[Eeglablist] help with events

Ilana Hairston hairston at berkeley.edu
Tue Mar 28 17:31:57 PST 2006


Hi -
I need advice.

I acquire data using compumedics PSG software. The basic experiment  
is measuring auditory ERPs during sleep. I have 2 types of events -  
pairs of clicks and and an oddball pair. The subject is presented  
with these tones for about 11 minutes (approx. 1300 presentations).  
The events are recorded in the PSG software in an event channel which  
records a voltage change each event. The difference between the  
oddball and the standard pairs of clicks is the duration of the  
voltage change.
The event line looks something like this:

______|    |_____|   |_________________|         |_____|         | 
_________etc.

The EEG data is exported as an edf file, with the event line into a  
separate file.

The first hitch is that the event line is acquired at a sampling rate  
of 64 Hz (the highest possible) while the the EEG is at 128. I wrote  
a script that up-samples the event line and converts it into a text  
file that EEGLAB can read, basically adding the average between every  
2 points. However, it doesn't look as though the events match up. Its  
hard to gage, since if i import the event line (as edf) into EEGLAB,  
it wont upsample to 128, but i would imagine that the time stamp  
would take care of the discrepancy; still the event marks from the  
text file don't match up in a consistent manner with the imported  
event channel. In addition, the ERPs look really noisy, which  
suggests to me that they are not aligned porperly.

So my first question is what happens when i import into EEGLAB (using  
the edf import tool) data that contains different sampling rates?
Second, can i somehow use the raw event line to mark events in EEGLAB?

I'm open to any suggestion.

ilana







More information about the eeglablist mailing list