graph2net.m

function W = graph2net(sA,e,d)

% function W = graph2net(sA,e,d)
%
% sA = binary adjacency matrix for a directed graph
% e = epsilon value to control synaptic strengths (default: e = 0.25)
% d = delta value for inhibitory weights (default: d = 2*e)
%
% create matrix W for a threshold-linear network from directed graph 
%
% modified june 22, 2015
% updated dec 30, 2020 to allow generalized sA matrices for interpolation
% updated apr 23, 2022 to fix bug: A was unassigned, now fixed
% updated aug 10, 2023 to remove A since unnecessary

if nargin < 2 || isempty(e)
    e = 0.25;
end;

if nargin < 3 || isempty(d)
    d = 2*e;
end;

% create matrix W from sA
W = zeros(size(sA));
n = size(sA,1);
for i=1:n
    for j=1:n
        t = sA(i,j);
        W(i,j) = t*(-1+e) + (1-t)*(-1-d);
    end
    W(i,i) = 0; % make sure diagonal is 0
end