神经元固有特性对针刺网络编码影响的机制研究

As the most important component of the traditional Chinese medicine, the acupuncture has increasingly attracted attentions from various researchers. The related studies show significant theoretical and practical meanings for the quantization and standardization of acupuncture techniques. The encoding and decoding of acupunctured neuron networks is regarded as one of the most important research topics to scientifically explain the acupuncture mechanism. However, corresponding research results are seldom reported in public. This study first proposes a neuronal-network-based method to explain the encoding mechanism of the acupuncture technology, aiming to explore the impact of the intrinsic properties on the network encoding mechanism, thereby establishing the quantitative relationship between the acupuncture and neuron responses. By synthetically analyzing the experimental data and mathematical model, this study demonstrates the rule of impacts among the intrinsic properties, input-output characteristics, synchronization, and neuronal network encoding. Also, by utilizing the state space method to realize the classification of multi-array acupuncture electric signals and correlation-based analysis, this study explains the characteristics of neuron responses to the acupuncture input. Based on the feed-forward network, additionally, the acupuncture conduction can be successfully emulated in this study. Accordingly, the virtual network method is proposed and implemented to describe the interactions between the neuronal intrinsic properties and acupuncture and synaptic inputs, as well as the corresponding impact on the synchronized conduction. Then, the encoding rules of the acupunctured neuronal network can be successfully quantized according to corresponding characteristics. Besides, this study integrates the experimental data, statistical analysis and computational model, establishes the biophysical neuron model, describes the impact of the neuronal intrinsic properties on the synchronized conduction and encoding mechanisms of neuronal network. This study may provide an approach that realizes the quantization of the acupuncture technique and the prediction of the encoding.

针灸是中医的重要组成部分,其作用机制研究对手法量化、标准化具有重要的理论意义和实用价值；针刺神经元网络的编解码是阐释针刺机制的重要科学问题之一，但目前研究结果鲜见报道。首次提出从神经网络角度刻画针刺的编码机制，探索神经元固有特性对网络编码影响的机制。将实验数据分析、模型分析相结合描述神经元固有特性对其输入输出特性的影响、输入输出特性对同步的影响及同步对网络编码影响的规律。利用状态空间实现多阵列针刺电信号的分类和相关分析，量化针刺输入与神经元网络的响应特性；用前馈网络模拟针刺电信号传导通路，采用虚拟网络刻画神经元固有特性与针刺输入、突触输入相互作用的机制及对同步的影响；用相关性特征量化针刺神经元网络同步编码规律；将实验数据、统计分析和计算模型相结合，建立具有生物物理特征的神经元和前馈网络模型，刻画神经元固有特性对网络同步和编码的影响机制。本研究将为针刺手法量化、标准化提供一条新思路。

研究神经元固有特性对神经网络编码的影响机制具有重要意义，本项目首次从神经网络角度刻画针刺的编码机制，对于科学研究针刺的作用机理提供了理论基础，为针刺手法量化、标准化提供新思路。. 通过设计动物足三里穴实验，获取脊髓背根、背角、丘脑核团和皮层神经电信号，提出了一种新的基于模型的分类算法，解决了多神经元放电同步重合问题。利用状态空间和非线性分析，发现了针刺能够影响神经元的放电阈值、分形和相关性等特性。基于针刺响应的概率模型，将针刺效应量化为被激活的内在状态，通过贝叶斯解码实现了针刺神经元编码量化。.基于实验测得的神经元放电时间序列，并加入具有生物物理意义的LIF模型，在模型计算和神经元形态之间取得一个平衡，然后运用贝叶斯解码估计放电特征参数，建立了针刺作用下神经元的放电响应模型。研究单神经元固有特性对放电编码的影响机制，建立了神经元类型、树突特性和适应电流与动作电位能量效率、神经元计算以及放电起始机制之间的生物物理联系，得到了神经元离子通道噪声对放电可靠性的促进规律。.将基于误差反传的学习机制引入到放电神经元网络模型中，研究了神经元固有特性对网络功能的影响，得到了神经元固有特性是影响网络稳定编码信息和功能鲁棒性的关键因素。将非线性系统可控可观的概念应用于神经元网络，分析并发现了神经元固有特性与神经元网络可控可观性之间的关系。将针刺信号等效为不同类型的神经元输入，研究随机噪声对神经元网络放电时空特性的影响，发现噪声诱导的随机共振现象。随机共振是神经系统对针刺信号进行检测和传递的重要方式，是针刺神经网络编码的重要方式之一。.依据针刺信号传导通路的生理结构特征，建立了前馈神经元网络模型，研究针刺信号的传导规律，得到了神经元固有特性对同步编码和频率编码机制的影响。构建了针刺传导通路动态特性的实时估计系统和智能针刺量化机器人系统，实现了针刺手法量化和针刺效应量化。