Opening the book…
During the EEG preprocessing steps, the low-frequency voltage drifts in the continuous EEG data were reduced by applying a high-pass filter. However, there is also some high-frequency noise (small but rapid deflections in the waveforms). You can see this at the most anterior scalp sites in the grand average waveforms (especially if you specify a zoom factor of 200 using the zoom controls in the bottom left corner of the eStudio window). The waveforms just look a little fuzzy.
My lab takes great pains to reduce the major sources of high-frequency noise in the recording environment (e.g., induced voltages from electrical devices in the environment and muscle activity). As a result, there isn’t a lot of high-frequency noise in the grand averages. But there’s a little, and you’ll see much more of it in other experiments (especially in the single-participant waveforms). The present exercise will show you how to filter out this high-frequency noise using a low-pass filter. We’ll apply it to the grand average ERP waveform, but you could instead apply it to the single-subject ERPs, the epoched EEG data, or even the continuous EEG data (see Chapter 7 in Luck, 2014 for information about when different filters should be applied).
Make sure that the Grand_N400 ERPset is still loaded and selected in the ERPsets panel, and then go to the Filtering panel. Set it up as shown in Screenshot 3.6, which should mainly involve setting the low-pass cutoff to 20 Hz. Then click Run to apply the filter. You can name the new ERPset Grand_N400_filt.

Take a look at the filtered data. The first thing you'll notice is that the standard error is no longer shown. And if you look in the Plot Settings panel, you'll see that the option for plotting the standard error is grayed out. When you filter the data, the original standard error values are no longer valid—they’re the standard error of the unfiltered mean voltage at each time point, not the standard error of the filtered values. Similarly, the original data quality metrics are no longer valid, so they're no longer available in the View Data Quality Metrics panel. In a moment, we'll see how to get standard errors for filtered waveforms.
Now compare the filtered waveforms to the original unfiltered waveforms in Grand_N400. The filtered waveforms look a little smoother than the unfiltered waveforms (but the difference is small because we had such a small amount of high-frequency noise in the original data). In Chapter 4, we'll take a closer look at how filtering works (and how it can distort your data if you're not careful).
Filtering the Single-Participant ERPsets
To obtain valid standard error values for a filtered grand average, you need to filter the single-participant ERPsets before making the grand average. Fortunately, this is quick and easy in ERPLAB Studio.
Go to the ERPsets panel and select the ERPsets for the 10 individual participants (leaving out the grand average ERPsets). Now go to the Filtering panel (which should still be set up as shown in Screenshot 3.6) and click Run.
You're going to create a new filtered ERPset for each of the 10 selected ERPsets, so the window for naming the new ERPsets has 10 ERP names in it. ERPLAB has simply added a suffix of _filt onto each of the original ERPset names. Let's try changing this suffix by changing the suffix in the text box from _filt to _filt20 and clicking the Add suffix to ERPset Names button (as shown in Screenshot 3.7). Click Okay to save the new ERPsets.

You should now see the 10 filtered ERPsets in the ERPsets panel, and they should all be selected. Now go back to the Average Across ERPsets panel and click Run. Name the new ERPset Grand_N400_filt20. Pretty efficient, isn't it?
Finally, select Show standard error in the Plot Settings panel (and click Apply). You should now see the standard error shading around the filtered grand average ERPs.

If you look in the View Data Quality Metrics panel, you'll see that we still don't have data quality metrics for Grand_N400_filt20. This is because filtering the single-participant ERPsets invalidated the data quality metrics for those ERPsets, and the data quality metrics in a grand average are created by combining the data quality metrics from the single-participant ERPsets. This isn't a problem for the standard error values, because the standard error at a single time point in a grand average is not computed from the single-participant standard error values. Instead, it is computed as the standard deviation across the single-subject waveforms divided by the square root of the number of subjects.
To get the data quality metrics for filtered data, you would need to filter the single-trial EEG data for each participant prior to averaging. The data quality metrics would then be calculated when the single-participant averaged ERPs are computed. And they would then be available when the grand average is computed.
This page titled 3.6: Exercise- Low-Pass Filtering was last modified on Thu, 01 Oct 2026 18:52:35 GMT and is shared under a CC BY-NC 4.0 license and was authored, remixed, and/or curated by Steven J Luck directly on the LibreTexts platform.
The book could not be loaded just now.