Finer discrimination of brain activation with local multivariate distance

The organization of human brain function is diverse on different spatial scales. Various cognitive states are always represented as distinct activity patterns across the specific brain region on fine scales. Conventional univariate analysis of functional MRI data seeks to determine how a particular cognitive state is encoded in brain activity by analyzing each voxel separately without considering the fine-scale patterns information contained in the local brain regions. In this paper, a local multivariate distance mapping (LMDM) technique is proposed to detect the brain activation and to map the fine-scale brain activity patterns. LMDM directly represents the local brain activity with the patterns across multiple voxels rather than individual voxels, and it employs the multivariate distance between different patterns to discriminate the brain state on fine scales. Experiments with simulated and real fMRI data demonstrate that LMDM technique can dramatically increase the sensitivity of the detection for the fine-scale brain activity patterns which contain the subtle information of the experimental conditions.     

基于任务态和静息态功能核磁共振信号的抑郁症识别

为了提高抑郁症识别的准确率,将功能核磁共振成像的任务态数据和静息态数据相结合,建立基于数据驱动的模型以提取识别特征.在没有任何先验知识的条件下,采用独立成分分析法提取任务态数据和静息态数据的独立成分;然后,利用相关遍历分析法获取功能信号集,利用频谱分析法识别并获取功能信号成分;最后,将功能信号成分作为贝叶斯分类器的特征...     

急性脑梗死偏瘫患者踝关节运动功能障碍的fMRI研究

目的:利用功能磁共振成像(fMRI)技术,观察急性脑梗死患者的运动相关皮层的激活图,探讨卒中后踝功能障碍及恢复的脑功能相关性改变.方法:采用1.5T磁共振成像系统对10例急性脑梗死患者和5例正常志愿者进行BOLD-fMRI检查,采用健侧、患侧踝关节背伸、趾屈运动作为刺激任务,比较正常志愿者与急性脑梗死患者踝关节运动时脑部激活情况的异同,并采用LI为指标,观察急性脑梗死患者健侧、患侧踝关节运动时脑部激活的差异.结果:正常对照组踝关节运动主要激活对侧SM1区,激活区形态呈斑片状或者不规则状,此外对侧PM区、SMA区也有不同程度的激活.急性脑梗死患者健侧踝关节运动主要激活对侧SM1区,此外对侧PM区、SMA区、颞叶、枕叶、小脑半球、额叶等也有不同程度的激活,与正常对照组激活区相似;患侧踝关节运动(主观运动)时除了对侧半球SM1区激活之外,同侧SM1区、PM区也可见不同程度的激活,对侧半球的激活体积大部分大于同侧.根据公式计算出LI值,结果显示健侧、患侧踝关节运动LI值之间存在显著性差异(P<0.05).结论:fMRI可以客观的显示急性脑梗死后患者大脑运动皮层的激活情况,提示存在脑功能代偿与重组,fMRI作为监测和研究急性脑梗死后运动功能恢复的有用工具,在临床治疗中具有重要的指导意义.     

Localization of the brain calculation function area with MRI

The aim of this study is to define the anatomical localization of corresponding brain function area during calculating. The activating modes in brain during continuous silent calculating subtraction and repeated silent reading multiplication table were compared and investigated. Fourteen volunteers of right-handedness were enrolled in this experiment. The quite difference of reaction modes in brain area during the two modes of calculation reveal that there are different processing pathways in brain during these two operating actions. During continuous silent calculating, the function area is localized on the posterior portion of superior and middle gyrus of frontal lobe and the lobule of posterior parietal lobe (P < 0.01,T = 5.41). It demonstrates that these function areas play an important role in the performance of calculation and working memory. Whereas the activating of visual cortex shows that even in mental arithmetic processing the brain action is having the aid of vision and visual space association.     

脑卒中后椎体束华勒变性的功能磁共振研究

目的应用扩散张量与磁共振波谱技术,对急性脑卒中患者进行分析,探讨其反映患者病情变化的价值.方法取一侧脑卒中患者26例,健康志愿者30例.在常规扫描后再进行DTI,MRS扫描,采用NIHSS和BI评分评价患者功能,并定期复查.分析患者梗死病灶、大脑脚ADC值、FA值及MRS的代谢产物数值变化.结果 1）病灶FA值低于正常值,与NIHSS呈正相关,与BI呈负相关.脑干FA值减小,提示神经纤维出现华勒变性;2）病灶NAA下降,低于正常侧,P〈0.01;病灶侧Cr,Cho较正常侧改变不明显,P〉0.05;大脑脚NAA 1个月后降低.结论对脑卒中患者进行DTI与MRS扫描,能较早反映患者病情严重程度,还能判断预后,并指导临床康复治疗.     

基于局部二元模式算子的人脸性别分类方法

提出了两种基于局部二元模式(Local Binary Pattern,LBP)算子的人脸性别分类方法:级联LBP方法和boosting LBP方法.前一种方法遵循从局部到整体的级联策略,使用LBP算子对由小波分解得到的细节图像进行特征提取,以达到扩充特征提取范围和增强所提取特征的有效性的目的,随后采用自适应加权机制对人脸图像的各个分块赋以不同的权值.后一种方法采用可变尺寸的子窗口对人脸图像进行扫描,在扫描所得的每个子窗口中,使用LBP算子对该子窗口提取LBP直方图.计算样本图像的LBP直方图和模板的LBP直方图之间的度量,并由此构建弱分类器集.利用Adaboost算法选取最有效的若干个弱分类器集组合成为强分类器.进行了三个基于LBP算子的人脸性别分类实验,实验所使用的训练集和测试集皆选自FERET人脸数据库.实验结果证明:LBP算子能有效地从人脸图像中提取出针对人脸性别分类的特征,并可以达到人脸性别分类的目的.所提出的两种基于LBP算子的方法可以有效的解决传统LBP方法所存在的特征提取范围有限、加权机制客观性不足等问题.     

A new area in the brain associated with learning and memory

A new subdivision, named marginal division (MrD)> consisting of spindle-shaped neurons, has been identified at the caudomedial margin of the neostriatum in the brains of the rat, cat, monkey and human. It is distinguishable from the rest of striatum by special neural connections and many intensely expressed neuropeptides and some monoamines in the fibers, terminals and neuronal somata. Three-dimensional reconstruction of the rat brain reveals that the MrD is a flat pan-shaped area between the neostriatum and globus pal-lidus. Chemical lesions of bilateral MrD in rats will result in severely impaired learning and memory functions, as was demonstrated by double blind Y-maze test. The function of MrD has been shown to be associated with learning and memory by functional magnetic resonance imaging (fMRl) technique in human brain in vivo . Functional neuronal connections are observed between the MrD and hippocampus, amygdala, as well as the basal nucleus of Meynert by chemically induced c-Fos immunohistochemical staining. MrD is a newly discovered part and a universal structure in the neostriatum of the mammalian brain. MrD might very possibly play an important role in processes of the learning and memory.     

联合局部和全局稀疏表示的磁共振图像重建方法

针对在压缩传感中独立使用全局或局部稀疏字典所分别导致的图像细节或整体图像结构信息的丢失,提出了一种联合利用局部和全局稀疏约束来捕捉磁共振图像细节和整体结构信息的磁共振图像重建算法。该算法首先从特定的磁共振图像中训练出稀疏字典,然后利用该字典进行局部稀疏编码。其次,利用预定义的全局字典来加强磁共振图像的全局稀疏性。最后,在局部和全局稀疏的共同约束下,利用非线性共轭梯度算法来对重建模型进行求解。整个重建过程可以重复迭代以逐步改善重建质量。实验结果表明:当下采样因子达到10时,相比于字典学习算法(dictionary learning MRI,DLMRI),提出的算法在重建质量上可以提高1-6dB。     

基于β估计校正值的低场功能磁共振成像分析方法

针对低场功能磁共振成像(fMRI)非脑组织信号变化的影响,提出一种有效的图像分析方法以改进大脑功能定位的结果.综合利用基于形变模型的图像分割算法和基于像素形态测量学(voxel-basedmorphometry,VBM)的统计分析方法提取脑组织,生成显式掩模,并构建分析掩模图像以修正一般线性模型的β参数估计值,使得VBM的统计推断只针对大脑组织,消除了非脑组织信号变化的影响.应用实例表明,这种方法可有效去除背景、头骨及头皮信号变化引起的伪激活区,提高了低场fMRI分析的可靠性和准确性.     

稳态视觉诱发电位频率与相位特性的脑电研究

以往的研究发现,头皮脑电在枕区电极处所测得的稳态视觉诱发响应的幅度随着刺激频率的变化而变化。为了解释这一现象,该文采用一个3层的边界元头模型以及基于该模型的求逆算法,通过真实脑电数据分析,发现位于左、右侧枕区的皮层源在稳态视觉诱发响应上存在着时间差,由此提出假设:左、右侧枕区的皮层源在不同刺激频率下,在时间相位上存在着差异,这是造成不同刺激频率下脑电响应幅度存在差异的原因之一。通过仿真实验进一步验证了该假设的合理性。     

Survey of nutrients and quality assessment of crab paste by 1H NMR spectroscopy and multivariate data analysis

Crab paste is a traditional and popular food for people in the coastal area of China.Visual inspection is currently the only method of quality assessment.We developed a new analytical method of a comprehensive survey of nutrients and quality assessment for crab paste based on a combination of 1 H NMR spectroscopy with multivariate data analysis.The aqueous extract of crab paste was dominated by amino acids,sugars,carboxylic acids,nucleotides and amines(including 19 first-time-reported compounds such as choline,uracil and guanosine).Two grades of crab paste had significant compositional differences in terms of amino acids,lactate,N-acetylglutamate,choline,dimethylamine,uridine,1-methylnicotinamide and 2-pyridienmethanol.These results provided important information on the grade-dependence of crab-paste composition,and demonstrated that NMR-MDA was effective not only for the comprehensive survey of nutrients,but also for quality assessment of crab paste.     

多元统计过程控制方法的平方预测误差分析

作为多元统计过程控制方法中的常用统计量,平方预测误差（ Ｓ Ｐ Ｅ）的变化规律有待深入研究。介绍了主元分析建模方法,推导了 Ｓ Ｐ Ｅ均值公式,分析了 Ｓ Ｐ Ｅ均值和过程变量均值向量、协方差矩阵之间的解析关系,用来自 ３ 阶液位系统的仿真数据验证了分析的结果。给出了 Ｓ Ｐ Ｅ随过程变量均值向量、协方差矩阵变化而变化的若干规律,说明了这些规律在生产过程监控应用中的意义。     

Detection of dynamic brain networks modulated by acupuncture using a graph theory model

Neuroimaging studies involving acute acupuncture manipulation have already demonstrated significant modulatory effects on wide limbic/paralimbic nuclei, subcortical gray structures and the neocortical system of the brain. Due to the sustained effect of acupuncture, however, knowledge on the organization of such large-scale cortical networks behind the active needle stimulation phase is lacking. In this study, we originally adopted a network model analysis from graph theory to evaluate the functional connectivity among multiple brain regions during the post-stimulus phase. Evidence from our findings clearly supported the existence of a large organized functional connectivity network related to acupuncture function in the resting brain. More importantly, acupuncture can change such a network into a functional state underlying both pain perception and modulation, which is exhibited by significant changes in the functional connectivity of some brain regions. This analysis may help us to better understand the long-lasting effects of acupuncture on brain function, as well as the potential benefits of clinical treatments.     

Functional brain laterality for sequential movements： Impact of transient practice

The impact of learning on brain functional laterality has not been systematically investigated. We employed an event-related functional magnetic resonance imaging combined with a delayed sequential movement task to investigate brain activation pattern and laterality during a transient practice in 12 subjects. Both hemispheres, involving motor areas and posterior parietal cortex, were engaged during motor preparation and execution, with larger activation volume in the left hemisphere than in the right. Activation volume in these regions significantly decreased after a transient practice, with more reduction in the right hemisphere resulting increase in left lateralization. The theoretical implications of these findings are discussed in relation to the physiological significance of brain functional laterality.     

基于Granger因果检验和PCA的脑网络效应连接方法

为了提高脑功能效应连接网络检测的可靠性,提出了一种基于Granger因果关系检验和主成分分析(PCA)的功能磁共振(fMRI)数据效应连接方法.该方法首先通过PCA提取感兴趣区域内fMRI信号的时间主成分,以此特征作为时间参考信息,然后计算参考信息与大脑其余每个体素之间的Granger因果关系,并映射到全脑,形成Granger因果图(GCM).理论推导阐明了所提方法的有效性.采用该方法研究人脑运动功能脑区在手动任务下的效应连接GCM,结果验证了运动功能神经网络理论.     

基于主成分分析和Granger因果检验的手动任务脑功能效应连接

提出了一种基于Granger因果关系检验和主成分分析的fMRI数据效应连接方法,并将两种方法有效联合研究手动任务下激活皮层的因果效应连接。该方法首先由统计参数图法(SPM)组分析得到任务激活脑空间图,选取感兴趣区域(ROI);其次通过主成分分析提取ROI内fMRI信号的时间主成分,以此特征作为时间参考信息;最后计算参考区域与大脑其余每个体素之间的Granger因果关系,并映射到全脑,形成Granger因果图(Granger Causality Map,GCM)。理论和试验结果阐明了提出方法的有效性及可靠性。运用该方法研究人脑运动功能脑区在手动任务下的效应连接GCM,验证了运动功能神经网络理论。     

多变量统计过程控制在浮选生产过程中的应用

针对浮选过程的被控对象复杂、数学模型不确定以及控制要求高等特点,提出一种用于阳离子浮选过程的多变量统计过程控制药剂流量方法,使系统具备一定的自学习和自组织的能力。仿真研究和现场运行试验表明,提出的控制方法能够满足浮选过程控制的要求。     

基于支持向量机的人脸姿态判定

对于多姿态人脸检测中的姿态判定问题,提出了一种基于支持向量机(SVM)的人脸姿态判定算法。将人脸姿态划分成6个类别,从一个多姿态人脸库中手工标定出1800幅人脸图像作为训练样本集,分别训练基于支持向量分类(SVC)和基于支持向量回归(SVR)2种姿态分类器。另外标定出300幅人脸图像作为测试样本。SVC方法和SVR方法分别取得了1.67%和3.33%的分类错误率。其中SVC方法的分类效果明显优于在传统方法中效果最好的人工神经元网络(ANN)方法(分类错误率为3.33%)。对比实验结果表明,SVM方法对于解决姿态判定问题是很有效的。     

SPM的数学基础及其在脑功能成像研究中的应用

探讨SPM的数学基础,给出了SPM中利用一般线性模型(General LinearModel,GLM)构建统计参数图的方法;为了适应处理功能磁共振成像(functionalMagnetic Resonance Imaging,fMR I)数据的需要,对一般线性模型做了扩展,推导出扩展的GLM的检验统计量和自由度.最后将扩展的一般线性模型应用于一个具体的fMR I实验,说明了方法的有效性.     

成年高血压与血压正常者颅脑MRI的对比观察

目的 :通过MR影像的分析探讨高血压与慢性脑损害关系。方法 :搜集颅脑MR扫描的成年高血压和血压正常者资料 2 37例 ,按年龄段将其MRI表现进行详细对比分析。结果 :本组病例在脑内多数部位 ,其高信号灶均随年龄增大而增多 ;6 0～ 6 9岁年龄段的高血压患者 ,在底节内外囊、放射冠、半卵圆中心、侧脑室体部和前后角周围 ,其高信号灶的出现率高于血压正常组 (P <0 0 5 ) ,并且更容易发生脑室扩大 (P <0 0 5 ) ;脑梗死的发生率与血压正常与否无关 (P >0 0 5 ) ,但随着年龄增大而增高。结论 :高血压患者多在 6 0～ 6 9岁时表现明显脑损害。