摘要
神经系统是一个庞大的系统，系统的结构不但从横向可以分为若干功能子系统，而且从纵向也可分为许多功能层次，各层次间不但物理构成有很大的差异，而且功能上也有相当的联系，下一层次是上一层次的基础。本文即是由此观点出发，研究在离子通道（亚细胞层次）的原理上，用仿真方法推导出神经元（细胞层次）的主要特性，由此可以建立神经元的模型（包括空间总和，时间总和，阈值，不应期，输出静特性，适应性），本文还讨论了不能由离子通道推导出的那些神经元特性，最后从计算开销上说明，研究哪一层次上的功能就应该在哪一层次上建模推导，企图由上一层次的原理推导出上一层的功能（进一步就是企图建立一个万能的神经系统功能原理，象力学中的牛顿定理一样），不仅是不必要的，也是不可能的。
eural system is huge.The structure of is can be divided not only into severalfunctional subsystems horizontally but also into many functional levels verticalyy.There arenot only great differences of physical structures between different levels, but also there aremuch connection functionally.Each level is the basis of its upper one.In this paper,wededuce the main features of neuron (cell level),based on the ionic channel principles(sub-cell level),using emulation method,and we construct a neron model(including space summation,temporal summation, threshold,refractory period, adaptability and output staticcharacteristics).We also discuss what cannot be deduced from the ionic channel. At last,we explain why we should construct a model and deduce it on the level which we arestudying from the point of view of expenditure on computation.It is unnecessary andimpossible to deduce the functions of an upper level from its lower level(and try toconstruct a omnipotent neural system in the manner of Newton's laws in mechanics).
出处
《北京生物医学工程》
1997年第2期102-106,共5页
Beijing Biomedical Engineering
基金
国家自然科学基金
863计划
攀登计划资助
