Tackling EEG signal classification with least squares support vector machines: A sensitivity analysis study

The electroencephalogram (EEG) signal captures the electrical activity of the brain and is an important source of information for studying neurological disorders. The proper analysis of this biological signal plays an important role in the domain of brain-computer interface, which aims at the construction of communication channels between human brain and computers. In this paper, we investigate the application of least squares support vector machines (LS-SVM) to the task of epilepsy diagnosis through automatic EEG signal classification. More specifically, we present a sensitivity analysis study by means of which the performance levels exhibited by standard and least squares SVM classifiers are contrasted, taking into account the setting of the kernel function and of its parameter value. Results of experiments conducted over different types of features extracted from a benchmark EEG signal dataset evidence that the sensitivity profiles of the kernel machines are qualitatively similar, both showing notable performance in terms of accuracy and generalization. In addition, the performance accomplished by optimally configured LS-SVM models is also quantitatively contrasted with that obtained by related approaches for the same dataset.     

颅脑手术可视化中手术场空间与图像空间之间的四点配准算法

手术场空间与图像空间之间的配准是影响颅脑手术可视化系统精度的一个主要因素。给定两个三维空间之间的对应两组点（ｐｉ）和（ｐ＾１ｉ），ｉ＝１，２，…，Ｎ。两个空间中点之间的关系可由，ｐ′ｉ＝Ｒｐｉ＋Ｔ表示，其中Ｒ是旋转矩阵，Ｔ是平移矢量。本文提供了一种基于矩阵单值分解（ＳＶＤ）求解Ｒ和Ｔ的最小二乘解的四点算法，并对算法进行了推导，该算法与迭代相相比，具有速度快，用点少的优点，尤其适用于实际应用的颅脑手     

Plasma phospholipid dysregulation in patients with cystathionine-β synthase deficiency

Background & aimsPatients with cystathionine β-synthase deficiency (CBSD) exhibit high circulating levels of homocysteine and enhanced lipid peroxidation. We have characterized the plasma lipidome in CBSD patients and related lipid abnormalities with reactions underlying enhanced homocysteine levels.Methods and resultsUsing an ultra-high-performance liquid chromatography-electrospray ionization-quadrupole-time of flight-mass spectrometry method, plasma lipids were determined with an untargeted lipidomics approach in 11 CBSD patients and 11 matched healthy subjects (CTRL). Compared to CTRL, CBSD patients had a higher medium and long-chain polyunsaturated fatty acids (PUFA) content in phosphatidylethanolamine (PE) and lysophosphatidylethanolamine (LPE) species (p < 0.02), and depletion of phosphatidylcholine (PC; p = 0.02) and of lysophosphatidylcholine (LPC; p = 0.003) species containing docosahexaenoic acid (DHA), suggesting impaired phosphatidylethanolamine-N-methyltransferase (PEMT) activity. PEMT converts PE into PC using methyl group by S-adenosylmethionine (SAM) thus converted in S-adenosylhomocysteine (SAH). Whole blood SAM and SAH concentrations by liquid chromatography tandem mass spectrometry were 1.4-fold (p = 0.015) and 5.3-fold (p = 0.003) higher in CBSD patients than in CTRL. A positive correlation between SAM/SAH and PC/PE ratios (r = 0.520; p = 0.019) was found.ConclusionsA novel biochemical abnormality in CBSD patients consisting in depletion of PC and LPC species containing DHA and accumulation of PUFA in PE and LPE species is revealed by this lipidomic approach. Changes in plasma SAM and SAH concentrations are associated with such phospholipid dysregulation. Given the key role of DHA in thrombosis prevention, depletion of PC species containing DHA in CBSD patients provides a new direction to understand the poor cardiovascular outcome of patients with homocystinuria.     

使用最小二乘支持向量机技术预测传染病发病率的研究

目的将最小二乘支持向量机(LS-SVM)技术应用到传染病预测中,寻找更加理想的预测结果。方法以某市1991—2002年乙型肝炎(乙肝)月发病率数据建立最小二乘支持向量机预测模型,对2003年1—6月的月发病率进行预测。结果 IS-SVM预测值分别为0.709 9,0.668 1,0.502 5,0.685 1、0.578 5,0.773 7,通过与径向基函数(RBF)神经网络模型和累积式自回归动平均模型(ARIMA)预测结果进行比较,预测精度明显高于RBF网络模型和ARIMA模型,相对误差明显减少,仅为ARIMA模型的23.62%,RBF网络模型的54.69%。结论 LS-SVM模型对乙肝发病率的预测精度更高,效果更好,也验证了支持向量机方法预测能力出色的理论优点,证明了支持向量机技术在传染病预测领域同样有着良好的表现。     

Body composition analysis of obesity and hepatic steatosis in mice by relaxation compensated fat fraction (RCFF) MRI

Purpose:

To develop and validate a quantitative magnetic resonance imaging (MRI) methodology for phenotyping animal models of obesity and fatty liver disease on 7T small animal MRI scanners.

Materials and Methods:

A new MRI acquisition and image analysis technique, relaxation‐compensated fat fraction (RCFF), was developed and validated by both magnetic resonance spectroscopy and histology. This new RCFF technique was then used to assess lipid biodistribution in two groups of mice on either a high‐fat (HFD) or low‐fat (LFD) diet.

Results:

RCFF demonstrated excellent correlation in phantom studies (R² = 0.99) and in vivo compared to histological evaluation of hepatic triglycerides (R² = 0.90). RCFF images provided robust fat fraction maps with consistent adipose tissue values (82% ± 3%). HFD mice exhibited significant increases in peritoneal and subcutaneous adipose tissue volumes in comparison to LFD controls (peritoneal: 6.4 ± 0.4 cm³ vs. 0.7 ± 0.2, P < 0.001; subcutaneous: 14.7 ± 2.0 cm³ vs. 1.2 ± 0.3 cm³, P < 0.001). Hepatic fat fractions were also significantly different between HFD and LFD mice (3.1% ± 1.7% LFD vs. 27.2% ± 5.4% HFD, P = 0.002).

Conclusion:

RCFF can be used to quantitatively assess adipose tissue volumes and hepatic fat fractions in rodent models at 7T. J. Magn. Reson. Imaging 2012;35:837–843. © 2011 Wiley Periodicals, Inc.     

非参数Monotonic回归及其应用

本文针对不满足经典回归条件的一对变量，讨论并介绍了非参数Ｍｏｎｏｔｏｎｉｃ回归的建模思想、条件与方法。通过实例说明，Ｍｏｎｏｔｏｎｉｃ回归比最小二乘回归拟合误差小，模型精度高，计算简单，有一定的实用价值。     

Scatplot: A computer program for determination of binding parameters of non-linear non-cooperative ligand-substrate interactions

The importance of computer-assisted analysis of a non-linear binding phenomenon through Scatchard equation has been widely acknowledged. While several user-friendly softwares [LIGAND, SCTFIT, ALLFIT] are available for determining the binding parameters of nonlinear Scatchard phenomenon, there is no easily available software covering the class of phenomena described by the McGhee and von Hippel formalism [J. Mol. Biol.86, 469–489 (1974)]. We report here user-friendly software, SCATPLOT, developed in Turbo BASIC, for the numerical estimation of binding parameters of a non-cooperative ligand-substrate interaction doing best fit to the experimental data on the basis of McGhee and von Hippel equation for a nonlinear Scatchard plot. A new parameter has also been incorporated to guide the process of least square analysis and subsequent determination of binding parameters.     

偏最小二乘-紫外分光光度法测定布渣叶中浸出物含量

目的:提出一种快速测定布渣叶浸出物含量的新方法。方法:运用偏最小二乘法(PLS)建立紫外光谱(UV)与浸出物含量测定值之间的多元校正模型,对未知样品进行含量预测。结果:校正模型相关系数(R2)为0.964 1。验证集预测平均相对偏差为1.73%。结论:PLS-UV法具有分析速度快、预测结果准确等优点,适合对组成复杂的中药浸出物进行快速分析,可用于布渣叶药材浸出物的快速测定。     

Importance of advanced analytical techniques and methods for food quality control and pollution analysis for more sustainable future in the least developed countries

Chemistry education plays a central role in solving the world's pressing societal issues by equipping and inspiring chemists. However, chemistry education and research in the least developed countries (LDCs) is challenging due to a lack of physical infrastructure and human resources. Among a range of issues and challenges, food and herbal adulteration, microplastics pollution, and chemical hazards are more common in the LDCs, and these issues would impede the achievement of the United Nations sustainable development goals (SDGs). These societal issues are frequently highlighted in the academia and social media of the LDCs, and we believe that at least some aspects of these issues could be solved by advancement in chemistry education and research. However, the current chemistry education and research efforts are inadequate to address these societal issues in these countries. In this article, we have summarized the present scenario and offered examples where developments in chemistry education, particularly advancement in analytical techniques and methods, will provide technological solutions to these issues. Our analysis revealed that sophisticated analytical laboratories are unavailable or unaffordable in the LDCs. The governments of these countries are facing extreme challenges in their pursuit of sustainable development primarily due to a lack of financial resources and shortage of skilled personnel. Revised chemistry curricula and sophisticated analytical testing facilities are urgently required in higher education institutions (HEIs) in the LDCs in helping society achieve the SDGs. A concerted effort between policymakers, chemical societies, funding agencies, chemists, and industries is required to advance chemistry education.