Technical aspects of auditory stimulation in EEG studies

In research experiments using electroencephalography (EEG), it is very important to have a precise trigger or marker from the stimulus presentation device to your EEG system. During analysis, these trigger markers will be used as reference point for your analysis, as in the case of event-related potentials. Given that the EEG signal is very accurate in time, the marker also needs to be. At first this may seem an easy task, but that's because in our daily tasks we do not need millisecond accuracy when we press play on a sound software to start the execution of the sound, but there is some delay between the play command and the moment we start enjoying the music.

To highlight the importance of precision in trigger markers for EEG, the image bellow illustrates what happens when you have a variable delay: the average signal of the trials has a lower amplitude a less characterized peak than would be expected in the case that all trails were correctly marked in regard to the stimulus onset. The same effect would happens in cases of variable neural responses to a stimulus/task (Enhanced response inhibition in experienced fencers)

In the case of auditory stimulation, the required precision can be even higher as is the case when studying the Auditory Brainstem Response (see Auditory brainstem response to complex sounds: a tutorial), which is a very fast response, compared to the most common evoked-potentials that start around 100 ms after stimulus presentation. Also, it is important to know the onset of the main sound of interest in the audio file. For example, when using sentence sounds there maybe be a key word that triggers the neural response studied in the middle of the audio file, and this time difference between the sound onset (usually used as trigger marker for the EEG) and the main word should be considered and corrected in the analysis.

Even with all these special requirements, there are several solutions to present triggers with high precision to allow this types of studies:

1) Using a stimulus presentation device with a special sound card prepared for research applications. Devices like, Chronos, from Psychology Software Tools or like AudioFile Stimulus Processor, from Cambridge Research Systems, offer the possibility to properly load and present the sound along with sending triggers with millisecond accuracy. Here is an example of auditory research done using Chronos: Auditory Spatial Perception

2) Using sound detection devices, which detect by themselves the onset of the sound send a trigger marker for the EEG. As examples in these category: StimTrak, from Brain Products; and StimTracker, from Cedrus. These devices have great accuracy to detect the sound onset, and have the advantage of sound threshold adjustments in a way that the trigger really is related to the sound onset (and not the file sound start). On the other hand, these devices usually offer only a 1 bit trigger, in a way that you cannot send different triggers for different sounds of interest. However, combining this trigger with another trigger used only for trial classification (without the time accuracy requirement) can be an easy way to provide a wider triggering combination. In this case, in the processing steps instead of selecting one trigger number as the reference, a logical combination of triggers would be used. Some software already offer a solution to combine and redefine triggers (as Analyzer 2.0), but in most cases the trigger marker file can be correctly edited with Microsoft Excel, or Google Sheets (selecting the timing of the required event and attributing it to the markers that classify the trials, or vice-versa).

3) Creating a trigger in the sound file. This is a less common option, but highly accurate, specially when you want to trigger a part of your sound file, and not the sound file onset. One possibility is to use one os the sound channels to create the trigger, and the other(s) have the sound that will be presented to the subject. This solution requires some engineering skills to split the output of the sound card and connecting one channel directly to the EEG trigger entrance (with the proper voltage specifications) and the other channel(s) to the earphone of the subject. Also, the variability of sound cards, might make this a specific solution for the computer in which it is developed and tested. However, it can be the most precise trigger to a part of the sound file.

So there are a few challenges when using auditory Stimulus and EEG, but there are a lot of opportunities as well.