[Eeglablist] Question about filtering in eeglab

[Andreas Widmann]

Min Bao wrote:
> Dear Andreas,
> 
> I saw your reply on a post in the eeglablist named 'Filtering trouble'.  
> In my instance, the sampling rate is 500 and the number of frames is 
> 1385355. I tried both 0.1Hz and 0.5Hz for the lower edge of my accepted 
> band when doing a highpass filtering. However, the below error messages 
> happened all the time for 0.1Hz and happened still quite frequently for 
> 0.5Hz. Do you have any suggestion to avoid that kind of error? Thank you 
> very much!
Relevant is not the total number of frames, but the number of frames per
*continuous* data epoch/segment (separated by 'boundary' events usually
inserted by the data acquisition system to mark DC offsets). As far as I
remember the EEGLAB built-in "Basic FIR filter" requires the segment to
be at least three times longer than the number of filter coefficients.
You find 'boundary' event indices/latencies fastest from the commandline:

>> boundArray = strmatch('boundary', {EEG.event.type})
>> [EEG.event(boundArray).latency]

You might want to try the "Windowed sinc FIR filter" from the firfilt
plugin which does not have this restriction. Reasonable starting values
for filter order are 8250 for 0.1 Hz highpass (Hamming window; 0.2 Hz
transition band width) or 1650 for 0.5 Hz highpass (Hamming window; 1 Hz
transition band width).

Please cc further questions to eeglablist. Please also report back to
the list how you solved the problem.

Best,
Andreas
-------------- next part --------------
A non-text attachment was scrubbed...
Name: smime.p7s
Type: application/x-pkcs7-signature
Size: 3257 bytes
Desc: S/MIME Cryptographic Signature
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20080616/15bceedb/attachment.bin>