Re: PDC

From: Alois Schlögl <alois.schloegl ***AT***tugraz.at>
Date: Tue, 13 Jun 2006 18:09:58 +0200
Subject: Re: PDC
Newsgroups: tu-graz.biosig
Organization: Technische Universitaet Graz
Message-id: <448EE356.4020502@tugraz.at>
User-agent: Mozilla Thunderbird 0.9 (Windows/20041103)
References: <000601c68bb1$ecf85dd0$0200a8c0@Elias> <448971FC.2040805@tugraz.at> <000901c68d00$0d33c1d0$0200a8c0@Elias> <448D1239.1070304@tugraz.at> <000c01c68ee0$47359430$0200a8c0@Elias>
Heracles Panagiotides wrote:

I tried the following code with the attached data file (12 channels X 320 time points) and got an error.

load ('lface.mat');

P=5;

[AR,RC,PE] = mvar(lface,P);


The data need to be one channel per column; and your first channel is always zero, yielding a singularity - remove it. Then, this is
[AR,RC,PE] = mvar(lface(2:end,:)',P);

M = size(AR,1); % number of channels

A = [eye(M),-AR];

B = eye(M);

C = PE(:,M*P+1:(M+1)*P);


This should be
    C = PE(:,M*P+1:M*(P+1));


[S,h,PDC,COH,DTF,DC,pCOH,dDTF,ffDTF,pCOH2,PDCF] = mvfreqz(B,A,C,64,512);



Visualization is supported by PLOTA. You can do the following:

[AR,RC,PE] = mvar(lface(2:end,:)',P);
X.A = [eye(M),-AR]; X.B = eye(M); X.C  = PE(:,(1-M:0)+end);
X.datatype = 'MVAR';
X.SampleRate = 512;
plota(X,'PDC');


Alois

----- Original Message ----- From: "Alois Schloegl" <alois.schloegl***AT***tugraz.at>
To: "Heracles Panagiotides" <hercp***AT***u.washington.edu>
Sent: Monday, June 12, 2006 12:05 AM
Subject: Re: PDC


help mvfreqz




Heracles Panagiotides wrote:

Dear Alois,

Thank you for the reply.  It is most appreciated.

I am wondering if I can ask you a very specific uestion. If I have a matrix, where the columns are channels and the rows time points, what would be the most direct way for obtaining PDCs using the biosig library? These are intracranial ERP data.

Best,

Heracles Panagiotides, Ph.D.


----- Original Message ----- From: "Alois Schloegl" <alois.schloegl***AT***tugraz.at>
To: "Heracles Panagiotides" <hercp***AT***u.washington.edu>
Sent: Friday, June 09, 2006 6:05 AM
Subject: Re: PDC


Heracles Panagiotides wrote:

Dear Dr. Schoegl,
I found a PowerPoint presentation by you and colleagues on the internet on Partial Directed Coherence.. Are you aware of any matlab code that can compute this function? You mention in one of the slides a PDC function. Is it part of the BIOSIG package? I could not locate it. Any help in this direction would be greatly appreciated.
 Kind regards,
 Heracles Panagiotides, Ph.D.




Yes, it is in BioSig. See MVFREQZ.M


Kind regards,
   Alois Schlögl