sA2soln.m

function soln = sA2soln(sA,T,X0,epsilon,delta,theta)

% function soln = sA2soln(sA,T,X0,epsilon,delta,theta)
%
% sA = n x n binary adjacency matrix for a directed graph
% T = length of time for ode solution, in units of leak timescale tau
% -> default is T=70
% X0 = n x 1 vector of initial conditions (default is random near all-zeros)
% epsilon/delta are the values for weights -1+epsilon, -1-delta in W
% -> defaults are those from graph2net.m
% theta = n x 1 vector of external stimulus or scalar value
%
% Note: theta = b in threshlin_ode, and T is also passed directly
% to threshlin_ode, so added flexibility there applies.  In particular
% we can have theta change values to make a pulse train of "kicks."
%
% calls: graph2net.m and threshlin_ode.m
% last modified Aug 8, 2020, to change defaults to match quick_plot ones
% modified Aug 16, 2023 to revert to e=0.25 and d=0.5 defaults

n = size(sA,1); % no. of neurons

if nargin < 2 || isempty(T)
    T = 70;
end;

if nargin < 3 || isempty(X0)
    X0 = zeros(n,1) + .01*rand(n,1); % break symmetry on init conds
end;

if nargin < 4 || isempty(epsilon)
    epsilon = 0.25;
end;

if nargin < 5 || isempty(delta)
    delta = 0.5; 
end;

if nargin < 6 || isempty(theta)
    theta = 1;
end;

% create network W from sA
W = graph2net(sA,epsilon,delta);

% create external input vector b from theta
if length(theta) == 1
    b = theta*ones(n,1);
else
    b = theta;
end;

% simulate activity with constant b and initial conds X0
soln = threshlin_ode(W,b,T,X0);

% add adjacency matrix to soln struct
soln.sA = sA;
soln.eps = epsilon;
soln.delta = delta;