摘要
KⅢ模型真实地模拟了整个嗅觉神经系统,包括嗅上皮、嗅球层和嗅皮层.梨状皮质是嗅球层投射的最大区域,在气味信息处理、嗅觉产生的过程中起着非常重要的作用.但是KⅢ模型简化了对梨状皮质的模拟,这部分的模型结构仅使用一个KⅠ模型和一个KⅡ模型来表示梨状皮质中神经元的连接情况.为了完善KⅢ模型,本文结合嗅上皮层细胞、嗅球层中的嗅小球细胞和嗅皮层梨状皮质中锥体细胞之间的神经元比例关系,以及梨状皮质内部的神经回路、细胞结构,从仿生学原理对KⅢ模型进行改进.同时,本文采用小世界网络理论分析了输入通道数从5逐渐增加到100时,改进前后KⅢ模型的平均路径长度L和聚类系数C的变化情况.分析结果表明,与原KⅢ模型相比,改进KⅢ模型在输入通道数大于16时,具有较小的平均路径长度和更大的聚类系数;当输入通道数超过20时,改进KⅢ模型网络的C与对应等效随机图的Crand的比值γ增长更快且远大于改进前模型,即改进后的KⅢ模型较改进前的模型具有更强的小世界特性.
The KⅢmodel truly mimics the entire olfactory nervous system,including the olfactory epithelium,olfactory bulb,and olfactory cortex.The piriform cortex is the largest region projected by the olfactory bulb and plays an important role in the process of odor information processing and olfactory production.However,the KⅢmodel simplifies the simulation of the piriform cortex,and this part of the model structure only uses a KⅠmodel and a KⅡmodel to represent the connections of neurons in the piriform cortex.To improve the KⅢmodel,this paper combine the olfactory epithelial layer cells,the olfactory glomerulus cells in the olfactory bulb layer,and the neuronal ratios between the pyramidal cells in the piriform cortex of the olfactory cortex,as well as the neural circuits and cell structure within the piriform cortex,toimprove the KⅢmodel from the principle of bionics.At the same time,this paper uses small-world network theory to analyze the changesin the average path length L and clustering coefficient C of the KⅢmodel before and after the improvement when the number of input channels increases from5to100.The results show that compared with the original KⅢmodel,the improved KⅢmodel has a smaller average path length and a larger clustering coefficient when the input channel number is greater than16.When the number of channel inputs exceeded20,the ratioγof the improved KⅢmodel network C to Crand in the corresponding equivalent random graph increased more rapidly and significantly than the model before the improvement,which means the improved KⅢmodel had stronger small-world characteristics than the model before the improvement.
作者
张锦
陈玲钰
田森
刘宏
田恬恬
ZHANG Jin;CHEN Ling-yu;TIAN Sen;LIU Hong;TIAN Tian-tian(College of Information Science and Engineering,Hunan Normal University,Changsha 410081,China;School of Mathematics and Statistics,Hunan Normal University,Changsha 410081,China;College of Information Engineering,Zhengzhou Institute of Science and Technology,Zhengzhou 450064,China)
出处
《小型微型计算机系统》
CSCD
北大核心
2022年第5期1027-1032,共6页
Journal of Chinese Computer Systems
基金
湖南省科技厅高新技术产业科技创新引领项目(2020GK2009)资助
军委装发预研项目(31511010105)资助
国防科工局国防基础科研计划项目(WDZC2020550011)资助
湖南省交通运输厅科技进步与创新计划项目(201927)资助
湖南省研究生培养创新实践基地项目(湘教通[2019]248号)资助
湖南省研究生教改项目(JG2018A012)资助
国家教育部产学合作协同育人项目(201901051021,202002004002,202002140003)资助。