Home > functions > leadfield > eeg > vb_berg_param.m

vb_berg_param

PURPOSE ^

Berg Parameter calculation for EEG multilayer spherical forward model

SYNOPSIS ^

function [mu, lam, f, fval, flag, nsteps] =vb_berg_param(R, sigma, NJ, Nmax, Nopt, mu_ini, lam_ini)

DESCRIPTION ^

 Berg Parameter calculation for EEG multilayer spherical forward model 
   [mu,lam]   = vb_berg_param(R, sigma)
   [mu,lam,f] = vb_berg_param(R, sigma, NJ, Nmax, Nopt)

 Computes the Berg parameters:
 
 INPUTS (Required):
    R    : Radii of sphere from innermost to outermost    ( NL x 1 )
    sigma: Conductivity    from innermost to outermost    ( NL x 1 )

 INPUTS (Optional):
   Nmax : # of Legendre Expansion 
            to fit Berg Parameters   (Default: 200)
     NJ : # of Berg Dipoles          (Default: 3)
   Nopt : # of optimization search
  mu_ini: Initial value for radial factors           ( NJ x 1 )
 lam_ini: Initial value for magnitude factors        ( NJ x 1 )
  
  NL = # of sphere layers; 
  NJ = # of Berg Dipoles

 OUTPUTS:
    mu: Berg radial factors            ( NJ x 1 )
   lam: Berg magnitude factors         ( NJ x 1 )
     f: Legendre Expansion Weights used to fit Berg Parameters ( Nmax x 1 )


 Single Dipole in a multilayer sphere can be approximated 
        by multiple dipoles in a single shell. 
 Z. Zhang,  Phys. Med. Biol., vol.40, pp. 335-349, 1995.


 Copyright (C) 2011, ATR All Rights Reserved.
 License : New BSD License(see VBMEG_LICENSE.txt)

CROSS-REFERENCE INFORMATION ^

This function calls: This function is called by:

SOURCE CODE ^

0001 function [mu, lam, f, fval, flag, nsteps] = ...
0002          vb_berg_param(R, sigma, NJ, Nmax, Nopt, mu_ini, lam_ini)
0003 % Berg Parameter calculation for EEG multilayer spherical forward model
0004 %   [mu,lam]   = vb_berg_param(R, sigma)
0005 %   [mu,lam,f] = vb_berg_param(R, sigma, NJ, Nmax, Nopt)
0006 %
0007 % Computes the Berg parameters:
0008 %
0009 % INPUTS (Required):
0010 %    R    : Radii of sphere from innermost to outermost    ( NL x 1 )
0011 %    sigma: Conductivity    from innermost to outermost    ( NL x 1 )
0012 %
0013 % INPUTS (Optional):
0014 %   Nmax : # of Legendre Expansion
0015 %            to fit Berg Parameters   (Default: 200)
0016 %     NJ : # of Berg Dipoles          (Default: 3)
0017 %   Nopt : # of optimization search
0018 %  mu_ini: Initial value for radial factors           ( NJ x 1 )
0019 % lam_ini: Initial value for magnitude factors        ( NJ x 1 )
0020 %
0021 %  NL = # of sphere layers;
0022 %  NJ = # of Berg Dipoles
0023 %
0024 % OUTPUTS:
0025 %    mu: Berg radial factors            ( NJ x 1 )
0026 %   lam: Berg magnitude factors         ( NJ x 1 )
0027 %     f: Legendre Expansion Weights used to fit Berg Parameters ( Nmax x 1 )
0028 %
0029 %
0030 % Single Dipole in a multilayer sphere can be approximated
0031 %        by multiple dipoles in a single shell.
0032 % Z. Zhang,  Phys. Med. Biol., vol.40, pp. 335-349, 1995.
0033 %
0034 %
0035 % Copyright (C) 2011, ATR All Rights Reserved.
0036 % License : New BSD License(see VBMEG_LICENSE.txt)
0037 
0038 % # of sphere
0039 NL = length(R);
0040 
0041 if nargin < 3  
0042     % Default # of Berg Parameters
0043     NJ=3;       
0044 end
0045 
0046 if nargin < 4    
0047     % Default # of Legendre expansion
0048     Nmax = 200;  
0049 end
0050 
0051 if nargin < 5    
0052     % Default # of iteration in optimum search
0053     Nopt = 1000;  
0054 end
0055 
0056 if nargin < 6
0057     % Default Value for Initial Eccentricity Parameters
0058     mu_ini = (1/(NJ+2))*[1:NJ];  
0059     % Default Value for Initial Magnitude Parameters
0060     lam_ini = 0.2*ones(1,NJ);      
0061 end
0062 
0063 %
0064 % Set minimization option
0065 %
0066 %OPTIONS = optimset('MaxFunEvals',  NJ*Nopt, ...
0067 %                   'MaxIter',      NJ*Nopt, ...
0068 %                   'TolFun', [0  1.e-9  1.e-9]);
0069 
0070 OPTIONS = optimset('MaxFunEvals',  NJ*Nopt, ...
0071                    'MaxIter',      NJ*Nopt, ...
0072                    'TolFun', [0  1.e-8  1.e-8]); 
0073 
0074 % Computes the weights f(n) for legendre expansion
0075 
0076 f = vb_multi_shell_param(R, sigma, Nmax);
0077 
0078 %
0079 % Minimize the error function defined by Zhang
0080 %
0081     
0082 [berg, fval, flag, outputs]= ...
0083     fminsearch('vb_berg_error',[mu_ini lam_ini(2:NJ)], OPTIONS, R, f); 
0084 
0085 nsteps = outputs.iterations;
0086 
0087 mu        = berg(1:NJ);
0088 lam(2:NJ) = berg(NJ+1:2*NJ-1);
0089 lam(1)    = f(1) - sum(lam(2:NJ));

Generated on Tue 27-Aug-2013 11:46:04 by m2html © 2005