Rodriguez Campos Francisco A, Enderle John
University of Connecticut, Storrs, Connecticut 06269-2157, USA.
Biomed Sci Instrum. 2003;39:93-8.
This paper describes the implementation of a neural network for the adaptative control of the saccadic system. The model shows the cerebellum plays an important role in the adaptive control of the saccadic gain. Using only eye position input through the granule cells, the cerebellum projects this signal to the other cerebellar structures and then to motor neurons responsible for the saccade. The generation of an adjustment signal occurs in the inferior olive as a result of the error sensory signal created by the open loop saccade system from propioceptive position inputs from the last eye movement generated by the network until the movement towards the target is completed. In addition, a memory component has been defined in the error system to achieve the adaptation. This neural network involves only the horizontal saccade component modeled with Matrix Laboratory language (MATLAB), in conjunction with the Simulink tool.
本文描述了一种用于眼球扫视系统自适应控制的神经网络的实现。该模型表明，小脑在眼球扫视增益的自适应控制中起着重要作用。小脑仅通过颗粒细胞接收眼位输入，将该信号投射到其他小脑结构，然后再投射到负责扫视的运动神经元。由于开环扫视系统根据网络生成的上一次眼球运动的本体感受位置输入产生的误差感觉信号，在下橄榄核中产生调整信号，直到向目标的运动完成。此外，在误差系统中定义了一个记忆组件以实现适应性。该神经网络仅涉及用矩阵实验室语言（MATLAB）结合Simulink工具建模的水平扫视组件。
