基于脑电样本熵的测谎分析

测谎分析在刑讯侦查和法律审判中具有重要意义.为了区分是否说谎,根据脑电信号的非线性特征,本文首次使用非线性动力学的样本熵方法分析30名受试者处于诚实和说谎两种状态时脑电信号的样本熵值.研究发现：受试者处于诚实状态时的熵值波动范围明显小于说谎状态下的波动范围,更重要的是说谎时的熵值显著高于说实话时的熵值,表明样本熵可以区分诚实和说谎两种不同状态下的脑电信号,该研究为基于脑电的测谎提供了一种新的途径.     

基于相位延迟指数的脑功能网络及测谎研究

在脑认知科学领域,越来越多的研究开始专注于利用不同导联脑电信号之间的相互依赖关系来研究大脑整体认知功能.相位延迟指数可有效减少由容积导体引起的误差,该方法已被广泛应用.而基于图论的脑网络研究方法在测谎方面还少见报道.本文通过对30名（诚实和说谎）受试者的脑电信号进行网络拓扑分析,将网络参数作为判别指标,使用支持向量机对实验数据进行分类.研究发现,两组受试者的小世界指标表现出显著的统计学差异,且得到较高的测谎准确率,结果证明了利用相位延迟指数方法进行图论分析的测谎有效性.     

基于互信息的脑网络及测谎研究

互信息分析方法是基于信息论提出的一种描述两信号间信息交互情况的算法,其在脑电信号领域的有效性已得到了充分证实.针对当前测谎方法中脑电信号特征提取困难以及大脑整体认知功能分析在脑认知科学研究中越来越被重视的情况,本文首次将互信息分析方法应用到脑电测谎领域中,使用互信息量化大脑各节点之间的相关性,对计算结果进行统计分析,选取出在两类人群中具有显著性差异的电极对的互信息作为分类特征,进行模式识别,得到了99.67%的准确率.这一结果表明,互信息分析方法是一种有效的脑功能连接分析方法,为基于脑电信号连接分析的测谎研究提供了一种新的途径.另外,对说谎与诚实两类受试者的大脑功能网络的分析结果表明：处于说谎状态时,大脑的额叶、顶叶、颞叶及枕叶之间协同实现谎言功能,并在躯体行为所对应的脑区与其他脑区的连接上也表现出相对诚实组的显著性差异,以上结果均有助于进一步揭示谎言的神经活动机制.     

利用相锁值算法的脑电相同步测谎研究

相锁值（Phase Locking Value,PLV）是由相同步的概念下提出一种描述不同信号相关性（同步性）的算法,在脑电信号领域,其有效性已经得到了验证.本文针对当前测谎方法中脑电信号特征提取困难的问题,首次将相锁值的算法应用到脑电测谎领域中,研究谎言脑活动下不同脑区之间的相关性,通过相关性发现谎言的认知机制,并利用该相关性作为特征,使用支持向量机对说谎者和诚实者的两类信号进行模式识别,得到了88.50%的准确率,提出的方法验证了PLV在测谎应用中的有效性,为基于脑电信号的测谎提供了一种新的途径.     

虚拟现实视觉体验对事件相关电位影响的研究

虚拟现实技术具有广阔应用前景,但它对大脑信息处理及认知影响尚不清楚。本文结合头皮脑电,设计了平面和虚拟现实两种模式的视觉任务,对比分析虚拟现实对事件相关电位（ERP）的影响,探索沉浸式视觉体验过程中大脑的认知加工过程。参考电极的选择是研究事件相关电位的关键,为获得更客观的脑电信息,本文首先使用参考电极标准化技术将记录的脑电信号的参考近似转换为理想的零点,经过必要的预处理,再采用叠加平均的方法从中提取ERP。结果发现,与平面模式相比,在虚拟现实模式下ERP成分中没有出现显著的P300成分,这可能与大脑产生疲劳感相关;虚拟现实模式下N100成分的潜伏期提前、幅值增大,反映了虚拟现实环境更容易引起注意,并使人产生沉浸感。      

基于脑功能网络和样本熵的脑电信号特征提取

针对脑-机接口(BCI)研究中采用单一特征对运动想象脑电信号(EEG)识别率不高的问题,该文提出一种结合脑功能网络和样本熵的特征提取方法.根据事件相关同步/去同步(ERS/ERD)现象以及皮层与肢体运动想象间的对侧映射机制,选取小波包变换消噪重构后的μ节律脑电信号,用左侧27个通道、右侧27个通道分别对左半球脑区和右半球脑区构建脑功能网络,计算网络的平均节点度和平均聚集系数作为运动想象的脑功能网络特征,并结合C3,C4通道μ节律的样本熵构筑分布性和指向性相结合的特征向量.选用支持向量机(SVM)对左右手运动想象脑电信号进行分类,结果表明基于脑功能网络和样本熵的特征提取方法能够实现更优的分类效果,分类准确率最高可达90.27％.     

左右手运动想象脑电信号的特征分析1

针对运动想象的两种思维任务的脑电信号,本文提出了一种基于能量的特征提取方法。该方法利用想象运动中,脑电信号Mu/Beta节律事件相关同步化/去同步化特性,采用BCI2003竞赛数据,对C3/C4通道脑电信号进行时域分析,分析得到特征较为明显的脑电信号时间段。     

基于PCANet和SVM的谎言测试研究

主成分分析网络（Principal Component Analysis Network,PCANet）是基于深度学习理论的一种非监督式的特征提取方法,它克服了手工提取特征的缺点,目前其有效性仅仅在图像处理领域中得到了验证。本文针对当前谎言测试方法中脑电信号特征提取困难的缺点,首次将PCANet方法应用到一维信号的特征提取领域,并对测谎实验的原始脑电信号提取特征,然后使用支持向量机（Support Vector Machine,SVM）将说谎者和诚实者的两类信号进行分类识别,将实验结果和其它分类器及未使用特征提取的分类效果进行了比较。实验结果显示相对未抽取任何特征的方法,提出的方法PCANet－SVM可以获得更高的训练和测试准确率,表明了PCANet方法对于脑电信号特征提取的有效性,也为基于脑电信号的测谎提供了一种新的途径。     

基于脑功能网络和样本熵的脑电信号特征提取

针对脑-机接口(BCI)研究中采用单一特征对运动想象脑电信号(EEG)识别率不高的问题,该文提出一种结合脑功能网络和样本熵的特征提取方法。根据事件相关同步/去同步(ERS/ERD)现象以及皮层与肢体运动想象间的对侧映射机制,选取小波包变换消噪重构后的\begin{document}$ \mu$\end{document}节律脑电信号,用左侧27个通道、右侧27个通道分别对左半球脑区和右半球脑区构建脑功能网络,计算网络的平均节点度和平均聚集系数作为运动想象的脑功能网络特征,并结合C3, C4通道节律的样本熵构筑分布性和指向性相结合的特征向量。选用支持向量机(SVM)对左右手运动想象脑电信号进行分类,结果表明基于脑功能网络和样本熵的特征提取方法能够实现更优的分类效果,分类准确率最高可达90.27%。     

基于小波的模拟自然阅读事件相关电位的单次提取

选择合适的信号处理方法从脑电信号中提取用户的信息和命令,是改进脑计算机接口通讯速率的方法之一.由于自发脑电信号(electroencephalograph,EEG)是非高斯有色噪声,且自发脑电信号的频谱不规则、与有效信号的频谱相重叠,传统的滤波方法难以取得较好的效果.基于白噪声与有效信号的小波变换模极大值在不同尺度下的传播行为不一样的原则,本文提出了一个基于自回归模型和小波变换多尺度分析的模拟自然阅读事件相关电位的单次提取方法.经对实际脑电信号处理的实验表明,该算法能较好的提取脑电信号.     

EEG-based assessment of driver cognitive responses in a dynamic virtual-reality driving environment

Accidents caused by errors and failures in human performance among traffic fatalities have a high death rate and become an important issue in public security. They are mainly caused by the failures of the drivers to perceive the changes of the traffic lights or the unexpected conditions happening accidentally on the roads. In this paper, we devised a quantitative analysis for assessing driver's cognitive responses by investigating the neurobiological information underlying electroencephalographic (EEG) brain dynamics in traffic-light experiments in a virtual-reality (VR) dynamic driving environment. The VR technique allows subjects to interact directly with the moving virtual environment instead of monotonic auditory and visual stimuli, thereby provides interactive and realistic tasks without the risk of operating on an actual machine. Independent component analysis (ICA) is used to separate and extract noise-free ERP signals from the multi-channel EEG signals. A temporal filter is used to solve the time-alignment problem of ERP features and principle component analysis (PCA) is used to reduce feature dimensions. The dimension-reduced features are then input to a self-constructing neural fuzzy inference network (SONFIN) to recognize different brain potentials stimulated by red/green/yellow traffic events, the accuracy can be reached 87% in average eight subjects in this visual-stimuli ERP experiment. It demonstrates the feasibility of detecting and analyzing multiple streams of ERP signals that represent operators' cognitive states and responses to task events.     

基于3D U-Net的轻量级脑肿瘤分割网络北大核心CSCD

针对现有脑肿瘤核磁共振成像(magnetic resonance imaging, MRI)分割神经网络的参数量和计算量较大且对肿瘤区域小目标分割精度不高的问题,提出一种改进的轻量级脑肿瘤分割网络MF-RES2Net(multiple fiber residual-like networks)。该网络以3D U-Net为基础架构,将多纤模块(multi-fiber, MF)和类残差模块(RES2)相结合代替传统卷积模块。MF将特征图像的通道进行混合,增加了通道间信息的交流融合;RES2将通道均分,单通道的卷积结果相加到相邻通道,在扩大图像感受野的同时保留了细节特征,同时降低网络参数量。此外,为改善数据不平衡问题,提出一种改进的加权损失函数,提高了网络对小目标的分割精度。将MF-RES2Net在BRATS 2019数据集进行验证,完整肿瘤、核心肿瘤和增强肿瘤分割的平均Dice系数分别为89.98%、84.02%、77.62%,参数量和浮点数分别为3.16 M和16.24 G,结果表明:该网络在降低参数量和计算量的同时进一步提升了分割性能,有效地降低了网络运行时的设备要求。     

An energy aware event-driven routing protocol for cognitive radio sensor networks

Cognitive radio sensor network (CRSN) is an intelligent and reasonable combination of cognitive radio technology and wireless sensor networks. It poses significant challenges to the design of topology maintenance techniques due to dynamic primary-user activities, which in turn decreases the data delivery performance of the network as well as it’s lifetime. This paper aims to provide a solution to the CRSN clustering and routing problem using an energy aware event-driven routing protocol (ERP) for CRSN. Upon detection of an event, the ERP determines eligible nodes for clustering according to local positions of CRSN nodes between the event and the sink and their residual energy levels. Cluster-heads are selected from the eligible nodes according to their residual energy values, available channels, neighbors and distance to the sink. In ERP, cluster formation is based on relative spectrum awareness such that channels with lower primary user appearance probability are selected as common data channels for clusters. For data routing, ERP employs hop-by-hop data forwarding approach through the CHs and primary/secondary gateways towards the sink. Through extensive simulations, we demonstrate that the proposed ERP provides with better network performances compared to those of the state-of-the-art protocols under a dynamic spectrum-aware data transmission environment.     

A Low Complexity Suboptimal Energy-Based Detection Method for SISO/MIMO Channels

In this paper, we propose an actually novel and simple method for detection of transmitted symbols in MIMO channels. This method is based on the energy level of the received signals. At the receiver, we assume the knowledge of channel state information which can be estimated by different methods, e.g. by sending pilots. So, we can determine all possible levels of energy. This computation of energy levels is done only once for the quasi-static channels. Energy of the received signals is a criterion by which we can estimate the transmitted symbols. Detection of transmitted signal is made based on the nearest energy level and the points which lie on it. In other words, we have restricted our search space to a new smaller space with different levels of energy. Simulation results confirm approximately the same performance between the maximum-likelihood detector and the proposed approach especially in high signal-to-noise ratios with a remarkable reduction in the computational complexity.     

用于语音控制的低资源关键词检索系统

基于深度神经网络的低资源条件下关键词检索已经取得了很大的进展，但这些方法仍旧需要较多的参数才能保证模型的精度。为了进一步减少模型的参数量，本文将Squeeze-and-Excitation网络和深度可分离卷积应用在关键词检索任务中。首先利用Squeeze-and-Excitation网络对不同特征通道之间的相互依赖关系建模的能力进一步提升模型的精度，然后通过将标准卷积替换为深度可分离卷积来有效的减少模型所需要的参数。在谷歌语音命令数据集上的实验证明我们的模型可以在保证高精度的同时把参数量限制在一定的范围内。      

Spatiotemporal emotion recognition based on 3D time-frequency domain feature matrix

The research of emotion recognition based on electroencephalogram (EEG) signals often ignores the relatedinformation between the brain electrode channels and the contextual emotional information existing in EEG signals,which may contain important characteristics related to emotional states. Aiming at the above defects, aspatiotemporal emotion recognition method based on a 3-dimensional (3D) time-frequency domain feature matrixwas proposed. Specifically, the extracted time-frequency domain EEG features are first expressed as a 3D matrixformat according to the actual position of the cerebral cortex. Then, the input 3D matrix is processed successivelyby multivariate convolutional neural network (MVCNN) and long short-term memory (LSTM) to classify theemotional state. Spatiotemporal emotion recognition method is evaluated on the DEAP data set, and achievedaccuracy of 87.58% and 88.50% on arousal and valence dimensions respectively in binary classification tasks, aswell as obtained accuracy of 84.58% in four class classification tasks. The experimental results show that 3D matrixrepresentation can represent emotional information more reasonably than two-dimensional (2D). In addition,MVCNN and LSTM can utilize the spatial information of the electrode channels and the temporal context information of the EEG signal respectively.     

普遍意义下的干涉仪通道间相位差测量精度分析

传统的干涉仪通道间相位差测量模型是建立在单频电磁波的基础之上的,在应用于调制信号测量时,其精度分析结果有时会出现较大偏差。针对这一问题,利用调制信号相关接收的方法来提取干涉仪通道间的相位差信息,推导并建立了普遍意义下的干涉仪通道间相位差测量精度的理论计算式,并且新的精度计算式对传统计算式具有向下兼容性,并通过仿真对其有效性进行了验证,这对于电子侦察的测向和定位等应用中与干涉仪相关的测量精度分析提供了新的参考。