Interactions between working memory and selective attention

Event-related potential (ERP) was used to examine the interactions between working memory and se- lective attention. We combined two unrelated tasks, one requiring working memory and the other se- lective attention, which were performed by some undergraduates. The ERP results revealed that both congruent and incongruent stimuli in the selective attention task evoked an N400 component, reaching the peak point at around 500 ms. The N400 evoked by incongruent stimuli was more negative than that of congruent, which indicated the difference of semantic N400. Furthermore, working memory load had a significant influence on the N400 evoked by selective attention task in parietal region. And working memory load showed difference in the ERPs of working memory retrieval in central and parietal regions. The ERPs of probe under high working memory load were more positive from 350 to 550 ms post-stimulus; however, stimulus type of selective attention had no influence on working memory re- trieval. The present study shows that working memory does not play a major role in the selective at- tention, especially in ignoring distracter, but it influences the performance of the selective attention as the background. The congruency of target and distracter in the selective attention task does not influ- ence the working memory retrieval.     

ERP dissociation and connection between implicit and explicit memory

ERP dissociations between implicit and explicit memory have been confirmed by a large amount of evidence in theories of human memory. However, similarities between the two forms of memory have scarcely been studied. A possibility is that while implicit memory and explicit memory have independ- ent components, they might additionally have shared components. To explore this question, an ERP experiment was conducted with a study-to-test paradigm, in which participants performed a "shallow" (color) study task or a "deep" (pleasant) study task, followed by either a lexical decision (implicit) test (Section 1) or a recognition (explicit) test (Section 2). An interference task was performed concurrently with either the encoding or the retrieval phase of the memory task for encoding interference condition or retrieval interference condition. We compared ERP signatures of implicit and explicit memory as a function of depth of processing or interference. Under the action of the same variables, 300―500 ms old/new ERP effects of implicit and explicit memory showed the same trend. These effects maybe all link with a perceptual representational system. 500―700 ms old/new ERP effects of the two memories were dissociated. They may probably reflect voluntary and involuntary recollection respectively. These results suggested that implicit and explicit memories are not completely independent of each other, but have both independent and shared components.     

不同年级学生的工作记忆研究

从小学三年级、五年级、初中二年级、高中二年级中分别挑选被试对象各80人，进行工作记忆广度测量。结果发现：1．工作记忆广度的发展随着年级的增长而发展；2．各年级学生的两项工作记忆广度测试成绩与语文、数学成绩有较高的正相关。结论：工作记忆与学生的学业成绩关系密切，即工作记忆是有效学习的前提。     

巴甫洛夫学说的逻辑漏洞和理论假设的片面性——从伦理学和现代神经科学看巴甫洛夫学说的终结

巴甫洛夫创立了大脑高级神经活动学说。80多年来,该学说在哲学、生理学和心理学界产生了广泛的影响。人们已往对于该学说的批评经常被忘却或置之脑后,致使该学说一直畅行无阻地流行而没有得到认真地清算。笔者经过长期的研究和思考,认为巴氏学说在逻辑和理论前提上存在着严重的问题,即对于大脑高级神经活动基本机能的设想不仅是片面的,而且在逻辑上有严重的漏洞。     

基于马氏体重定向的铁磁形状记忆合金数值分析

基于铁磁形状记忆合金的磁场诱发应变来自磁场诱发马氏体变体的特征,本文在建立铁磁形状记忆合金的力—磁—热耦合理论的基础上,对铁磁形状记忆合金的磁致应变效应、磁滞效应以及磁致应力等特性进行数值分析,以揭示模型的合理性和适用性.     

无界区域上具有记忆项的随机波动方程的拉回吸引子的存在性

在无界区域Rⁿ(n≤3)上研究了如下具有线性记忆项的随机波动方程的渐进行为u_tt+αu_t－k(0)Δu+λu+f(x,u)－∫^∞₀k′(s)Δu(t－s)ds=g(x)+h(x)dωdt。其中, 当n=3时非线性项f具有次临界增长率, 当n=1,2时f可具有任意增长率。运用解的一致估计方法在H¹(Rⁿ)×L²(Rⁿ)×M¹(Rⁿ)上证明了对应的随机动力系统拉回吸引子的存在性。     

IEEE 1394事务层接口的设计与实现

本文设计的IEEE 1394事物层接口连接了PLB（Processor Local Bus）总线和1394链路层模块,主要实现1394事物层模块数据的发送和接收.首先对事物层接口进行系统结构设计,然后对PLB从接口、FIFO接口和主机访问寄存器接口这3个子模块分别进行逻辑设计,最后采用硬件描述语言对设计进行可综合描述,并使用软件QuestaSim进行仿真验证.该设计使事务层能够跨时钟域同链路层进行数据包的传输,并进行有效的数据位宽转换,提高传输效率.经过仿真验证,结果满足设计要求,表明IEEE 1394事物层接口能够正常发送和接收等时及异步数据包.     

基于语义分析的恶意JavaScript代码检测方法

JavaScript是一种动态脚本语言,被用于提高网页的交互能力.然而攻击者利用它的动态性在网页中执行恶意代码,构成了巨大威胁.传统的基于静态特征的检测方式难以检测经过混淆后的恶意代码,而基于动态分析检测的方式存在效率低等问题.本文提出了一种基于语义分析的静态检测模型,通过提取抽象语法树的词法单元序列特征,使用word2vec训练词向量模型,将生成的序列向量特征输入到LSTM网络中检测恶意JavaScript脚本.实验结果表明,该模型能够高效检测混淆的恶意JavaScript代码,模型的精确率达99.94%,召回率为98.33%.     

基于磁记忆法的焊接缺陷检测技术研究

研究焊缝缺陷处的磁记忆信号变化规律. 建立了地磁场下焊缝的磁偶极子模型,并模拟了其在有、无缺陷时的变化曲线,根据模型曲线的变化情况总结出了焊缝处存在缺陷时磁场强度的变化规律. 通过检测自制焊接试件的磁记忆信号与模型进行对比. 缺陷处的检测曲线与模型模拟的曲线具有很好的一致性. 建立的磁偶极子模型能够反映实际焊缝处的磁记忆信号情况,可以对焊缝处缺陷进行评价.     

用水迷宫建立大鼠空间辨别性学习记忆模型

目的探索脑形态与脑功能结合的动物模型建立方法.方法用水迷宫训练大鼠建立空间辨别性学习记忆模型.结果用水迷宫每天训练10次,连续7 d即可达到预定的模型标准;在建立模型的过程中发现在训练的第5天有学习的平台期出现.结论水迷宫训练实验能有效的使大鼠建立空间辨别性学习记忆能力.