Table of contents
1. Load and display data
Observed event values (Event values): The data set contains 2400 events, and each event specifies the fields Type (type), position (position) and latency (delay) of the EEG.event structure.
If it is a .set file type , directly File->Load existing dataset.
Browse (and visualize) EEG channel data : Plot > Channel data (scroll).
On the right side of the plot window is the vertical scale value (and its unit, microvolts) , which indicates the "magnitude" of the vertical scale bar. In this case, the value is 4217 (microvolts). The same value is also shown in the edit box at the bottom right as shown below where we can make changes. The value of the "Scale" edit text box can be modified
by clicking the "-" button repeatedly or by editing the text value via the keyboard , and pressing the Enter key updates the scrolling window. In the eegplot() interface, Settings > Zoom off/on > Zoom on. Then using the mouse, drag a rectangle around the data area to enlarge it. The scrolling window may now look similar to the window below. Click the right mouse button to zoom out again. Setting > Zoom off/on > Zoom off to turn off zoom . Draw grid lines: Display > Grid > X grid on to draw horizontal lines, set Display > Grid > Y grid on to set vertical axis. If you repeat the corresponding operation, you can turn off the corresponding grid line.
2. Draw an EEG scalp map
Load channel location information: Edit>Channel locations
At this point, the loaded channel labels and polar coordinates are displayed in the interface below (pop_chanedit.m window), and it is recommended to use the default settings. [ Electrodes drawn outside the head animator are electrodes that lie below the midline of the head (i.e., negative z (height) coordinates, with 0 being the center of the head. By convention, they are drawn outside the head animator. Want to only To draw a scalp map in a head animation, it is necessary to enter 0.5 in the "Plot radius" edit box. In this case, when calculating an interpolated 2D scalp map for display or 9 in some cases) for further processing , the two eye electrodes will not be displayed or considered . These settings are used for all scalp topographic maps drawn in eeglab. This dialog can also be set to 1.0 to interpolate and display a scalp map that includes all possible scalp channel positions , and the portion of the head that is (0.5) below the head equator, showing 'skirt' outside the cartoon head bounds or in the 'halo' region (more precise control over where the channels are plotted is available from the command line: see the "Help" of the scalp plot plotting function topoplot.m.)]
Visualize channel location: Plot > Channel location > By name
It can also be drawn in digital form (the number corresponds to the channel number): Plot > Channel location > By number
3. Draw the channel spectrum
Plot channel spectra: Plot > Channel spectrum and maps
Set the parameters according to the requirements. The default settings are used here . Click OK to jump to the spectopo.m interface:
This picture is the result of sampling 15% of the data, [15% is set in the red box in the first picture]. It can also be set to 100%, and the effect picture of setting 100% is:
In the figure above, each colored trace represents a spectrum of data channel activity. The scalp map on the far left shows the energy distribution of the scalp at 6 Hz, and these data are centered in the midline of the frontal lobe. Other scalp maps show energy distribution at 10 Hz and 22 Hz.
If you want to see detailed information, you can click each small picture, for example, click the 6 Hz brain map.
You can operate: Plot > Channel properties to draw the scalp position of the selected channel , its activity range and the ERP image of its activity in a single period.
The red part of the above picture is the scalp position of the selected channel, and the above picture also includes the range of activity and the ERP image of its activity in a single period.
4. Preprocessing tools
Re-reference Data: Right Ear Reference AFz Tools > Re-reference
For a given dataset the first time this menu item is invoked, the following window will pop up.We do not want to include this reference channel (neither in the data center nor in the average reference), so do not click the " Add current reference channel in data " checkbox. (Click this checkbox when the record reference is on the scalp).
Filter: 0.1-30HZ Tools > Filter the data > Basic FIR filter, input 0.1(Hz) as the lower edge frequency, 30(Hz) as the upper edge frequency, then click "OK".
After the above steps are finished, a pop_newset.m window will pop up asking for the name of the new dataset. We choose to modify the dataset name and override the parent dataset. Finally click "OK"
Eliminate the DC offset (DC offset) in the signal: On the eegplot interface, click Display > Do remove DC offset
In signal processing, DC offset is the difference between the mean value of a signal and a baseline of zero frequency. DC offsets are usually caused by sensor bias, noise, or other system issues. DC offsets have an impact on signal analysis and interpretation and therefore need to be removed.
A common method of removing the DC offset is by subtracting the average value of the signal so that the average value of the signal is close to zero. In this way, the baseline of the signal can be adjusted to an appropriate position, making subsequent signal processing and analysis more accurate and effective.
In EEG signal processing, removing DC offset is one of the very common preprocessing steps. By removing the DC offset, the signal offset that may interfere with subsequent analysis can be eliminated, the signal quality and reliability can be improved, and we can better study and understand the characteristics and patterns of EEG signals.
5. ICA removes artifacts
ICA removes artifacts: Tools > Run ICA (this process is super slow), the state changes after running.
After running ICA, remove artifacts: Tools > Reject data using ICA > Reject components by map
Click OK.
5. Extract data epoch
(You can run ICA in segments first, or run ICA first and then segments; if the stimulation time is longer than 2/3 seconds, you can segment first and then ICA)
To study event-related EEG dynamics in continuously recorded data, data time periods locked to events of interest (e.g., data time periods locked to a set of experimental stimuli) can be extracted .
In the pop_epoch.m window above, click the upper right button of the red box above, which will invoke a browser box that lists the available event types.
