Why does Matlab eeglab pop_select create a boundary event at the start of the selected segment? Does the boundary event signal a discontinuity in this case?

Hi all,

I am working on discontinuous raw EEG files (.EDF+D) including multiple event types.

I want to select specific segments of the EEG recordings that I know are continuous in order to create multiple continuous files that are ready for filtering (as they do not contain discontinuities anymore).

To custom-select the appropriate data segments, I am using the following:

EEG = pop_select(EEG, 'time', [(EEG.event(16).latency/EEG.srate+0.5) (EEG.event(16).latency/EEG.srate-0.5)] ); % /EEG.srate because my true latency (as displayed on eegplot, in seconds) = EEG.event.latency/srate (with srate = 256 or =500 depending on recordings).

EEG = eeg_checkset(EEG);

eeglab redraw;

EEG = pop_saveset(EEG, 'RKS186_RawData_clip5.set', '/Volumes/Ashur DT2/EEG/RKS186/');

In my newly-created (continuous) data segment (clip5 above), however, I find a boundary event at the start of the segment, at EEG.event.latency = 0.5.

Why does pop_select create a boundary event at the start of the selected segment? I would expect that this segment is continuous, if I selected data that did not include discontinuities.

So does the boundary event signal a discontinuity in this case, or is it just a random marker?

I would like to understand if I can ignore such boundary event when running further processing (e.g. filtering or applying PREP pipeline) or if I need to be aware of any discontinuities at that point.

Many thanks for your help!

Irene

Answer 1

[Vatsal]

Hi Irene,

The “pop_select” function in EEGLAB is used to select specific segments of EEG data. Upon using this function to create a new continuous data segment, it creates a ‘boundary’ event at the start of the selected segment. This is a standard behavior of the function and does not necessarily indicate a discontinuity in the data.

The ‘boundary’ event is essentially a marker indicating the start of the new data segment. It does not signal a discontinuity unless the selected data segment itself includes discontinuities. If selected segment of data is continuous, the ‘boundary’ event at the start of the segment can be considered as a marker and does not represent a discontinuity.

Therefore, if you are certain that the selected data segment is continuous and does not contain any discontinuities, you can proceed with further processing (e.g., filtering or applying the PREP pipeline) without worrying about the ‘boundary’ event.

I hope this helps!

[Irene Faiman]

Thank you very much @Vatsal for the clarification!

Best wishes,

Irene