神经的

The sutureless techniques with fibrin glue or laser welding seems to be only an adjunction in nerve repair because the connection quality between nerve stumps is unreliable.

鉴于运动轴突对肌皮瓣再神经化的意义，区分感觉和运动神经束的选择性神经缝合，在游离神经血管肌肉移植术中取得了较好的效果。

METHODS: The extracts of forebrain, midbrain,afterbrain and cerebellum of neonatal mice were prepared, and the NSCs of newborn mice were cultured in vitro. Neurospheres were observed, immunocytochemical staining of characteristic protein, nestin, and MTT assay were performed to identify NSCs and their proliferative properties.

制备新生鼠的前脑、中脑、后脑及小脑组织提取液，体外培养新生鼠神经干细胞，观察神经球的生长情况，采用MTT细胞活性检测法和特异性蛋白质免疫细胞化学染色鉴定神经干细胞的增殖能力。

The top of neural tube forms basic encenphalic shape,namely endbrain, interbrain, midbrain and afterbrain. The abnormal closure of neural tube will result in neural tube detects. Encephalocoele is a common congenital malformation whose mechanism was no known. NO participates in the physiologic process of brain.

胚胎脑发育的早期阶段主要是神经板隆起、弯曲、闭合形成神经管，神经管的头部从前向后依次发育形成脑的基本组织形态，即端脑、间脑、中脑、后脑，在神经管闭合过程中出现异常就会导致NTD。

It is related to the variational inequality problem, bilinear programming, nonlinear equation. So there is actual meaning and theories value to study how to solve it.In recent years, many algorithms for solving the generalized linear complementarity problem are put forward. But they basically belong to the traditional iterative methods, and may not be efficient since the computing time required for a solution is greatly dependent on the dimension, structure of the problem and the complexity of the algorithm used. Different from the traditional algorithms, the neural network has many advantages on calculations and real-time applications for its inherent massive parallelism and electric circuit implementation.

近年来，人们对广义线性互补问题提出了许多算法，但这些方法基本上属于传统的迭代法，由于它们的计算时间极大地依赖于问题的规模、维数以及所使用的算法，因此很难满足实时并行的要求，与传统数值方法不同，由于内在的并行分布处理信息的特点及电路实现的潜能，神经网络有着许多计算上的优势和实时性的应用，自Hopfield提出著名的人工神经网络模型-Hopfield神经网络，并将其成功应用于优化问题后，用神经网络求解优化问题得到了相当深入的研究，并取得了许多重要的成果。

Aimed at the characteristic of multiple types of faults possibly happened in nuclear power plant, large scale of training sample, and requirement of quick and accurate diagnosing, after the analysis of large sample number generated by large fault type exceeds limitation of 64K paragraph in DOS and large number of local minmum in error surface, measurements of same format of data file, Win95/NT operating system platform, resetting weight learning rate, dynamic training set in quick learning algorithm and improvement on quick learning algorithm using homotopy method which can avoid local minmum points in error surface have been adopted to ensure quickly and effective process of the course of neural network's training and testing.

针对核动力装置可能发生的故障种类多，训练样本规模大，故障诊断需要快速准确的特点等，分析了样本多和故障种类多产生的训练样本量超出DOS段大小和网络误差曲面上局部极小值多的情况，采取了使用相同格式数据文件、Win95/NT操作系统平台，对快速学习算法采用重置神经网络权值学习率和动态训练集、并采用能够有效克服网络误差曲面上局部极小点的同伦方法对学习算法进行改进等措施保证神经网络训练测试过程的快速、有效地进行；针对装置发生的故障须快速、准确诊断的需要，分析了故障的产生对装置参数变化的影响及操纵员对故障诊断的基础，在核动力装置发生故障时参数曲线的变化量与正常运行时参数曲线的变化量存在明显差异的基础上，提出采用参数曲线的变化量作为神经网络的输入，并围绕参数变化量的方法采用二次曲线拟合滤波求变化量和阈值技术来保证神经网络得到精确的装置参数变化量，从而得到准确的诊断结果。

Methods The rat MSCs were isolated primarily from the femurs and tibias of the Wistar rats. MSCs were cultured, proliferated and purified by passage culture. After being induced by bFGF, cultuered MSCs were divided into experimental groups (GM1 group, GDNF group and GDNF+ GM1 group) and control group.

取雄性Wistar大鼠股骨和胫骨骨髓，进行MSCs的体外培养、传代扩增及纯化。bFGF预诱导24 h后，依据加入的神经营养因子不同分为单唾液酸四己糖神经节苷脂（GM1）组、胶质源性神经营养因子组和GDNF+GM1组，以及对照组。

Also traditional neurograms for representing the neural activities within auditory nerves are used as the corresponding evidences for supporting this framework.

本文并用传统代表听觉神经里神经活动的神经图来作为支持此架构的相对证据。

Firstly we propose a layered model of neural computing systems, and conclude the requirements of neural computation to neurochip.

首先作者提出神经计算系统的层次模型，系统、全面地分析神经计算对神经芯片的需求。

Based on the selection of similar structure, a method for predicting the rock drillability of wildcat well is proposed.

该方法首先应用已钻相似构造井的地震资料、测井资料以及岩心测试资料，建立岩石可钻性钻前预测遗传神经网络模型，然后将神经网络与遗传算法有机结合起来，以神经网络理论为基础，利用遗传算法优化隐含层神经元个数和网络连接权值，最后利用野猫井的地震资料进行岩石可钻性钻前预测。

The main contents are as follows:Firstly, starting with the general situation of soil erosion and the harms caused by it, the causes leading to the local soil erosion problem are analyzed comprehensively in the paper. And connecting with the measures taking place, sticking points towards the career of soil and water conservation are expatiated upon.Secondly, Back-Propagation Neural Network, One of Artificial Neural Network is used to set up a modal about the connection of the soil erosion modulus and seven factors impacts on it, such as, rainfall, rainfall largest intensity in 30 minutes, runoff coefficient, vegetation cover percent, rate of granule, rate of physical viscidity-clay, the rate of organic matter. Through the comparison with linear regression model, the second regression model, the Chinese Soil Loss Equation, it illustrated that BP Neural Network modal is more accurate than the other three modals in forecasting the mount of soil erosion, and the BP Neural Network will have some applicability in forecasting in soil erosion.

本文以霍山县作为皖西大别山区的典型区域，主要研究了以下内容：(1)从介绍霍山县土壤侵蚀状况以及所造成的危害入手，全面分析了导致当地土壤侵蚀发生的原因，并结合当地采取的水土保持相关措施，阐述了当地水土保持工作的症结所在；(2)结合上土市水土保持试验站多年实测资料和2005年实验资料，应用BP神经网络理论，建立了次降雨土壤侵蚀量与次降雨量、最大30min雨强、径流系数、植被覆盖度等因子之间关系的模型，并通过BP神经网络的预测模型与一次回归模型、二次回归模型、CSLE模型之间的对比分析，说明了建立的BP神经网络模型在土壤侵蚀预测可以取得较回归模型和CSLE模型更高的精度，也说明了BP神经网络理论在土壤侵蚀预报中具有一定的适用性。

For a big chunk of credit-card losses; the number of filings (and thus charge-off rates) would be rising again, whether

年美国个人破产法的一个改动使得破产登记急速下降，而后引起了信用卡大规模的亏损。

Eph. 4:23 And that you be renewed in the spirit of your mind

弗四23 而在你们心思的灵里得以更新