Shan-chen model

Dear all,
http://www.lbmethod.org/doku.php?id=numerics:matlab_samples
In the Shan-Chen model,

drho = 0.001;
delta_rho = -drho*(1-2.0*rand(lx));

for i=1:9
fIn(i,1:lx,1:ly) = tNS(i).(1.0 + delta_rho);
gIn(i,1:lx,1:ly) = tNS(i).
(1.0 - delta_rho);
end

Why are plus and minus used in the function of fIn and gIn respectively?

Nash

Is this connected with the initial condition?
Had someone read this?