[Eeglablist] Fourier transform on small segments

Makoto Miyakoshi mmiyakoshi at ucsd.edu
Wed Jan 6 12:46:52 PST 2021

Dear Eric,

Happy new year!

The window length is determined by the lowest frequency of interest for you.
If your data is sampled at 1000Hz, 1000 samples means 1000 ms. The first
100 samples is 0.1 second, which is 10 Hz. If you use window length of 100
ms, you can obtain 10 Hz data as the lowest frequency. If you are only
interested in gamma range (40 Hz) for example, the shortest window you can
use is basically 1000/40 = 25 ms.

Overlap of 50% is a standard procedure. There may be a mathematical proof
for this which I have not seen. I heard my mentor (Scott) recommend this
value (though EEGLAB's default is overlap of 0%)


On Wed, Jan 6, 2021 at 9:43 AM Eric HG <erichg2013 at gmail.com> wrote:

> Hi Eeglablist,
> We have an epoch of 1 second with 1000 samples and are trying to use
> fourier transform on small segments like the first 100 samples (0.1
> second).
> When doing this, how would you set the window length and overlap if you
> want to calculate the spectral power for each 100 point segment
> independently?
> Thanks a lot in advance!
> Best regards,
> Eric
> _______________________________________________
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to
> eeglablist-unsubscribe at sccn.ucsd.edu
> For digest mode, send an email with the subject "set digest mime" to
> eeglablist-request at sccn.ucsd.edu

More information about the eeglablist mailing list