Hindmarsh-Rose神经元模型在小世界网络里的非线性耦合

研究了Hindmarsh-Rose神经元模型在小世界网络里的非线性耦合问题.通过改变邻居半径m和重连概率P来观察模型的同步问题.同步误差主要是通过解网络里的微分方程来实现的,从而揭示了该神经元模型在小世界网络里的同步规律.     

耦合神经网络的分形边界

提出一个被常数刺激驱动的耦合神经网络模型，在这个网络的参数空间中能出现具有分形结构的阶梯，讨论发现这阶梯是一临界边界．跨越它，系统的其中一层网络的总活性从部分兴奋向完全兴奋过度．数值计算分析了这阶梯具有分形结构的原因。     

最近邻耦合动力学网络的同步

对最近邻节点耦合网络的同步进行了研究,得出了达到同步所满足的条件;并进行了数值模拟．     

非线性耦合Hindmarsh-Rose神经元同步迁移

研究了Hindmarsh-Rose神经元在非线性耦合作用下的同步问题,进一步讨论了非线性耦合与线性耦合作用下同步的迁移,采用灰度图的方式表示出Laypunov指数,判定了混沌系统的同步;然后给定暂态过程值,通过计算误差函数给出二级耦合下线性耦合与非线性耦合同步参数区域,发现非线性耦合对线性耦合有比较强的调制作用.     

噪声诱导电耦合FHN神经元同步

通过噪声对2个电耦合神经元施加影响,讨论了噪声强度和耦合强度对耦合神经元系统同步的影响.研究发现,噪声和耦合的双重作用下,系统可以更快地达到同步.当耦合强度g=1.0>0.5时,电耦合连接的系统相当稳定,噪声并不能彻底改变系统状态.     

Division and subtraction by distinct cortical inhibitory networks in vivo

Brain circuits process information through specialized neuronal subclasses interacting within a network. Revealing their interplay requires activating specific cells while monitoring others in a functioning circuit. Here we use a new platform for two-way light-based circuit interrogation in visual cortex in vivo to show the computational implications of modulating different subclasses of inhibitory neurons during sensory processing. We find that soma-targeting, parvalbumin-expressing (PV) neurons principally divide responses but preserve stimulus selectivity, whereas dendrite-targeting, somatostatin-expressing (SOM) neurons principally subtract from excitatory responses and sharpen selectivity. Visualized in vivo cell-attached recordings show that division by PV neurons alters response gain, whereas subtraction by SOM neurons shifts response levels. Finally, stimulating identified neurons while scanning many target cells reveals that single PV and SOM neurons functionally impact only specific subsets of neurons in their projection fields. These findings provide direct evidence that inhibitory neuronal subclasses have distinct and complementary roles in cortical computations.     

GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus

The hippocampus, in particular the neocortex-hippocampus-neocortex circuit, is widely believed to be crucial in memory. Information flow in this circuit is strongly influenced by relatively sparse afferents derived from subcortical centres, such as the septum, involved in arousal, emotions and autonomic control. A powerful mechanism, by which numerically small inputs can produce profound effects, is feed-forward inhibition, that is, the activation of local inhibitory interneurons, which, in turn, control the activity of large populations of principal cells in the hippocampus. An example is the cholinergic input to the hippocampus from the septum, which is likely to be involved in feed-forward operations. Here, we demonstrate the existence of a circuit underlying another powerful mechanism of subcortical control of hippocampal information processing. We show that GABA-containing afferents originating in the septum innervate most of the GABA-containing interneurons in the hippocampus, making many synaptic contacts with each of them. Activation of the GABA-containing neurons in the septum is likely to lead to disinhibition of the principal neurons in the hippocampal formation and so this pathway is probably crucial in the induction of hippocampal electrical activity patterns, and may be involved in NMDA (N-methyl-D-aspartate) receptor-mediated functions, such as memory, in a permissive manner.     

时滞耦合随机神经网络全局指数同步

应用矩阵的Kronecker积、线性矩阵不等式及随机微分方程等有关工具,研究了一类具离散时滞与分布时滞的随机神经网络耦合大系统的均方全局指数同步问题.综合考虑了参数的不确定性、分布时滞、随机扰动、非线性耦合等的影响,并弱化了相关文献中激励函数为严格Lipshtz或Sigma型的假定,得到该耦合大系统均方全局指数同步的一些充分性判据.所得到的结果易于通过Matlab线性矩阵不等式工具箱(LMI)进行检验.     

基于平均场的全局耦合神经集群的去同步控制

针对一些疾病中出现的神经元放电的病理性同步,本文提出了一种去同步控制方法.在神经元受到电流和噪声的刺激引发放电时,引入平均场对神经元集群进行点对点的全局耦合,将会使神经集群达到同步状态.在此过程中将平均场进行滤波以及线性处理反馈回神经集群,可以控制集群去同步.通过对集群参数与控制器参数分别研究,表明这种控制方法对噪声,耦合强度,控制器参数都具有鲁棒性.而且此方法有实用性,易临床实现,可以用作深度大脑刺激的新颖控制方法来治疗精神疾病.     

混合时滞随机神经网络全局渐近同步

研究了一类具混合时滞神经网络耦合大系统的同步问题.应用矩阵的Kronecker积与线性矩阵不等式,得到耦合大系统同步的一些充分性判据,所得到的结果易于通过Matlab线性矩阵不等式工具箱(LMI)进行检验,所讨论的神经网络具有参数的不确定性、分布时滞、随机扰动以及非线性耦合等,而且弱化了相关文献中激励函数为严格Lipshitz或Sigma型的假定,使所得结果更为精细.     

陶瓷涂层材料多模态数据表征学习

陶瓷涂层具有耐高温、耐腐蚀、耐磨损等特性, 其热膨胀系数和热导率等参数与其性能息息相关. 为解决陶瓷涂层性能实验成本高、测试困难等问题, 提出了陶瓷涂层材料多模态数据表征学习的性能预测方法. 首先利用高斯混合模型虚拟样本生成(Gaussian mixture model virtual sample generation, GMMVSG)算法生成符合真实陶瓷涂层数据分布的样本来扩充数据集; 其次利用卷积神经网络 VGG16 对陶瓷涂层的显微结构图像数据进行特征提取, 利用 TabNet 对结构化数据进行特征提取, 将提取到的图像数据特征与结构化数据特征融合; 最终根据多模态数据表征建立基于K-最近邻(K-nearest neighbor, KNN)、支持向量机回归(support vector regression, SVR)和多层感知机(multi-layer perceptron, MLP) 3 种机器学习算法的预测模型, 对陶瓷涂层的性能指标, 即热膨胀系数和热导率进行了预测. 实验结果表明: 提出的多模态数据表征学习模型的预测结果要优于单模态数据表征学习模型, 其中基于 MLP 算法训练的多模态数据表征学习模型对陶瓷涂层性能的预测效果最好; 在测试集中, 对陶瓷涂层热膨胀系数预测的平均绝对误差(mean absolute error, MAE)和均方误差(mean square error, MSE)分别为 0.026 6 和 0.001 7, 对热导率预测的 MAE 和 MSE 分别为 0.017 9 和 0.000 7. 所提出的陶瓷涂层材料多模态数据表征学习方法有效融合了结构化数据与非结构化数据, 联合学习了各模态数据的潜在共享信息, 成功提升了对陶瓷涂料层材料性能预测的准确度.     

时延脉冲耦合神经网络在最大流问题中的应用

 在时延脉冲耦合神经网络(DPCNN-Delay Pulse Coupled Neural Network)的基础上,提出了mDC-DPCNN(modified Dual Channels DPCNN)模型,并将其成功地用于最大流网络的求解.该方法通过快速并行计算,经过反复迭代有限步后可以得到所求解的网络,与传统的算法相比体现了较好的优势.     

节点含时滞的输出耦合复杂网络的自适应脉冲同步

针对节点含时滞的输出耦合动态网络模型,基于Lyapunov稳定性理论和时滞脉冲微分方程理论,设计了复杂网络同步自适应脉冲控制器,给出网络同步的充分条件,保证动态网络渐近同步于任意指定的网络中单独节点的状态,进而实现该网络的同步.理论推导和数值结果均表明该同步方法的有效性及可行性.     

Channel noise-induced phase transition of spiral wave in networks of Hodgkin-Huxley neurons

The phase transition of spiral waves in networks of Hodgkin-Huxley neurons induced by channel noise is investigated in detail.All neurons in the networks are coupled with small-world connections,and the results are compared with the case for regular networks,in which all neurons are completely coupled with nearest-neighbor connections.A statistical variable is defined to study the collective behavior and phase transition of the spiral wave due to the channel noise and topology of the network.The effect of small-world connection networks is described by local regular networks and long-range connection with certain probability p.The numerical results confirm that (1) a stable rotating spiral wave can be developed and maintain robust with low p,where the breakup of the spiral wave and turbulence result from increasing the probability p to a certain threshold;(2) appropriate intensity of the optimized channel noise can develop a spiral wave among turbulent states in small-world connection networks of H-H neurons;and (3) regular connection networks are more robust to channel noise than small-world connection networks.A spiral wave in a small-world network encounters instability more easily as the membrane temperature is increased to a certain high threshold.     

带有非线性耦合节点和时滞的复杂网络的指数同步研究

在耦合矩阵不对称并且假设耦合指数已知的条件下,利用 Lyapunov Krasovskii稳定理论设计了一个非线性反馈控制器,使得复杂网络达到全局指数同步。通过数值仿真,证明了该同步定理的可行性和有效性。