[Eeglablist] pop_eegfilt

[Andreas Widmann]

Hi,

> Thanks much for your reply.  That comes with an external plugin, right?  I found a link, but it's not available anymore.  Where did you find the link to download the plugin?
The firfilt-plugin (providing pop_firws) is still available and under active development. Which link is unavailable (all links I'm aware should be working)?
Homepage: http://www.uni-leipzig.de/~biocog/content/widmann/eeglab-plugins/
Direct download (v1.5): https://github.com/downloads/widmann/firfilt/firfilt1.5.zip

> Also, do you by any chance know how the pop_firws is different from the regular pop_eegfilt? 
pop_firws "designs" filters, pop_eegfilt "fits" filters using MATLAB provided functions. Designing is much more powerful, flexible and precise imho, fitting sometimes gives wrong result (infamous bandpass filter bug; more details here: https://sccn.ucsd.edu/eeglab/bugzilla/show_bug.cgi?id=631). Designing filters allows precise estimates of required filter lengths and other filter parameters.

> When I used pop_eegfilt( EEG, 0.01,80,[ ],0), it gave me the error that "Maximum variable size allowed by the program is exceeded."
You can not beat physics. Filter length is a linear function of transition bandwidth. Highpass cutoff of 0.01Hz allows a max transition bandwidth of 0.02 Hz. Assuming 500 Hz sampling rate and hamming window this will result in a 82501 point filter which is unreasonably long (computation and memory intensive; carrying artifacts through the data for almost three minutes recording time). Using 0.5 Hz (or min 0.1 Hz) cutoff as suggested by Arno is reasonable and sufficient for most purposes in EEG analysis.

Best,
Andreas