This paper studies possible bifurcations and synchronization of subpopulations of a class of macroscopic models called neural mass models. These models describe the mean activity of entire neural populations, represented by their averaged firing rates and membrane potentials. Connections between the nodes represent the static nonlinear sigmoidal function. We study local bifurcations and make a global stability analysis for one subpopulation of the neural mass model. Also we consider the behavior of two coupled subpopulations and find the sufficient conditions of their synchronization. (C) 2017, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.