[Eeglablist] a question on averaging in EEGLAB

Alexander Degerman alexander.degerman at helsinki.fi
Sat Oct 8 05:05:03 PDT 2005


I'm a beginner in using eeglab. I've measured EEG data with Biosemi using 
a 128-channel electrode cap. I have an experiment that is divided into 5 
different BDFs, which I intend to merge at a later stage. My question is 
about averaging, though.

For example, let's say I take one of the BDFs, import it, and eventually 
epoch it using an imported event file. My event file is the format:
Latency	Type	Position
9.578125	12	1
9.964844	133	2
10.101563	111	3

The Position in this case is just a runing number, I won't be using it in 
later analysis. I get an epoched dataset, which looks fine and gives me 
all the 1547 Epochs I have set based on my event file. After artifact 
rejection I end up with a data set with 1467 epochs. The number of events 
is far greater than the number of events in my event file, but this number 
probably corresponds to the sum of events in the different epocs. Or does 
it? Well, so far so good, anyway, as far as the epochs are concerned.

Yet, when I try to make a simple averege file using only one trigger code 
(e.g., type 12), I get a data set containing 57 epochs. The way I do this 
I open the epoched data set in the Gui > Edit > Select Epochs/Events > 
type: 12, Remove epochs not referenced by any selected event > ok

It is strange that I get 57 epochs, because there were only 27 type 12 
events in my event file to begin with. If I scroll the averaged data set, 
I can see that there are event types 134, 133, 132, 131, 16, 15, 11, and 1 
in addition to event type 12. I can get rid of the extra events by 
choosing the 'Keep only selected events and remove all other events' 
option when selecting the events, but this will not get rid of the extra 
epochs. There are still 57 epochs instead of under 27. 

What could I be doing wrong? I would appreciate any help on this matter.

Thank you,

Alexande Degerman
Attention and Memory Networks
Department of Psychology
University of Helsinki 

More information about the eeglablist mailing list