The Observer Design Of The Neural Mass Models

Electroencephalography signal contains a large number of brain information, providing a basis for the diagnosis and treatment of some brain diseases. The modeling of brain neural system can deepen our understanding for the brain structure and function, and the model can also be to simulate EEG signal, which is helpful to study the working principle of the brain neural system, and provides the theoretical basis for the mechanism analysis of some diseases in the brain neural system. The neural mass model, as a classic model of modeling the brain neural system, has made some achievements in the several years, but it is still one of the hot spots in the neuroscience research.This paper studies the observer design of the neural mass model, a classic model of modeling the brain neural system, designing an observer of nonlinear systems with the theory of Lurie systems and matrix manipulation. The all contents are in the following:Firstly, we design the observer of the neural mass model that is the nonlinear system, and the constraint inequality condition to the observer matrix is put forward. The inequality condition is solved and the observer matrices are obtained by using YALMIP toolbox of MATLAB. The simulation shows that the designed observer performance is better than the previous results. Secondly, the designed observer is optimized by using the theory of Lurie systems, to reduce the influence of the input disturbance and measurement noise for the observer performance. Two gain value and the comparison between the neural mass model real state and the estimate state of observer, both illustrate the performance of the designed observer has been further improved. Finally, on the hardware the FPGA, based on SOPC it implements the brain dynamic model and producing spikes with neural population models, then lay the foundation for the realization of the hardware the observer.