[Eeglablist] CSD maps.

[Darren Weber]

I've come the conclusion that it is -conductivity.  The scalp conductivity
is often estimated at 0.33 Siemens/meter.  (Note this is often also used for
brain areas).  There are various values toward the bottom of the notes.

I've noticed that some publications use -1, and then indicate that the
values plotted are "proportional" to SCD, where the proportionality depends
on conductivity.  I guess that is OK, but as a novice I found it initially
perplexing to see the same polarity as the potential data, with units of
uV/cm^2, yet discussion as though the values are current (or SCD, CSD).  It
took some digging into one or another paper to find the comment about
proportionality in the methods (not all SCD papers do that).  There are also
variations in the units used, such as V/m^2 or uV/m^2 or uV/cm^2, etc.  I
know topography depends a lot on colormaps, but it can make a difference to
waveform plots of SCD.  Most reports of scalp potential give units of uV,
but I'm not clear that a similar convention has been adopted for SCD.

See Nunez for discussions on the relationship between the surface Laplacian
and cortical sources.  Also see Perrin and Pernier, as well as Oostendorp
(some useful references should be in those notes).  The simulation work
shows the Laplacian can "estimate" shallow cortical sources better than the
scalp potential.  The other advantage is that it is a reference free value
(by "reference" I mean the EEG acquisition reference, eg, mastoid).

Best, Darren


----- Original Message ----- 
From: "Christopher Summerfield" <summerfd at paradox.psych.columbia.edu>
To: "Darren Weber" <darren.weber at radiology.ucsf.edu>
Cc: "Arnaud Delorme" <arno at salk.edu>; "eeglablist"
<eeglablist at sccn.ucsd.edu>
Sent: Wednesday, November 05, 2003 9:14 AM
Subject: Re: [Eeglablist] CSD maps.


> thanks very much darren.
>
> if my understanding of the link is correct, then if you want to use CSD
> (or SCD) maps to 'sharpen' your image - or to show activity which is most
> likely due to a cortical source - then you should multiply the result of
> the del2map function by -1 (correct me if I'm wrong)
>
> best
> chris
>
>
>
>
> Christopher Summerfield
> summerfd at psych.columbia.edu
>
> On Wed, 5 Nov 2003, Darren Weber wrote:
>
> >
> > The Laplacian is the inverted polarity of the potential.  I have had
> > problems understanding this, so I went through the process of better
> > understanding the mathematical physics involved.  My notes on this are
here:
> >
> > http://dnl.ucsf.edu/users/dweber/dweber_docs/eeg_scd.html
> >
> > This document has some matlab examples of how the potential polarity is
> > inverted for the Laplacian.  Maybe it will help.  I've been confused by
some
> > figures in the literature, please be careful to get the polarity and
units
> > of your figures correct.
> >
> > I expect the eeglab code is working correctly, in that it is giving the
> > Laplacian.  If my understanding is correct, we should multiply
> > by -conductivity to get the "scalp current density".  In this case, the
term
> > 'scalp current density' actually refers to the divergence of the current
> > density in the scalp.
> >
> > BTW, I guess I don't like the term "current source density" because the
> > surface Laplacian has very little to do with "sources" and "sinks" (the
> > application of Laplace's equation specifically assumes no sources or
sinks,
> > otherwise I think we should be considering Poison's equation).  I
prefer,
> > "scalp current density" but even this is not entirely satisfactory.
> >
> > I must confess that my understanding of the mathematical physics is not
> > perfect.  My qualifications on this topic are inadequate, so please
consider
> > these comments and the notes with reasonable scepticism.
> >
> > Best, Darren
> >
> > ----- Original Message -----
> > From: "Arnaud Delorme" <arno at salk.edu>
> > To: "Christopher Summerfield" <summerfd at paradox.psych.columbia.edu>
> > Cc: "eeglablist" <eeglablist at sccn.ucsd.edu>
> > Sent: Friday, October 31, 2003 5:27 PM
> > Subject: Re: [Eeglablist] CSD maps.
> >
> >
> > > >
> > > >
> > > >strange thing is, the code below plots a CSD map that looks exactly
like
> > > >the inverse of my scalp map.
> > > >
> > > What do you mean? Del2map uses the Matlab del2 function to compute the
> > > 2-D laplacian. I checked the code of del2map and there does not seem
to
> > > be any sign inversion.
> > >
> > > >is there a problem with the code?
> > > >
> > > >
> > > This is a good point. Unfortunatelly, there can always be a problem
with
> > > the code :-). If you have doubts and have the capacity to understand
the
> > > code, you should definetelly check that it does what you expect (if it
> > > doesn't, then contact us).
> > >
> > > Arno
> > > --
> > >
> > > *Arnaud Delorme, Ph.D.*
> > > Computational Neurobiology Lab, Salk Institute
> > > 10010 North Torrey Pines Road
> > > La Jolla, CA 92037 USA
> > >
> > > *Tel* : /(+1)-858-458-1927 ext 15/
> > > *Fax* : /(+1)-858-458-1847/
> > > *Web page *: www.sccn.ucsd.edu/~arno
<http://www.sccn.ucsd.edu/%7Earno>
> > >
> > >
> > > _______________________________________________
> > > Eeglablist mailing list Eeglablist at sccn.ucsd.edu
> > > Eeglablist page: http://www.sccn.ucsd.edu/eeglab/eeglabmail.html
> > > To unsubscribe, send an empty email to eeglablist-unsub at sccn.ucsd.edu
> > > To switch to digest mode, send an empty email to
> > eeglablist-digest at sccn.ucsd.edu
> > >
> >
>