I was engaged at Moscow in pioneer research concerned with developing numerical algorithms to investigate the bifurcations of steady states and limit cycles in nonlinear dynamical systems under parameter variation. These investigations showed that cooperative effects like physical phase-transitions and synchronisation phenomena occur normally in biological neural networks despite a very chaotic spike activity of single neurons. In addition, metastable states of neural networks were proved to be useful to model short-term memory in a series of theoretical and simulation works.