We present a novel integration technology for replacing damaged microcircuits in the rat brain with electronic neuronal networks. This technology will allow simulating important brain functions, such as memory, through the adaptive stimulating effects of neural interfaces at cellular and neuronal network levels. The proposed system includes electronic devices to register and process neuronal signals and to adaptively stimulate damaged brain areas according to the online algorithm decoding the registered neuronal activity. The stimulation algorithm is implemented in hardware neural network with memristive connections that allows both network learning and adaptive re-learning.The basic idea is to develop components of hardware neural network with memristive connections for adaptive generation of electrical signals to stimulate particular brain areas. Moreover, the uniqueness of the system is adaptive closed-loop connection between two hippocampal neuronal networks. Such a system is very promising for restoring functionality of damaged areas in the human brain to replace lost functions in patients with brain pathologies or trauma.