Now let’s take a look at the averaged ERP data from each participant. Select ERPset 1 from the ERPsets menu so that we can examine the data from Subject 1. Go to the Matlab command window and type ERP (followed by Return or Enter). You should see something like the information shown in Screenshot 3.2.
Screenshot 3.2
ERP is the name of a Matlab variable that stores the current ERPset (the one you selected in the ERPsets menu). When you typed the name of it in the command window, Matlab printed out the contents of the variable. It’s a complicated variable with many different fields, including the erpname field (the name of the ERPset), the nchan field (which stores the number of channels), the chanlocs field (which stores the names and 3-D locations of the electrodes), and the bindata field (which stores the actual ERP data, in binary format). There is also an EEG variable that stores the current EEG dataset (if one is loaded).
The available variables can be seen in the Workspace pane of the Matlab GUI. You can also see the contents of a variable like ERP or EEG by double-clicking the name of the variable in the Workspace pane. This causes the variable to be shown in a separate window. You can then double-click on the fields of the ERP variable to see those fields in more detail. For example, try double-clicking the chanlocs field to see what information it holds.
Now type ERP.ntrials in the command window. Matlab will print out information about the number of accepted trials, the number of rejected trials, and the number of invalid trials for each bin. You should see that there were 47 accepted trials and 13 rejected trials for Bin 1. You can also get a slightly nicer table with the same information by selecting EEGLAB > ERPLAB > Summarize artifact detection > Summarize ERP artifacts in a table. You’ll want to know this information before looking at the ERP waveforms for a given participant. For example, you’ll want to know if there were any bins without a reasonable number of accepted trials.
Now plot the ERP waveforms with EEGLAB > ERPLAB > Plot ERP > Plot ERP waveforms. You can use the default settings, except just plot Bins 3 and 4 (the related and unrelated targets). Just as in the grand averages shown in Figure 2.1 in the previous chapter, you should see that the unrelated targets elicited a more negative voltage around 400 ms than the related targets. However, this participant’s waveforms are a bit noisier than those of those of the participant we looked at in Chapter 2 (Subject 6). If you don’t remember what Subject 6’s waveforms looked like, you can select that participant’s ERPset in the ERPsets menu and plot the waveforms.
To look at the data quality (the analytic standardized measurement error or aSME), select EEGLAB > ERPLAB > Data Quality options > Show Data Quality measures in a table.
Now repeat this sequence of steps to look at the data from all 10 participants. Here are some questions you should answer:
Do you see the N400 effect (more negative for unrelated than related targets) in each participant’s waveforms?
Are there any participants who have an unusually small or large numbers of trials rejected?
Are there any participants with appreciably worse aSME values?
These 10 participants were chosen because they all have pretty good data. They all have an N400 effect, and they all have pretty similar levels of data quality. If we looked at all 40 participants in the full study, we’d see a wider range of effects, numbers of rejected trials, and data quality.
