[Eeglablist] Extract absolute power using ERSP (Thang Le)

[Marco Montalto]

Apologies Thang Le for not being more clear. The problem I found with the baseline being 1 is that your ERSP data is not normalised to baseline in the first place - that's what you said in your post on eeglablist.

The formula is:

(A-R)/R

Which is equivalent to:

A/R - 1

A/R is your ERSP data (power values) normalised to baseline.

But since you did not normalise to baseline, you have to extract power values for both baseline and experimental condition and then input both in the above formula.

Had your data been normalised to baseline, I believe you could then use the formula as your power values would be equivalent to A/R.

Let me know if I am still not being understandable.

Marco

On 14 Jan 2013, at 21:40, Thang Le <thang.le at stonybrook.edu> wrote:

> Hi Marco,
> 
> I'm not sure that I understand your question about the baseline being as 1. Could you please elaborate?
> 
> I did indeed use the same formula as in the post you forwarded to me. 
> 
> TL
> 
> On Sat, Jan 12, 2013 at 5:24 PM, Marco Montalto <montaltomarco at onvol.net> wrote:
> Dear Thang Lee,
> 
> I saw your post on eeglablist. I have a problem with part of your code, which is the following:
> 
> ClustERD_cond1 = (10.^(ClustErspData_cond1/10)-1)*100;
> 
> I am assuming you are using Pfurtscheller's method to compute ERD/ERS. If so, what is your justification for specifying the baseline as 1?
> 
> Pfurtscheller's method is:
> 
> (A-R)/R*100
> 
> Please note the brackets enclosing A-R!!!
> 
> >From your previous email I am now assuming that this was the reason why you were getting such large values for percentage change in alpha.
> 
> Thanks and regards,
> Marco
> 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://sccn.ucsd.edu/pipermail/eeglablist/attachments/20130114/8bbf3028/attachment.html>