs |
Output: Populates
obj.results.sigValues.(variable).(time)with the resulting tables.
📊 obj.plotNetwork(...)
Plots network based grouped bar charts for selected variables, frequencies, and time ranges.
| Parameter | Description |
|---|---|
'vars2plot' |
Cell array of variable names to include |
'freq2plot' |
Frequencies to analyze |
'times2plot' |
Time windows to compare (e.g., 'time1') |
'groups2plot ' |
1xN array of group indices to compare (e.g., [1 2]) |
'FDRflag' |
1 = apply FDR correction; 0 = uncorrected |
'isnormal' |
'auto', 1 (normal), or 0 (non-parametric) |
'color_list' |
Cell array indicating custom colors for each group |
obj.plotNetConnectivity(...)
Plots connectivity heatmaps marked with significance. Plots include individual groups and group difference comparisons.
| Parameter | Description |
|---|---|
'vars2plot' |
Cell array of variable names to include |
'freq2plot' |
Frequencies to analyze |
'times2plot' |
Time windows to compare (e.g., 'time1') |
'combinations' |
Nx2 array of group indices to compare (e.g., [1 2]) |
'FDRflag' |
1 = apply FDR correction; 0 = uncorrected |
'isnormal' |
'auto', 1 (normal), or 0 (non-parametric) |
Must run
calConnectivity()first.