[Eeglablist] calculation of spectral Power

[Arnaud Delorme]

Dear Don,

yes, if you rescale the output of the spectopo function to 10.^(x/10), then it should be identical to the output of the pwelch Matlab function (given the same set of parameters as Makoto pointed out). Note that this is only for continuous data that does not contain boundary events.

Thanks,

Arno

On 18 Oct 2012, at 09:07, Makoto Miyakoshi wrote:

> Dear Matthew and Arno,
> 
> At least for spectopo.m, there are parameters like nfft and winsize
> that changes results a lot. I want to know what values you used for
> the comparison test.
> 
> Makoto
> 
> 2012/10/17 Don Scott <dscott8201 at gmail.com>:
>> Hello,
>> 
>> I have a question about how channel spectra are calculated.  The Plot menu
>> calls pop_prop() which, in turn, appears to call (line 215) Spoectopo.m for
>> the actual calculations using Welch's method (line 277) .  Then in line 306+
>> the calculations occur, maybe rescaled by some kind of RMS value?   the plot
>> eventually displayed shows the ordinate label as "Power 10*log10(uv^2/Hz)"
>> 
>> I have tried to duplicate the EEGLAB results using the Matlab Signal
>> Processing toolbox but cannot get the same numbers given by EEGLAB.
>> Rescaling the sptool results by 10*log10(uv^2)/Hz) does not seem to help
>> either.
>> 
>> Can someone please help me clear this up, or point out what I am missing?  I
>> need to know which is the correct set of data.
>> 
>> Thank you
>> Don
>> 
>> 
>> --
>> Don Scott
>> 
>> (864) 415 - 1795
>> Dscott8201 at Gmail.Com
>> 
>> "Although the universe doesn't have an end, it had a beginning in the Big
>> Bang. One might ask what is before that, but the answer is that there is
>> nowhere before the Big Bang, just as there is nowhere south of the South
>> Pole."
>> 
>> --- Stephen Hawking
>> 
>> 
>> _______________________________________________
>> 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
> 
> 
> 
> -- 
> Makoto Miyakoshi
> JSPS Postdoctral Fellow for Research Abroad
> Swartz Center for Computational Neuroscience
> Institute for Neural Computation, University of California San Diego