Robust Nonlinear Mapping of Soil Contamination Using Support Regression

Spatial data analysis mapping and visualization is of great importance in various fields: environment, pollution, natural hazards and risks, epidemiology, spatial econometrics, etc. A basic task of spatial mapping is to make predictions based on some empirical data (measurements). A number of state-of-the-art methods can be used for the task: deterministic interpolations, methods of geostatistics: the family of kriging estimators (Deutsch and Journel, 1997), machine learning algorithms such as artificial neural networks (ANN) of different architectures, hybrid ANN-geostatistics models (Kanevski and Maignan, 2004; Kanevski et al., 1996), etc. All the methods mentioned above can be used for solving the problem of spatial data mapping. Environmental empirical data are always contaminated/corrupted by noise, and often with noise of unknown nature. That's one of the reasons why deterministic models can be inconsistent, since they treat the measurements as values of some unknown function that should be interpolated. Kriging estimators treat the measurements as the realization of some spatial randomn process. To obtain the estimation with kriging one has to model the spatial structure of the data: spatial correlation function or (semi-)variogram. This task can be complicated if there is not sufficient number of measurements and variogram is sensitive to outliers and extremes. ANN is a powerful tool, but it also suffers from the number of reasons. of a special type ? multiplayer perceptrons ? are often used as a detrending tool in hybrid (ANN+geostatistics) models (Kanevski and Maignank, 2004). Therefore, development and adaptation of the method that would be nonlinear and robust to noise in measurements, would deal with the small empirical datasets and which has solid mathematical background is of great importance. The present paper deals with such model, based on Statistical Learning Theory (SLT) - Support Vector Regression. SLT is a general mathematical framework devoted to the problem of estimation of the dependencies from empirical data (Hastie et al, 2004; Vapnik, 1998). SLT models for classification - Support Vector Machines - have shown good results on different machine learning tasks. The results of SVM classification of spatial data are also promising (Kanevski et al, 2002). The properties of SVM for regression - Support Vector Regression (SVR) are less studied. First results of the application of SVR for spatial mapping of physical quantities were obtained by the authorsin for mapping of medium porosity (Kanevski et al, 1999), and for mapping of radioactively contaminated territories (Kanevski and Canu, 2000). The present paper is devoted to further understanding of the properties of SVR model for spatial data analysis and mapping. Detailed description of the SVR theory can be found in (Cristianini and Shawe-Taylor, 2000; Smola, 1996) and basic equations for the nonlinear modeling are given in section 2. Section 3 discusses the application of SVR for spatial data mapping on the real case study - soil pollution by Cs137 radionuclide. Section 4 discusses the properties of the modelapplied to noised data or data with outliers.

Diagnostic and prognostic value of cerebral 31P magnetic resonance spectroscopy in neonates with perinatal asphyxia.

The impact of depressed neonatal cerebral oxidative phosphorylation for diagnosing the severity of perinatal asphyxia was estimated by correlating the concentrations of phosphocreatine (PCr) and ATP as determined by magnetic resonance spectroscopy with the degree of hypoxic-ischemic encephalopathy (HIE) in 23 asphyxiated term neonates. Ten healthy age-matched neonates served as controls. In patients, the mean concentrations +/- SD of PCr and ATP were 0.99 +/- 0.46 mmol/L (1.6 +/- 0.2 mmol/L) and 0.99 +/- 0.35 mmol/L (1.7 +/- 0.2 mmol/L), respectively (normal values in parentheses). [PCr] and [ATP] correlated significantly with the severity of HIE (r = 0.85 and 0.9, respectively, p < 0.001), indicating that the neonatal encephalopathy is the clinical manifestation of a marred brain energy metabolism. Neurodevelopmental outcome was evaluated in 21 children at 3, 9, and 18 mo. Seven infants had multiple impairments, five were moderately handicapped, five had only mild symptoms, and four were normal. There was a significant correlation between the cerebral concentrations of PCr or ATP at birth and outcome (r = 0.8, p < 0.001) and between the degree of neonatal neurologic depression and outcome (r = 0.7). More important, the outcome of neonates with moderate HIE could better be predicted with information from quantitative 31P magnetic resonance spectroscopy than from neurologic examinations. In general, the accuracy of outcome predictability could significantly be increased by adding results from 31P magnetic resonance spectroscopy to the neonatal neurologic score, but not vice versa. No correlation with outcome was found for other perinatal risk factors, including Apgar score.

Raman spectroscopy and microspectroscopy of reactive dyes on cotton fibres: analysis and detection limits

A collaborative study on Raman spectroscopy and microspectrophotometry (MSP) was carried out by members of the ENFSI (European Network of Forensic Science Institutes) European Fibres Group (EFG) on different dyed cotton fabrics. The detection limits of the two methods were tested on two cotton sets with a dye concentration ranging from 0.5 to 0.005% (w/w). This survey shows that it is possible to detect the presence of dye in fibres with concentrations below that detectable by the traditional methods of light microscopy and microspectrophotometry (MSP). The MSP detection limit for the dyes used in this study was found to be a concentration of 0.5% (w/w). At this concentration, the fibres appear colourless with light microscopy. Raman spectroscopy clearly shows a higher potential to detect concentrations of dyes as low as 0.05% for the yellow dye RY145 and 0.005% for the blue dye RB221. This detection limit was found to depend both on the chemical composition of the dye itself and on the analytical conditions, particularly the laser wavelength. Furthermore, analysis of binary mixtures of dyes showed that while the minor dye was detected at 1.5% (w/w) (30% of the total dye concentration) using microspectrophotometry, it was detected at a level as low as 0.05% (w/w) (10% of the total dye concentration) using Raman spectroscopy. This work also highlights the importance of a flexible Raman instrument equipped with several lasers at different wavelengths for the analysis of dyed fibres. The operator and the set up of the analytical conditions are also of prime importance in order to obtain high quality spectra. Changing the laser wavelength is important to detect different dyes in a mixture.

Determination of the size of quantum dots by fluorescence spectroscopy.

There has been a lack of quick, simple and reliable methods for determination of nanoparticle size. An investigation of the size of hydrophobic (CdSe) and hydrophilic (CdSe/ZnS) quantum dots was performed by using the maximum position of the corresponding fluorescence spectrum. It has been found that fluorescence spectroscopy is a simple and reliable methodology to estimate the size of both quantum dot types. For a given solution, the homogeneity of the size of quantum dots is correlated to the relationship between the fluorescence maximum position (FMP) and the quantum dot size. This methodology can be extended to the other fluorescent nanoparticles. The employment of evolving factor analysis and multivariate curve resolution-alternating least squares for decomposition of the series of quantum dots fluorescence spectra recorded by a specific measuring procedure reveals the number of quantum dot fractions having different diameters. The size of the quantum dots in a particular group is defined by the FMP of the corresponding component in the decomposed spectrum. These results show that a combination of the fluorescence and appropriate statistical method for decomposition of the emission spectra of nanoparticles may be a quick and trusted method for the screening of the inhomogeneity of their solution.

Brain energy metabolism measured by (13) C magnetic resonance spectroscopy in vivo upon infusion of [3-(13) C]lactate.

The brain uses lactate produced by glycolysis as an energy source. How lactate originated from the blood stream is used to fuel brain metabolism is not clear. The current study measures brain metabolic fluxes and estimates the amount of pyruvate that becomes labeled in glial and neuronal compartments upon infusion of [3-(13) C]lactate. For that, labeling incorporation into carbons of glutamate and glutamine was measured by (13) C magnetic resonance spectroscopy at 14.1 T and analyzed with a two-compartment model of brain metabolism to estimate rates of mitochondrial oxidation, glial pyruvate carboxylation, and the glutamate-glutamine cycle as well as pyruvate fractional enrichments. Extracerebral lactate at supraphysiological levels contributes at least two-fold more to replenish the neuronal than the glial pyruvate pools. The rates of mitochondrial oxidation in neurons and glia, pyruvate carboxylase, and glutamate-glutamine cycles were similar to those estimated by administration of (13) C-enriched glucose, the main fuel of brain energy metabolism. These results are in agreement with primary utilization of exogenous lactate in neurons rather than astrocytes. © 2014 Wiley Periodicals, Inc.

Assessment of metabolic fluxes in the mouse brain in vivo using 1H-[13C] NMR spectroscopy at 14.1 Tesla.

(13)C magnetic resonance spectroscopy (MRS) combined with the administration of (13)C labeled substrates uniquely allows to measure metabolic fluxes in vivo in the brain of humans and rats. The extension to mouse models may provide exclusive prospect for the investigation of models of human diseases. In the present study, the short-echo-time (TE) full-sensitivity (1)H-[(13)C] MRS sequence combined with high magnetic field (14.1 T) and infusion of [U-(13)C6] glucose was used to enhance the experimental sensitivity in vivo in the mouse brain and the (13)C turnover curves of glutamate C4, glutamine C4, glutamate+glutamine C3, aspartate C2, lactate C3, alanine C3, γ-aminobutyric acid C2, C3 and C4 were obtained. A one-compartment model was used to fit (13)C turnover curves and resulted in values of metabolic fluxes including the tricarboxylic acid (TCA) cycle flux VTCA (1.05 ± 0.04 μmol/g per minute), the exchange flux between 2-oxoglutarate and glutamate Vx (0.48 ± 0.02 μmol/g per minute), the glutamate-glutamine exchange rate V(gln) (0.20 ± 0.02 μmol/g per minute), the pyruvate dilution factor K(dil) (0.82 ± 0.01), and the ratio for the lactate conversion rate and the alanine conversion rate V(Lac)/V(Ala) (10 ± 2). This study opens the prospect of studying transgenic mouse models of brain pathologies.

An overview of the metabolic differences between Bradyrhizobium japonicum 110 bacteria and differentiated bacteroids from soybean (Glycine max) root nodules: An in vitro 13C-and 31P-nuclear magnetic resonance spectroscopy study

Bradyrhizobium japonicum is a symbiotic nitrogen-fixing soil bacteria that induce root nodules formation in legume soybean (Glycine max.). Using 13C- and 31P-nuclear magnetic resonance (NMR) spectroscopy, we have analysed the metabolite profiles of cultivated B.japonicum cells and bacteroids isolated from soybean nodules. Our results revealed some quantitative and qualitative differences between the metabolite profiles of bacteroids and their vegetative state. This includes in bacteroids a huge accumulation of soluble carbohydrates such as trehalose, glutamate, myo-inositol and homospermidine as well as Pi, nucleotide pools and intermediates of the primary carbon metabolism. Using this novel approach, these data show that most of the compounds detected in bacteroids reflect the metabolic adaptation of rhizobia to the surrounding microenvironment with its host plant cells.