FAQ: How to synchronize data when I have pauses/recording breaks in the EEG?

[olafdimigen]

Context:
While recording pauses in the recorded eye-tracking data are no problem, EYE-EEG requires the electrophysiological recording to be uninterrupted without any breaks/pauses in the recording. The underlying reason for this is that eye-trackers properly assign a time stamp (in micro-/milliseconds) to each recorded data sample. If you restart the recording after a pause, the timestamp of any new samples will still be on the same timeline. In contrast, commonly used EEG systems simply "count" samples (we therefore have to trust that the inter-sample intervals/sampling rate to be correct). As a side effect, not all EEG systems properly account for recording pauses when the recording is restarted.

Answer:
For the reasons stated above, try avoiding recording pauses in the EEG, which serves as the "master clock" for synchronization. If pauses did happen, additional (programming) effort is necessary at the moment. You can either:

1. Synchronize ET and EEG data separately for each block of data that were recorded in-between pauses. Afterwards, you can again concatenate the synchronized blocks of data into one pseudo-continuous EEG datasets (e.g. to run optimized ICA). This is currently the best solution but requires that you have enough synchronization events near the beginning/end of each block.

2. Some EEG system (and EEGLAB import functions) add a "boundary event" (in EEG.event) with a duration field to the EEG.event structure whenever the recording is paused. Before synchronization, you could fill up the duration of these pauses with dummy data (e.g. zeros) and synchronize the EEG with the ET data afterwards. However, pay extra attention to the synchronization quality feedback to make sure that the timelines are correct. We might add a function to do this to future EYE-EEG versions, but this has not yet been implemented.