[Eeglablist] eventfield propagation in pop_epoch

[Marius Klug]

Dear eeglablist,

I have found a strange thing today while I was playing around with epoch
lengths. I have a data set with thousands of events, using only some of
them for epoching. Those have additional eventfields where I stored
information about the participant and the specific trial. When I epoch
time-locked on those events, the information gets passed to all other
events that are present in the epoch, so it's not the epoch itself that
contains the information, but it's all of the events inside that epoch. Now
the strange thing is, when my epoch is so long that the next event with
stored information in the event field is present in the epoch, the events
of this epoch get the info of the next event and not of the event the epoch
has actually been created upon. Testing it with even longer epochs I found
that it is always the info of the last event of an epoch having event
fields that is stored in all events of the epoch. Now obviously this is
counterintuitive and I was just checking the code but didn't find any
direct hints where this all happens. I suspect it is actually later, in
eeg_checkset around the line 451.

I hope my explanations are not too confused and you can follow my thoughts!
If so, can anyone kindly point me to how I can fix this? I work with eeglab
v14.1.0. I will also file a bug report, but I need to fix this asap for my
data sets myself.

Regards,
Marius

PS.: While we're at it, it is a rather strange thing that the epoch
information is not stored as info of the epoch fields itself, but that the
epoch fields do have cell arrays that are as long as the number of events
in the epoch. It's all just the same value and just takes up space...
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20180123/17faeb77/attachment.html>