[Eeglablist] ERSP calculation with no baseline

[Arnaud Delorme]

Dear Caroline,

> 1. Question: Is the the timef() ,as I used it, correct?

Yes, this is a correct usage. You should use newtimef though which is more computationally efficient.

> 2. Question: Normally, the output ERSP values are log spectral diffs. from baseline, but what do they describe, if no baseline is available (baseline set NaN)? Do they then refelect Event related spektral power after the pulse (in dB?)?

Yes, this reflect the spectrogram. The unit it 10*log(power)/Hz unless you ask otherwise.

> 3. Question: I set cycles number to 0.25 (that I earn frequency range 1-4.5 with 0.25 Hz bins), can I use that?

It is better to start with numbers above 1 for cycles, otherwise you frequency resolution will be very poor.

> 4. Question: Can I assume, when the ersp_c1(i).ersp differ from the ersp_c2(i).ersp in any timepoint, that the EMF pulse leads to a change in spektral power (after the pulse) for this timepoint(s)?

Yes, this is a reasonable assumption if the difference proves to be significant. You can test this hypothesis when using the newtimef function.
Best,

Arno



-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20110728/b89e20a0/attachment.html>