Computational Studies on the Cognition and Memory Mechanisms of the Layered Neural Network with Asymmetric Structure

分析了具有奇-偶(或偶-奇)阶非线性的生物非对称网络的结构，用于网络计算。运动知觉是大脑分层视觉系统的一种基本机制。在基本不对称子网络计算中，推导了运动及其运动速度的不变量方程。作为视觉系统的模型，本文建立了分层的异质网络，给出了视皮层V1区和MT区运动矢量的生成模型。

The architecture of the biological asymmetric network with odd-even(or even-odd) order nonlinearities is analyzed for the network computations. Motion perception is a basic mechanism in the layered visual system of the brain. The invariant equation for evaluation of the movement and its movement velocity are derived in the fundamental asymmetric sub-network computations. As the model of visual system, the layered heterogeneous network is developed here, which shows a generation model of movement vectors in V1and MT area of the visual cortex.