Multivariate statistical analysis of diffusion imaging parameters using partial least squares: Application to white matter variations in Alzheimer's disease

Diffusion magnetic resonance imaging (dMRI) is a unique technology that allows the noninvasive quantification of microstructural tissue properties of the human brain in healthy subjects as well as the probing of disease-induced variations. Population studies of dMRI data have been essential in identifying pathological structural changes in various conditions, such as Alzheimer's and Huntington's diseases (Salat et al., 2010; Rosas et al., 2006). The most common form of dMRI involves fitting a tensor to the underlying imaging data (known as diffusion tensor imaging, or DTI), then deriving parametric maps, each quantifying a different aspect of the underlying microstructure, e.g. fractional anisotropy and mean diffusivity. To date, the statistical methods utilized in most DTI population studies either analyzed only one such map or analyzed several of them, each in isolation. However, it is most likely that variations in the microstructure due to pathology or normal variability would affect several parameters simultaneously, with differing variations modulating the various parameters to differing degrees. Therefore, joint analysis of the available diffusion maps can be more powerful in characterizing histopathology and distinguishing between conditions than the widely used univariate analysis. In this article, we propose a multivariate approach for statistical analysis of diffusion parameters that uses partial least squares correlation (PLSC) analysis and permutation testing as building blocks in a voxel-wise fashion. Stemming from the common formulation, we present three different multivariate procedures for group analysis, regressing-out nuisance parameters and comparing effects of different conditions. We used the proposed procedures to study the effects of non-demented aging, Alzheimer's disease and mild cognitive impairment on the white matter. Here, we present results demonstrating that the proposed PLSC-based approach can differentiate between effects of different conditions in the same region as well as uncover spatial variations of effects across the white matter. The proposed procedures were able to answer questions on structural variations such as: "are there regions in the white matter where Alzheimer's disease has a different effect than aging or similar effect as aging?" and "are there regions in the white matter that are affected by both mild cognitive impairment and Alzheimer's disease but with differing multivariate effects?"

A robust localized soft sensor for particulate matter modeling in Seoul metro systems

Developing accurate soft sensors to predict and monitor the indoor air quality (IAQ) of hazardous pollutants that accumulate in underground metro systems is of key importance. The just-in-time (JIT) learning technique possesses a local feature that can track the variations in the dynamic process more effectively, which is different from the traditional soft sensor modeling methods, such as partial least squares (PLS), which models the process in an offline and global way. In this study, a robust soft sensor that combined the JIT learning technique with a least squares support vector regression (LSSVR) method, named JIT-LSSVR, was derived in order to improve the prediction performance of a PM2.5 soft sensor in a subway station. Additionally, in order to eliminate the adverse effects caused by the outliers in the process variables, an outlier detection step was integrated into the JIT-LSSVR modeling procedure. The performance evaluation results demonstrated that the proposed robust JIT-LSSVR soft sensor has the capability to model nonlinear and dynamic subway systems. The root mean square error of the JIT-LSSVR soft sensor was improved by 55% in comparison with that of the LSSVR soft sensor.

Determination of persimmon leaf chloride contents using near-infrared spectroscopy (NIRS)

Early diagnosis of specific chloride toxicity in persimmon trees requires the reliable and fast determination of the leaf chloride content, which is usually performed by means of a cumbersome, expensive and time-consuming wet analysis. A methodology has been developed in this study as an alternative to determine chloride in persimmon leaves using near-infrared spectroscopy (NIRS) in combination with multivariate calibration techniques. Based on a training dataset of 134 samples, a predictive model was developed from their NIR spectral data. For modelling, the partial least squares regression (PLSR) method was used. The best model was obtained with the first derivative of the apparent absorbance and using just 10 latent components. In the subsequent external validation carried out with 35 external data this model reached r(2) = 0.93, RMSE = 0.16% and RPD = 3.6, with standard error of 0.026% and bias of -0.05%. From these results, the model based on NIR spectral readings can be used for speeding up the laboratory determination of chloride in persimmon leaves with only a modest loss of precision. The intermolecular interaction between chloride ions and the peptide bonds in leaf proteins through hydrogen bonding, i.e. N-H···Cl, explains the ability for chloride determinations on the basis of NIR spectra.

A best​ comprehension about the toxicity of phenylsulfonyl carboxylates in Vibrio fischeri using quantitative structure activity/property relationship methods

Aromatic sulfones comprise a class of chemicals used in agrochemical and pharmaceutical industries and as floatation and extractant agents in petrochemical and metallurgy industries. In this study, new QSA(P)R studies were carried out to predict the toxicity against Vibrio fischeri of a set of 52 aromatic sulfones. The same approach was used to evaluate the relationship between these endpoint and the water solubility, another important environmental endpoint. The study resulted in models of good statistical quality and mechanistic interpretation with a possible correlation between the two endpoints, but the toxic effect is also likely to depend on other physicochemical properties. The use of the PLS2, a method not commonly used in QSA(P)R studies, also produced models of greater reliability, and the relationship between the two endpoints was reinforced to some degree. These results are useful for better understanding the process by which these compounds exert their environmental toxicity, thus aiding in the development of industrially useful compounds with less potential environmental damage.

Estimation of melting points of large set of persistent organic pollutants utilizing QSPR approach

The presence of polyhalogenated persistent organic pollutants (POPs), such as Cl/Br-substituted benzenes, biphenyls, diphenyl ethers, and naphthalenes has been identified in all environmental compartments. The exposure to these compounds can pose potential risk not only for ecological systems, but also for human health. Therefore, efficient tools for comprehensive environmental risk assessment for POPs are required. Among the factors vital for environmental transport and fate processes is melting point of a compound. In this study, we estimated the melting points of a large group (1419 compounds) of chloro- and bromo- derivatives of dibenzo-p-dioxins, dibenzofurans, biphenyls, naphthalenes, diphenylethers, and benzenes by utilizing quantitative structure-property relationship (QSPR) techniques. The compounds were classified by applying structure-based clustering methods followed by GA-PLS modeling. In addition, random forest method has been applied to develop more general models. Factors responsible for melting point behavior and predictive ability of each method were discussed.

A protocol for analysing thermal stress in insects using infrared thermography

The study of insect responses to thermal stress has involved a variety of protocols and methodologies that hamper the ability to compare results between studies. For that reason, the development of a protocol to standardize thermal assays is necessary. In this sense, infrared thermography solves some of the problems allowing us to take continuous temperature measurements without handling the individuals, an important fact in cold-blooded organisms like insects. Here, we present a working protocol based on infrared thermography to estimate both cold and heat thermal stress in insects. We analyse both the change in the body temperature of individuals and their behavioural response. In addition, we used partial least squares regression for the statistical analysis of our data, a technique that solves the problem of having a large number of variables and few individuals, allowing us to work with rare or endemic species. To test our protocol, we chose two species of congeneric, narrowly distributed dung beetles that are endemic to the southeastern part of the Iberian Peninsula. With our protocol we have obtained five variables in the response to cold and twelve in the response to heat. With this methodology we discriminate between the two flightless species of Jekelius through their thermal response. In response to cold, Jekelius hernandezi showed a higher rate of cooling and reached higher temperatures of stupor and haemolymph freezing than Jekelius punctatolineatus. Both species displayed similar thermoregulation ranges before reaching lethal body temperature with heat stress. Overall, we have demonstrated that infrared thermography is a suitable method to assess insect thermal responses with a high degree of sensitivity, allowing for the discrimination between closely related species.

Determination of benzo[a]pyrene in cigarette mainstream smoke by using mid-infrared spectroscopy associated with a novel chemometric algorithm

Determination of benzo[a]pyrene (BaP) in cigarette smoke can be very important for the tobacco quality control and the assessment of its harm to human health. In this study, mid-infrared spectroscopy (MIR) coupled to chemometric algorithm (DPSO-WPT-PLS), which was based on the wavelet packet transform (WPT), discrete particle swarm optimization algorithm (DPSO) and partial least squares regression (PLS), was used to quantify harmful ingredient benzo[a]pyrene in the cigarette mainstream smoke with promising result. Furthermore, the proposed method provided better performance compared to several other chemometric models, i.e., PLS, radial basis function-based PLS (RBF-PLS), PLS with stepwise regression variable selection (Stepwise-PLS) as well as WPT-PLS with informative wavelet coefficients selected by correlation coefficient test (rtest-WPT-PLS). It can be expected that the proposed strategy could become a new effective, rapid quantitative analysis technique in analyzing the harmful ingredient BaP in cigarette mainstream smoke.

Green near-infrared determination of copper and mancozeb in pesticide formulations

A green analytical procedure has been successfully developed for the simultaneous determination of copper and mancozeb in phytosanitary products. The method is based on different direct measurements of diffuse reflectance near-infrared (DR-NIR) spectra. Accuracy of the method has been evaluated by comparison of the obtained copper and mancozeb concentrations with those provided by reference methodologies based on titrimetric procedures. The average relative prediction error was 0.7 and 1.6 % for copper and mancozeb, respectively. The evaluation of the greenness of the DR-NIR procedure provided 100 points, which is the maximum value in the Green Certificate ranking, because of the absence of consumed reagents and waste generation and energy consumption lower than 0.1 kWh.

Task-independent effects are potential confounders in longitudinal imaging studies of learning in schizophrenia

Learning impairment is a core deficit in schizophrenia that impacts on real-world functioning and yet, elucidating its underlying neural basis remains a challenge. A key issue when interpreting learning-task experiments is that task-independent changes may confound interpretation of task-related signal changes in neuroimaging studies. The nature of these task-independent changes in schizophrenia is unknown. Therefore, we examined task-independent “time effects” in a group of participants with schizophrenia contrasted with healthy participants in a longitudinal fMRI learning-experiment designed to allow for examination of non-specific effects of time. Flanking the learning portions of the experiment with a task-of-no-interest allowed us to extract task-independent BOLD changes. Task-independent effects occurred in both groups, but were more robust in the schizophrenia group. There was a significant interaction effect between group and time in a distributed activity pattern that included inferior and superior temporal regions, frontal areas (left anterior insula and superior medial gyri), and parietal areas (posterior cingulate cortices and precuneus). This pattern showed task-independent linear decrease in BOLD amplitude over the two scanning sessions for the schizophrenia group, but showed either opposite effect or no activity changes for the control group. There was a trend towards a correlation between task-independent effects and the presence of more negative symptoms in the schizophrenia group.

The strong interaction between group and time suggests that both the scanning experience as a whole and the transition between task-types evokes a different response in persons with schizophrenia and may confound interpretation of learning-related longitudinal imaging experiments if not explicitly considered.

•

A robust method was used to identify task-independent fMRI BOLD changes in a multiday learning experiment in schizophrenia

•

Task-independent effects were apparent in healthy control group and schizophrenia but differed in direction and magnitude

•

In schizophrenia they were greater in magnitude and most prominent in areas of the salience and default mode networks

•

Unless properly accounted for, these effects will compromise precise interpretation of fMRI learning data in schizophrenia.

Chemometrics meets homeopathy--an exploratory analysis of infrared spectra of homeopathic granules

10 homeopathic remedies commercially available (each in 3 dilutions) as sugar granules, where half of them were of organic (and half inorganic) origin were subjected to solid-state infrared spectroscopy, both in middle infrared (ATR-FTIR) and near infrared (NIR) range. Measurements were repeated six times (six days, each sample was measured once in the same day, samples were measured in random order). The obtained spectra was subjected to unsupervised (PCA) and supervised (PLS-DA) chemometric techniques to check any visible differnces in spectral data between homeopathic remedies, including also feature selection approaches. It can be concluded that the only one information encoded in this dataset is the atmospheric drift of spectra between consecutive measurement days. This proves that homeopathy is not "infrared visible" in the case of proper experimental design. These results can be useful in further investigations of possible mechanisms of action of homeopathy (if they exist).

Exploring simple, transparent, interpretable and predictive QSAR models for classification and quantitative prediction of rat toxicity of ionic liquids using OECD recommended guidelines

The present study explores the chemical attributes of diverse ionic liquids responsible for their cytotoxicity in a rat leukemia cell line (IPC-81) by developing predictive classification as well as regression-based mathematical models. Simple and interpretable descriptors derived from a two-dimensional representation of the chemical structures along with quantum topological molecular similarity indices have been used for model development, employing unambiguous modeling strategies that strictly obey the guidelines of the Organization for Economic Co-operation and Development (OECD) for quantitative structure-activity relationship (QSAR) analysis. The structure-toxicity relationships that emerged from both classification and regression-based models were in accordance with the findings of some previous studies. The models suggested that the cytotoxicity of ionic liquids is dependent on the cationic surfactant action, long alkyl side chains, cationic lipophilicity as well as aromaticity, the presence of a dialkylamino substituent at the 4-position of the pyridinium nucleus and a bulky anionic moiety. The models have been transparently presented in the form of equations, thus allowing their easy transferability in accordance with the OECD guidelines. The models have also been subjected to rigorous validation tests proving their predictive potential and can hence be used for designing novel and "greener" ionic liquids. The major strength of the present study lies in the use of a diverse and large dataset, use of simple reproducible descriptors and compliance with the OECD norms.

Pharmacist work stress and learning from quality related events

BACKGROUND

Among the many stresses faced by pharmacy staff, quality related event (QRE) learning can be among the most significant. In the absence of a supportive organizational culture, the potential for blaming individuals, versus identifying key process flaws, is significant and can be very intimidating to those involved in such discussions and may increase an already stressful work environment.

OBJECTIVE

This research develops and tests a model of the relationship between the work stress faced by pharmacists and the extent of QRE learning in community pharmacies. Building upon recent research models that explore job characteristics and safety climate, the model proposes that work stress captured by the effort that the pharmacist invests into job performance, the extent to which the pharmacist is rewarded for such efforts, and the extent of pharmacist work-related commitment to their job, influence pharmacist assessment of the working conditions within their community pharmacy. It is further proposed that working conditions influence the extent of a blame culture and safety focus in the pharmacy, which, in turn, influences organizational learning from QREs.

METHODS

This research formed part of a larger study focused on QRE reporting in community pharmacies. As part of the larger study, a total of 1035 questionnaires were mailed to community pharmacists, pharmacy managers, and pharmacy owners in the Canadian province of Saskatchewan during the fall of 2013 and winter and spring of 2014. Partial least squares (PLS) using SmartPLS was selected to test and further develop the proposed model. An examination of the statistical significance of latent variable paths, convergent validity, construct reliability, discriminant validity, and variance explained was used to assess the overall quality of the model.

RESULTS

Of the 1035 questionnaire sent, a total of 432 questionnaires were returned for an initial response rate of approximately 42%. However, for this research, only questionnaires from staff pharmacists were used thereby reducing the number of usable questionnaires to 265. The final model highlights that pharmacist work stress greatly influences perceptions of the working conditions in the pharmacy (R(2) = 0.52), which, in turn, influence assessments of the safety focus (R(2) = 0.27) and blame culture (R(2) = 0.14) in the pharmacy. The model also found that the extent of a safety focus and blame culture within the pharmacy both influence the extent of organizational learning from QREs (R(2) = 0.44) within the pharmacy.

CONCLUSIONS

In an environment where financial rewards are not always possible, ensuring that pharmacy staff feel respected and encouraged in providing safe care may help enhance QRE learning. Given the importance placed on organizational reporting of, and learning from, QREs in many jurisdictions in North America, the findings from this study suggest that a number of working conditions and perceptions of blame culture and organizational safety need to be explored before such processes can become entrenched in work flow.

Optimization of wavelength range and data interval in chemometric analysis of complex pharmaceutical mixtures

The performance of different chemometric approaches was evaluated in the spectrophotometric determination of pharmaceutical mixtures characterized by having the amount of components with a very high ratio. Principal component regression (PCR), partial least squares with one dependent variable (PLS1) or multi-dependent variables (PLS2), and multivariate curve resolution (MCR) were applied to the spectral data of a ternary mixture containing paracetamol, sodium ascorbate and chlorpheniramine (150:140:1, m/m/m), and a quaternary mixture containing paracetamol, caffeine, phenylephrine and chlorpheniramine (125:6. 25:1.25:1, m/m/m/m). The UV spectra of the calibration samples in the range of 200–320 nm were pre-treated by removing noise and useless data, and the wavelength regions having the most useful analytical information were selected using the regression coefficients calculated in the multivariate modeling. All the defined chemometric models were validated on external sample sets and then applied to commercial pharmaceutical formulations. Different data intervals, fixed at 0.5, 1.0, and 2.0 point/nm, were tested to optimize the prediction ability of the models. The best results were obtained using the PLS1calibration models and the quantification of the species of a lower amount was significantly improved by adopting 0.5 data interval, which showed accuracy between 94.24% and 107.76%.

Metabolomics of meat exudate: Its potential to evaluate beef meat conservation and aging

In this study we analyzed the exudate of beef to evaluate its potential as non invasive sampling for nuclear magnetic resonance (NMR) based metabolomic analysis of meat samples. Exudate, as the natural juice from raw meat, is an easy to obtain matrix that it is usually collected in small amounts in commercial meat packages. Although meat exudate could provide complete and homogeneous metabolic information about the whole meat piece, this sample has been poorly studied. Exudates from 48 beef samples of different breeds, cattle and storage times have been studied by (1)H NMR spectroscopy. The liquid exudate spectra were compared with those obtained by High Resolution Magic Angle Spinning (HRMAS) of the original meat pieces. The close correlation found between both spectra (>95% of coincident peaks in both registers; Spearman correlation coefficient = 0.945) lead us to propose the exudate as an excellent alternative analytical matrix with a view to apply meat metabolomics. 60 metabolites could be identified through the analysis of mono and bidimensional exudate spectra, 23 of them for the first time in NMR meat studies. The application of chemometric tools to analyze exudate dataset has revealed significant metabolite variations associated with meat aging. Hence, NMR based metabolomics have made it possible both to classify meat samples according to their storage time through Principal Component Analysis (PCA), and to predict that storage time through Partial Least Squares (PLS) regression.

Study of the rapid detection of γ-aminobutyric acid in rice wine based on chemometrics using near infrared spectroscopy

Rice wine, in which γ-aminobutyric acid is present, is beneficial to human health and is one of the three most well-known fermented wines in the world, and is very popular in China. The rapid detection of γ-aminobutyric acid was studied in rice wine using near infrared spectroscopy with an optical fibre probe. Through the selection of detection conditions, including a waveband range of 12500-4000 cm(-1), a scanning duration of 16 scans and a resolution of 8 cm(-1), the near infrared spectrum of rice wine was acquired three times, for every wine sample, with an optical fibre probe. The resulting average value of the spectrum was obtained and the corresponding data were analysed via normalization. By adopting a multivariate calibration partial least squares method (PLS) and establishing a calibration model, the highest precision for γ-aminobutyric acid in rice wine was predicted when the factor coefficient was 17. The overall results demonstrating the content of γ-aminobutyric acid in rice wine was predicted to be between 157.6696-317.5813 mg/L, with a relative standard deviation of prediction between 0.01-5 %, as well as the fact that the single sample measuring time was less than 20 s, prove that near infrared spectroscopy is a rapid, accurate and effective method to adopt for detecting the content of γ-aminobutyric acid in rice wine.

A new strategy to prevent over-fitting in partial least squares models based on model population analysis

Partial least squares (PLS) is one of the most widely used methods for chemical modeling. However, like many other parameter tunable methods, it has strong tendency of over-fitting. Thus, a crucial step in PLS model building is to select the optimal number of latent variables (nLVs). Cross-validation (CV) is the most popular method for PLS model selection because it selects a model from the perspective of prediction ability. However, a clear minimum of prediction errors may not be obtained in CV which makes the model selection difficult. To solve the problem, we proposed a new strategy for PLS model selection which combines the cross-validated coefficient of determination (Qcv(2)) and model stability (S). S is defined as the stability of PLS regression vectors which is obtained using model population analysis (MPA). The results show that, when a clear maximum of Qcv(2) is not obtained, S can provide additional information of over-fitting and it helps in finding the optimal nLVs. Compared with other regression vector based indictors such as the Euclidean 2-norm (B2), the Durbin Watson statistic (DW) and the jaggedness (J), S is more sensitive to over-fitting. The model selected by our method has both good prediction ability and stability.

Evaluation of the value of near infrared (NIR) spectromicroscopy for the analysis of glycyrrizhic acid in licorice

It has been reported that hyperspectral data could be employed to qualitatively elucidate the spatial composition of tablets of Chinese medicinal plants. To gain more insights into this technology, a quantitative profile provided by near infrared (NIR) spectromicroscopy was further studied by determining the glycyrrhizic acid content in licorice, Glycyrrhiza uralensis. Thirty-nine samples from twenty-four different origins were analyzed using NIR spectromicroscopy. Partial least squares, interval partial least square (iPLS), and least squares support vector regression (LS-SVR) methods were used to develop linear and non-linear calibration models, with optimal calibration parameters (number of interval numbers, kernel parameter, etc.) being explored. The root mean square error of prediction (RMSEP) and the coefficient of determination (R(2)) of the iPLS model were 0.717 7% and 0.936 1 in the prediction set, respectively. The RMSEP and R(2) of LS-SVR model were 0.515 5% and 0.951 4 in the prediction set, respectively. These results demonstrated that the glycyrrhizic acid content in licorice could barely be analyzed by NIR spectromicroscopy, suggesting that good quality quantitative data are difficult to obtain from microscopic NIR spectra for complicated Chinese medicinal plant materials.

Improving PLS-RFE based gene selection for microarray data classification

Gene selection plays a crucial role in constructing efficient classifiers for microarray data classification, since microarray data is characterized by high dimensionality and small sample sizes and contains irrelevant and redundant genes. In practical use, partial least squares-based gene selection approaches can obtain gene subsets of good qualities, but are considerably time-consuming. In this paper, we propose to integrate partial least squares based recursive feature elimination (PLS-RFE) with two feature elimination schemes: simulated annealing and square root, respectively, to speed up the feature selection process. Inspired from the strategy of annealing schedule, the two proposed approaches eliminate a number of features rather than one least informative feature during each iteration and the number of removed features decreases as the iteration proceeds. To verify the effectiveness and efficiency of the proposed approaches, we perform extensive experiments on six publicly available microarray data with three typical classifiers, including Naïve Bayes, K-Nearest-Neighbor and Support Vector Machine, and compare our approaches with ReliefF, PLS and PLS-RFE feature selectors in terms of classification accuracy and running time. Experimental results demonstrate that the two proposed approaches accelerate the feature selection process impressively without degrading the classification accuracy and obtain more compact feature subsets for both two-category and multi-category problems. Further experimental comparisons in feature subset consistency show that the proposed approach with simulated annealing scheme not only has better time performance, but also obtains slightly better feature subset consistency than the one with square root scheme.

Neural activity associated with self, other, and object-based counterfactual thinking

Previous research has shown that autobiographical episodic counterfactual thinking-i.e., mental simulations about alternative ways in which one's life experiences could have occurred-engages the brain's default network (DN). However, it remains unknown whether or not the DN is also engaged during impersonal counterfactual thoughts, specifically those involving other people or objects. The current study compares brain activity during counterfactual simulations involving the self, others and objects. In addition, counterfactual thoughts involving others were manipulated in terms of similarity and familiarity with the simulated characters. The results indicate greater involvement of DN during person-based (i.e., self and other) as opposed to object-based counterfactual simulations. However, the involvement of different regions of the DN during other-based counterfactual simulations was modulated by how close and/or similar the simulated character was perceived to be by the participant. Simulations involving unfamiliar characters preferentially recruited dorsomedial prefrontal cortex. Simulations involving unfamiliar similar characters, characters with whom participants identified personality traits, recruited lateral temporal gyrus. Finally, our results also revealed differential coupling of right hippocampus with lateral prefrontal and temporal cortex during counterfactual simulations involving familiar similar others, but with left transverse temporal gyrus and medial frontal and inferior temporal gyri during counterfactual simulations involving either oneself or unfamiliar dissimilar others. These results suggest that different brain mechanisms are involved in the simulation of personal and impersonal counterfactual thoughts, and that the extent to which regions associated with autobiographical memory are recruited during the simulation of counterfactuals involving others depends on the perceived similarity and familiarity with the simulated individuals.

Using FTIR spectra and pattern recognition for discrimination of tea varieties

In order to classify typical Chinese tea varieties, Fourier transform infrared spectroscopy (FTIR) of tea polysaccharides (TPS) was used as an accurate and economical method. Partial least squares (PLS) modeling method along with a self-organizing map (SOM) neural network method was utilized due to the diversity and heterozygosis between teas. FTIR spectra results of tea extracts after spectra preprocessing were used as input data for PLS and SOM multivariate statistical analyses respectively. The predicted correlation coefficient of optimization PLS model was 0.9994, and root mean square error of calibration and cross-validation (RMSECV) was 0.03285. The features of PLS can be visualized in principal component (PC) space, contributing to discover correlation between different classes of spectra samples. After that, a data matrix consisted of the scores on the selected 3PCs computed by principle component analysis (PCA) and the characteristic spectrum data was used as inputs for training of SOM neural network. Compared with the PLS linear technique's recognition rate of 67% only, the correct recognition rate of the PLS-SOM as a non-linear classification algorithm to differentiate types of tea reaches up to 100%. And the models become reliable and provide a reasonable clustering of tea varieties.

Predicting cytotoxicity of complex mixtures in high cancer incidence regions of the Huai River Basin based on GC-MS spectrum with partial least squares regression

Complex mixture exposures, such as those associated with water sources, are an important issue in health risk assessment. This study assessed the cytotoxicity of chemical mixtures extracted from water sources in regions of the Huai River Basin with high cancer incidences and built statistical models of cytotoxicity based on pollution profiles that were measured with gas chromatography-mass spectrometry (GC-MS). Both surface and ground waters were collected from rural water sources of Shenqiu County, Henan Province of China from 2008 to 2011 and extracted with XAD-2 resigns. Cytotoxicity was evaluated with Chinese hamster ovary K1 (CHO-K1) cells and compared against the pollution profiles of the extracts. IC50 of water samples ranged from 0.023 to 0.338L-eq/mL. The pollutants in waters determined by GC-MS are complex and some of the compounds that contributed to cytotoxicity lack toxicity data. A partial least squares (PLS) regression model of cytotoxicity was built based on linear aggregation of predictor variables (i.e., peaks for single compounds in the gas chromatograms). The PLS model contains 2 PLS factors extracted from 141 variables. The model was validated internally with training data permutation and externally with a test sample. The model explained 92% of the cytotoxicity in the training samples and 40% in the test sample. This approach provides a general, rapid method for relating water toxicity to GC-MS chromatograms and for predicting the compounds that contribute most to toxicity.

Old wine in new bottles: validating the clinical utility of SPECT in predicting cognitive performance in mild traumatic brain injury

The neural underpinnings of cognitive dysfunction in mild traumatic brain injury (TBI) are not fully understood. Consequently, patient prognosis using existing clinical imaging is somewhat imprecise. Single photon emission computed tomography (SPECT) is a frequently employed investigation in this population, notwithstanding uncertainty over the clinical utility of the data obtained. In this study, subjects with mild TBI underwent (99m)Tc-ECD SPECT scanning, and were administered a brief battery of cognitive tests and self-report symptom scales of concussion and emotional distress. Testing took place 2 weeks (n=84) and 1 year (n=49) post-injury. Multivariate analysis (i.e., partial least squares analysis) revealed that frontal perfusion in right superior frontal and middle frontal gyri predicted poorer performance on the Stroop test, an index of executive function, both at initial and follow-up testing. Conversely, SPECT scans categorized as normal or abnormal by radiologists did not differentiate cognitively impaired from intact subjects. These results demonstrate the clinical utility of SPECT in mild TBI, but only when data are subjected to blood flow quantification analysis.

Simultaneous quantification of Aroclor mixtures in soil samples by gas chromatography/mass spectrometry with solid phase microextraction using partial least-squares regression

Multivariate partial least-squares (PLS) method was applied to the quantification of two complex polychlorinated biphenyls (PCBs) commercial mixtures, Aroclor 1254 and 1260, in a soil matrix. PCBs in soil samples were extracted by headspace solid phase microextraction (SPME) and determined by gas chromatography/mass spectrometry (GC/MS). Decachlorinated biphenyl (deca-CB) was used as internal standard. After the baseline correction was applied, four data representations including extracted ion chromatograms (EIC) for Aroclor 1254, EIC for Aroclor 1260, EIC for both Aroclors and two-way data sets were constructed for PLS-1 and PLS-2 calibrations and evaluated with respect to quantitative prediction accuracy. The PLS model was optimized with respect to the number of latent variables using cross validation of the calibration data set. The validation of the method was performed with certified soil samples and real field soil samples and the predicted concentrations for both Aroclors using EIC data sets agreed with the certified values. The linear range of the method was from 10μgkg(-1) to 1000μgkg(-1) for both Aroclor 1254 and 1260 in soil matrices and the detection limit was 4μgkg(-1) for Aroclor 1254 and 6μgkg(-1) for Aroclor 1260. This holistic approach for the determination of mixtures of complex samples has broad application to environmental forensics and modeling.

Using variable combination population analysis for variable selection in multivariate calibration

Variable (wavelength or feature) selection techniques have become a critical step for the analysis of datasets with high number of variables and relatively few samples. In this study, a novel variable selection strategy, variable combination population analysis (VCPA), was proposed. This strategy consists of two crucial procedures. First, the exponentially decreasing function (EDF), which is the simple and effective principle of 'survival of the fittest' from Darwin's natural evolution theory, is employed to determine the number of variables to keep and continuously shrink the variable space. Second, in each EDF run, binary matrix sampling (BMS) strategy that gives each variable the same chance to be selected and generates different variable combinations, is used to produce a population of subsets to construct a population of sub-models. Then, model population analysis (MPA) is employed to find the variable subsets with the lower root mean squares error of cross validation (RMSECV). The frequency of each variable appearing in the best 10% sub-models is computed. The higher the frequency is, the more important the variable is. The performance of the proposed procedure was investigated using three real NIR datasets. The results indicate that VCPA is a good variable selection strategy when compared with four high performing variable selection methods: genetic algorithm-partial least squares (GA-PLS), Monte Carlo uninformative variable elimination by PLS (MC-UVE-PLS), competitive adaptive reweighted sampling (CARS) and iteratively retains informative variables (IRIV). The MATLAB source code of VCPA is available for academic research on the website: http://www.mathworks.com/matlabcentral/fileexchange/authors/498750.

ATR-FTIR spectroscopy and quantitative multivariate analysis of paints and coating materials

The applicability of ATR-FTIR spectroscopy with partial least squares (PLS) data analysis was evaluated for quantifying the components of mixtures of paint binding media and pigments, and alkyd resins. PLS methods were created using a number of standard mixtures. Validation and measurement uncertainty estimation was carried out. Binary, ternary and quaternary mixtures of several common binding media and pigments were quantified, with standard measurement uncertainties in most cases below 3g/100g. Classes of components - aromatic anhydrides and alcohols - used in alkyd resin synthesis were also successfully quantified, with standard uncertainties in the range of 2-3g/100g. This is a more demanding application because in alkyd resins aromatic anhydrides and alcohols have reacted to form a polyester, and are not present in their original forms. Once a PLS method has been calibrated, analysis time and cost are significantly reduced from typical quantitative methods such as GC/MS. This is beneficial in the case of routine analysis where the components are known.