Summary: The conceptual model and computer simulations results of path integration in free-scalable nonlinear oscillator neural networks with even cyclic inhibition (ECI-networks) are discussed. To estimate the phase shifting under the input impact the ECI-networks contain two subsystems, namely reference and information ones. The population of reference (nonencoding) oscillatory units has a significant role in generation and stabilization of numerous time scales despite that it does not assist directly in the phase pattern encoding of input signals. Multifunctionality is the main characteristics of ensemble encoding of the environment space because the same ensembles can encode (to present coherently) different events of the environment space. It was experimentally shown that the high-precision frequency-phase mechanism in the frameworks of ensemble hypothesis can be used in the navigation behavior.