This paper illustrates an earlier introduced method for systematic monitoring and control of slow electro-mechanical oscillations in electric power systems. The emphasis is on generalizing the two-area system concepts to a general multi-area power system. The method illustrates a transformed state space which explicitly models tie-line flow power dynamics between areas of interest (groups of interconnected equipment) and the rest of the system. Given parameters of an electric power grid, these tie-line power flow oscillations are expressed in terms of load and generation power deviations from given equilibrium, both controlled and uncontrolled. This new model sets the basis for identifying root causes of slow inter-area oscillations, needed to define the most important measurements and placement/enhancement of controllers for reducing these interactions. Notably, both conventional model and the transformed state space model have the same eigenvalues. The proposed approach does not require eigenmode calculations, and therefore lends itself to practical use of Phasor Measurmeent Units (PMUs). It is then shown how each area can monitor, compute and control inter-area oscillations either in an entirely decentralized way, much the same way as today's Automatic Generation Control (AGC) is carried out, or in a cooperative efficient way by jointly controlling inter-area fluctuations.