[Eeglablist] GUI vs command line
Brian Roach
Brian.Roach at yale.edu
Fri Oct 28 15:30:57 PDT 2005
I am trying to automate some processing of .cnt data files in EEGlab. I am
unable to use the EEG.history commands on multiple files. I am also
completely unable to use the epoch function at all. I can do everything
from the GUI, but something must be lost when I move to command line
function calls. I downloaded Matt Stead's updated loadcnt.m file into the
proper directory, but this did not help. Has anyone had similar
difficulties with .cnt data/epoching? Can anyone tell me (or show example
of) what the input format should be for the 'events' ( - vector events
(expressed in seconds or points) ). I have no idea about the proper input
for this argument because I have tried things like:
>> epocheddata = epoch(EEG, [3641], [-0.5 1]);
Epoching...
Warning: event 1 out of data boundary
or
>> epocheddata = epoch(EEG, [1:100000000], [-0.5 1]);
this produces 100000000 out of boundary warnings
Also, EEG is assigned by a seemingly successful call to:
EEG = loadcnt('myfile.cnt' , 'dataformat', 'int16');
where
EEG =
header: [1x1 struct]
electloc: [1x63 struct]
data: [63x111280 double]
Teeg: [1x1 struct]
event: [1x101 struct]
tag: 0
also, of all the fields in EEG.event:
ans =
1x101 struct array with fields:
stimtype
keyboard
keypad_accept
offset
type
code
latency
epochevent
accept
accuracy
only stimtype and offset have non zero values - is that a problem for epoch
or pop_epoch? When I load a file from the GUI, these extra messages are sent:
eeg_checkset note: upper time limit (xmax) adjusted so (xmax-xmin)*srate+1
= number of frames
eeg_checkset note: creating the original event table (EEG.urevent)
Creating a new dataset with index 1
Done.
In this EEG structure, EEG.event holds: 1x101 struct array with fields:
type
latency
urevent
stimtype = type, but the associated latency values, event(1).latency might
equal 3608, and event(1).offset (when the command line call is used) =
3607. Can I just set event(1).latency = event(1).offset for all events,
and urevent = 1 for all events?
thanks,
Brian
More information about the eeglablist
mailing list