[Eeglablist] fft filter

Kevin Guise guisek1 at mail.montclair.edu
Mon Dec 27 09:30:50 PST 2004


I am attempting to use a fft/ifft filter implemented in eeg lab on a fairly large dataset (about 1600000 data points and 32 channels), and i consistently get out of memory error messages.  To remedy this without running out to the store and buying more memory, I tried to perform a manual fft/ifft filter on individual channels.  For some reason, the resulting fourier transform contains all Inf+/-Inf(i) elements, and the ifft all zeros.  Might this be due to the large size of the dataset?  I have done this before with other datasets but never had this problem.  I believe that EEG lab uses the same method, so I don't expect that more memory would be an immediate help.

Any suggestions?
Kevin Guise
Cognitive Neuroimaging Lab
Montclair State University




More information about the eeglablist mailing list