We summarize the approach leading to the dynamical mean-field equation 
for the spike emission rate $\nu(t)$ of an interacting population of 
Integrate-and-Fire (IF) neurons derived in \cite{md02}. Building on 
the results concerning the stability conditions and finite-size effects, 
we investigate how such properties are affected by a non-trivial 
distribution of spike transmission delays. The main findings are: 
i) the stability of the collective asynchronous states is improved 
by widening the distribution of delays; ii) high-frequency components of 
the power spectrum of the collective activity are damped; 
iii) Details of the stability and spectral properties are strongly 
affected by the shape of the delay distribution. We present quantitative 
predictions from the theory and we demonstrate a very good agreement 
with simulations.