DSPRelated.com
Forums
   Forums    More Forums    comp.dsp  

Computing EEG power vs time

Started by cherrie ●October 2, 2007
Hi all,

I am working on some EEG data and I would like to compute the power vs
time. I know that i can perform an FFT to obtain the power, however, this
would be for specific frequencies. I read that the power = magnitude
square of the voltage. Is it valid then to compute the power at each time
point by squaring the voltage magnitude and then plotting these against
time to obtain the EEG power vs time plot?

Thank you for your help in advance.

cherrie
Reply by John Sampson ●October 2, 20072007-10-02
cherrie wrote:

> Hi all,
> 
> I am working on some EEG data and I would like to compute the power vs
> time. I know that i can perform an FFT to obtain the power, however, this
> would be for specific frequencies. I read that the power = magnitude
> square of the voltage. Is it valid then to compute the power at each time
> point by squaring the voltage magnitude and then plotting these against
> time to obtain the EEG power vs time plot?
> 
> Thank you for your help in advance.
> 
> cherrie
> 
> 


It is valid if you divide V^2 by the impedance.

John
Reply by Jerry Avins ●October 2, 20072007-10-02
John Sampson wrote:

> cherrie wrote:
>> Hi all,
>>
>> I am working on some EEG data and I would like to compute the power vs
>> time. I know that i can perform an FFT to obtain the power, however, this
>> would be for specific frequencies. I read that the power = magnitude
>> square of the voltage. Is it valid then to compute the power at each time
>> point by squaring the voltage magnitude and then plotting these against
>> time to obtain the EEG power vs time plot?
>>
>> Thank you for your help in advance.
>>
>> cherrie
>>
>>
> 
> It is valid if you divide V^2 by the impedance.


Provided the impedance has no imaginary part.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯¯
Reply by John ●October 2, 20072007-10-02
On Oct 2, 9:11 am, Jerry Avins <j...@ieee.org> wrote:

> John Sampson wrote:
> > cherrie wrote:
> >> Hi all,
>
> >> I am working on some EEG data and I would like to compute the power vs
> >> time. I know that i can perform an FFT to obtain the power, however, t=

his

> >> would be for specific frequencies. I read that the power =3D magnitude
> >> square of the voltage. Is it valid then to compute the power at each t=

ime

> >> point by squaring the voltage magnitude and then plotting these against
> >> time to obtain the EEG power vs time plot?
>
> >> Thank you for your help in advance.
>
> >> cherrie
>
> > It is valid if you divide V^2 by the impedance.
>
> Provided the impedance has no imaginary part.
>
> Jerry
> --
> Engineering is the art of making what you want from things you can get.
> =AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=

=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=
=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF=AF


Well if Z is complex then you can still talk about complex power,
divide Vrms^2 by conj(Z). But I don't think the OP cares about that.

John
Reply by mnentwig ●October 2, 20072007-10-02
As long as you don't need absolute power, V^2 will do.

-mn
Reply by Jerry Avins ●October 3, 20072007-10-03
Jerry Avins wrote:

> John Sampson wrote:
>> cherrie wrote:
>>> Hi all,
>>>
>>> I am working on some EEG data and I would like to compute the power vs
>>> time. I know that i can perform an FFT to obtain the power, however, 
>>> this
>>> would be for specific frequencies. I read that the power = magnitude
>>> square of the voltage. Is it valid then to compute the power at each 
>>> time
>>> point by squaring the voltage magnitude and then plotting these against
>>> time to obtain the EEG power vs time plot?
>>>
>>> Thank you for your help in advance.
>>>
>>> cherrie
>>>
>>>
>>
>> It is valid if you divide V^2 by the impedance.
> 
> Provided the impedance has no imaginary part.


cherriegeller@hotmail.co.uk wrote:
 > Hi, Thank you for your reply to my post \"compute EEG power vs time\".
 > You mentioned that its ok to do V^2 provided i divide by impedance. I
 > think this is from P=V^2/R ?? But how do i obtain R? This is EEG
 > recorded from the brain not from a circuitry?? Sorry i am really new
 > to this. Thanks

Please correspond via the newsgroup unless the message is personal. 
Power is V^2/R where R is a real resistance and V is an RMS quantity. A 
decibel is (strictly) a measure of relative power, but it is often used 
to relate voltages *as if* there were real power involved.

I imagine you aren't investigating how many watts the brain actually 
produces Isolated EEG signals wouldn't by themselves reveal that anyway. 
Use decibel measurements to provide relative fictitious power. For such 
an application, using magnitude or RMS merely changes an additive constant.

Jerry
-- 
Engineering is the art of making what you want from things you can get.
&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;
Reply by mnentwig ●October 3, 20072007-10-03
>> Use decibel measurements to provide relative fictitious power. For such


an application, using magnitude or RMS merely changes an additive
constant.

Yes... or in other words: the reference to "relative fictious power"
becomes a scaling factor c, as in c*V^2, hence: Simply use V^2.

It might be R or it might be an empirically found (constant) fudge factor,
as long as you don't require absolute powers.

-mn
Reply by cherrie ●October 3, 20072007-10-03
>>> Use decibel measurements to provide relative fictitious power. For

such

>
>an application, using magnitude or RMS merely changes an additive
>constant.
>
>Yes... or in other words: the reference to "relative fictious power"
>becomes a scaling factor c, as in c*V^2, hence: Simply use V^2.
>
>It might be R or it might be an empirically found (constant) fudge

factor,

>as long as you don't require absolute powers.
>
>-mn
>


Do you mean to just use V^2 to find the relative power and not dividing
this by R??

Thanks,cherrie
Reply by mnentwig ●October 3, 20072007-10-03
That's what I would do, without more specific requirements.

-mn
Reply by Rune Allnor ●October 3, 20072007-10-03
On 2 Okt, 12:12, "cherrie" <cherriegel...@hotmail.co.uk> wrote:

> Hi all,
>
> I am working on some EEG data and I would like to compute the power vs
> time. I know that i can perform an FFT to obtain the power, however, this
> would be for specific frequencies. I read that the power = magnitude
> square of the voltage. Is it valid then to compute the power at each time
> point by squaring the voltage magnitude and then plotting these against
> time to obtain the EEG power vs time plot?


That's generally the case, yes. I don't know if this is
valid for EEG signals, but for the types of signals
I usually work with I prefer to use an envelope detector
before squaring the amplitudes. The power curve becomes
somewhat smoother that way.

Rune