[Eeglablist] wavletcycles fastor

[arno]

Elena Orekhova wrote:
>
> Dear EEGLab users
>
>  
>
> Does anybody know the reason to use particular 'wavletcycles fastor' 
> in timef function?
>
> Certainly, the picture looks better with the default 0.5 factor, than 
> with both 1 or 0 factors, but can it be misleading? Intuitively, 
> factor 1 (no increase standard wavelets) seems to be more precise, is 
> not it?
>
Yes, "wavelet factor 1" has a better time resolution at all frequencies. 
We must use a low number of cycles at low frequencies (for example 3 for 
5-Hertz theta) otherwise the window size would be too big (3 cycles at 5 
Hertz is already 600 ms). However, at high frequency this is not a 
problem (3 cycles at 40 Hertz is 7.5 ms) and it is better to obtain a 
better frequency estimate by increasing the number of cycles (using the 
wavelet cycle factor), which in turn decreases time resolution. If you 
do not do that, some "blurring" effect appear at high frequency. 0.5 
means that the program will use a window size of half the size at the 
lowest frequency. It is not ideal as a parameter. A new version of the 
function we have designed (and will be releasing soon) allow to set 
directly the number of cycle at the lowest frequency and at the number 
of cycle at the highest plotted frequency.

Best,

Arno

> Elena
>
>  
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> eeglablist mailing list eeglablist at sccn.ucsd.edu
> Eeglablist page: http://sccn.ucsd.edu/eeglab/eeglabmail.html
> To unsubscribe, send an empty email to eeglablist-unsubscribe at sccn.ucsd.edu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20060419/319ca481/attachment.html>