Scalp maps (topoplot)

Why do I need this step?

Each ERP component has a roughly known distribution. So a map helps you tell apart two components that occur at the same time, and spot a bad channel that shows up as a lone, odd-looking spot.

Maps are also easy to misread: a different color scale on two maps can create a difference that does not exist. This lesson teaches plotting and reading together.

How do I do it in EEGLAB?

Start with the dataset D1 epochs clean open, as in “Starting point”. It has 32 channels with locations, epoched around the square event.

0. Keep the scalp channels only

EOG1 and EOG2 were left out of the average reference in the rereferencing lesson, so they are still on the original reference. They also have locations in D1, so EEGLAB would draw them on the map and could take its color limits from them. Make a copy without them first.

1. Open Edit › then Select data.
2. In the Channel(s) row, type EOG1 EOG2, and tick the box in the on->remove these column of that row. Press Ok.
3. In the window that follows, name the new dataset D1 epochs clean scalp and do not save it to disk. It has 30 channels, and all of them use the average reference.

Use this dataset for every map in this lesson.

1. The waveforms and their maps together

1. Open Plot › then Channel ERPs › then With scalp maps.
2. In Plotting time range (ms): type -200 800.
3. Leave Scalp map latencies (ms, NaN -> max-RMS) at NaN. EEGLAB then picks the time at which the ERP is strongest across channels.
4. Leave Window size at latencies in ms at 0. Press Ok.
5. The waveforms of all 30 channels appear on top of each other, with a map above them at the chosen time and a line linking it to the time axis. Click any point on the waveforms to draw the map at that time.
6. Repeat the steps, this time typing 100 200 400 in the latency box to see three maps above the waveforms.

2. A map series on one scale

1. Open Plot › then ERP map series › then In 2-D.
2. In Plotting ERP scalp maps at these latencies type 0:100:500. The box accepts any MATLAB expression, so this means 0, 100, 200, 300, 400 and 500 ms.
3. In Plot title keep the dataset name.
4. In Plot geometry (rows,col.); [] -> near square type [2 3].
5. Leave the extra options box at 'electrodes', 'on' so the electrode dots are shown. Press Ok.
6. Six maps appear on one color scale with one color bar. Click any map to enlarge it in its own window.

3. Read the map

How do I write it as code?

EEG = pop_loadset('D1_epochs_clean.set', pwd);

% Scalp channels only: the EOG channels are not on the average reference
EEG = pop_select(EEG, 'rmchannel', {'EOG1', 'EOG2'});
EEG.setname = 'D1 epochs clean scalp';

% All-channel ERP with a scalp map at the latency of maximum power
figure; pop_timtopo(EEG, [-200 800], NaN, 'D1 epochs clean scalp');

% ERP with scalp maps at chosen latencies
figure; pop_timtopo(EEG, [-200 800], [100 200 400], 'D1 epochs clean scalp');

% Map series every 100 ms, one shared color scale, 2 x 3 grid
pop_topoplot(EEG, 1, 0:100:500, 'D1 epochs clean scalp', [2 3], 0, 'electrodes', 'on');

% Mean map over a 300 to 500 ms window
erp = mean(EEG.data, 3);
win = EEG.times >= 300 & EEG.times <= 500;
winMap = mean(erp(:, win), 2);
lim = max(abs(winMap));
figure;
topoplot(winMap, EEG.chanlocs, 'maplimits', [-lim lim], 'electrodes', 'on');
cb = colorbar; ylabel(cb, '\muV');
title('300 to 500 ms');

Which value should I choose?

How do I check it worked?

• Nose up, with Fz at the front and Oz at the back when you use 'electrodes', 'labels'.
• One color bar for the series of six maps, with limits of equal size.
• No electrode dots below the eyes: the map has 30 dots, and EEG.nbchan is 30.
• The 0 ms map has calm colors, usually close to the zero color, because the -200 to 0 ms mean was set to zero by baseline correction and the visual response has not started yet.
• The window map changes when you move the window from 300–500 ms to 100–200 ms, and its limits change only when the amplitude changes.

What mistakes do people make?

• Comparing maps on different scales. Read the numbers on the color bar before the colors. Two maps with similar colors can differ tenfold in amplitude.
• Trusting the Window size at latencies in ms box. The value is half the window width, so 50 at 100 ms means 50 to 150 ms. In EEGLAB 2026.1 the first maps carry that “50 to 150 ms” label, but they are drawn from a single sample. The real window average only happens when you click the waveforms. For a reliable window map, use the script above.
• Reading the map as the source location. A red spot over Pz does not mean the source is under Pz. Conduction through the skull spreads the voltage, and the source orientation moves the peak. Source localization is the topic of the DIPFIT lesson.
• A map without correct electrode locations. If locations are missing or wrong, the whole map is wrong. Check them as in the channel locations lesson.
• Forgetting the reference. The map changes completely with the reference. Our maps use the average reference of 30 scalp channels, so say so under every figure.

How do I write it in my paper?

A template: edit the data, settings and versions to match what you ran.

Scalp topographies of the ERP at 30 average-referenced scalp channels (EOG channels excluded) were plotted at 0 to 500 ms in 100-ms steps and as the mean amplitude from 300 to 500 ms, using a symmetric color scale (pop_topoplot, topoplot, EEGLAB 2026.1).

Going further

• EEGLAB Tutorial: https://eeglab.org/tutorials/08_Plot_data/Data_Averaging.html (8. Plot data > a. ERPs: Plotting all-channel ERPs; Plotting a series of 2-D ERP scalp maps)
• EEGLAB Tutorial: https://eeglab.org/tutorials/11_Scripting/EEG_scalp_measures.html (11. Write scripts > d. Advanced plots)