Page 2 of 2

question about r^2

Posted: 11 May 2010, 11:52
by EliGC
Hello, I don´t know if you can answer my question but I will try to explain you what is my objective with the r^2. :oops:

I am trying to make a time-frequency analysis for my data during a sensorimotor rhythm experiment. The point is that if I want to be sure that I found a desynchronization or synchronization I need a reference to compare with, and this reference is usually the rest period (during intertrial interval). But in this case, I am not sure if is coherent to calculate the r^2 between two data conditions that has as dimentions time*frequency*power, and are separable in time, I mean, one condition is stimulus interval and the other one is the rest or intertrial interval.

Is that coherent? :?:

Thank you so much, sorry for the question but I am staring with this kind of analysis and don´t know how to deal with.
:?

time-frequency

Posted: 12 May 2010, 07:15
by gschalk
Eliana,

What you are trying to do is fine. Actually, you have the dimensions: time x frequency x trials, and that for two conditions, e.g., task and rest or task1 and task2.

Gerv

confussion about r^2

Posted: 07 Sep 2010, 10:49
by EliGC
Good Morning, I Have two main questions about r^2:

1. If I am doing a P300 experiment with two subjects, and I found that one has a bigger value of r^2 than the other one, what this exactly means? That one can evoked stronger the P300?

2. In the dissertation that is posted here to find more information about coefficient of determination, I found that the equation describes the correlation coefficient and no the coeficcient of determination, according to Wonnacott and Wonnacott (r^2=Sxy/SxSy). I know how you calculate this in the software but I would like to have more clarity of this concept, the latter because I tried to do it by myself and I don't know how to calculate the covariance (SxSy) between two vectors of different lenghts (at least this is no possible in Excel).

Thank you so much,

Re: Signal Processing questions ...

Posted: 14 Jun 2011, 19:06
by keliao
hi, Schalk,

I read your thesis about r^2 calculation (equation 2.1). I think this equation is the same as Pearson r (correlation coefficient) definition. And it's different with the definition from Wiki, which has the same as the book you refered in your thesis. I wonder how to calculate the r^2 and what's the meaning?

I am a newer for BCI. Maybe I am totally wrong. Would you please give me some information?

Best Regards,

keliao
gschalk wrote:Mitsuru,

The calculation of the r^2 values is described in a dissertation that you may download from:

http://www.bciresearch.org/tmp/thesis.pdf

You can also find a Matlab routine rsqu.m in \src\tools\analysis\P3 that you can use.

Thank you also for the suggestion to improve on the manuals. We are working on dramatically improved documentation at present.

Gerv

Re: r2 values

Posted: 17 Jun 2011, 14:27
by mellinger
Hi,

you are right, the r^2 value is indeed the square of the correlation coefficient.

When mentioning an r^2 definition on the wiki, are you referring to this page:
http://www.bci2000.org/wiki/index.php/G ... #r-squared
?

The formula there shows how to convert an r^2-value into a t-value (under the assumption of Gaussian distribution). A t-value is used in statistics to determine how significant a difference between two distributions is.

Best regards,
Juergen

Re: r2 values

Posted: 18 Aug 2011, 12:53
by keliao
hi, Juergen,

Thanks for your reply.

When i search the r^2, I find the wiki page named "coefficient of determination":
http://en.wikipedia.org/wiki/Coefficien ... ermination
It looks the "coefficient of determination " is not the square of coefficient of correlation. is that page the right definition of r^2 we use in BCI?

btw, from the page your send, i think there is no equation how to calculate r^2. am I correct?

Best Wishes,

keliao
mellinger wrote:Hi,

you are right, the r^2 value is indeed the square of the correlation coefficient.

When mentioning an r^2 definition on the wiki, are you referring to this page:
http://www.bci2000.org/wiki/index.php/G ... #r-squared
?

The formula there shows how to convert an r^2-value into a t-value (under the assumption of Gaussian distribution). A t-value is used in statistics to determine how significant a difference between two distributions is.

Best regards,
Juergen

Re: r2 values

Posted: 18 Aug 2011, 14:34
by mellinger
When i search the r^2, I find the wiki page named "coefficient of determination":
http://en.wikipedia.org/wiki/Coefficien ... ermination
It looks the "coefficient of determination " is not the square of coefficient of correlation. is that page the right definition of r^2 we use in BCI?
On that Wikipedia page, it is clearly stated that the "coefficient of determination" is the squared coefficient of correlation when correlation is used as a means to infer how much a measured outcome is determined by some underlying quantity. In BCI, one uses r^2 values to assess how much a brain signal is determined by a task that the user performs.
btw, from the page your send, i think there is no equation how to calculate r^2. am I correct?
The formula on the BCI2000 wiki page does not show how to compute an r^2 value but how to convert an r^2 value into a t value, as I wrote in my previous post.

Regards,
Juergen