Class-modelling in food analytical chemistry: Development, sampling, optimisation and validation issues - A tutorial

Qualitative data modelling is a fundamental branch of pattern recognition, with many applications in analytical chemistry, and embraces two main families: discriminant and class-modelling methods. The first strategy is appropriate when at least two classes are meaningfully defined in the problem under study, while the second strategy is the right choice when the focus is on a single class. For this reason, class-modelling methods are also referred to as one-class classifiers. Although, in the food analytical field, most of the issues would be properly addressed by class-modelling strategies, the use of such techniques is rather limited and, in many cases, discriminant methods are forcedly used for one-class problems, introducing a bias in the outcomes. Key aspects related to the development, optimisation and validation of suitable class models for the characterisation of food products are critically analysed and discussed.

Effects of pacing, status and unbalance in time motion variables, heart rate and tactical behaviour when playing 5-a-side football small-sided games

OBJECTIVES

To compare time-motion variables, heart rate and players' tactical behaviour according to game pace (slow, normal or fast), status (winning and losing) and team unbalance (superiority and inferiority) in football 5-a-side small-sided games. To identify the most discriminating variables in classifying performances according to these constraints.

DESIGN

Cross-sectional field study.

METHODS

The data were gathered using global positioning systems (5 Hz) in 5-a-side small-sided games (7 × 5 min) played by twenty-four footballers. The tactical performance was measured using dynamical positioning variables, processed by non-linear signal processing techniques (approximate entropy). ANOVA models were used to compare between constraints and discriminant analyses to identify the variables that best discriminate between pacing and status × unbalance constraints.

RESULTS

The fast paced games had the highest mean speed value, followed by normal and slow paced games (8.2 ± 0.6 km h(-1), 7.8 ± 0.5 km h(-1) and 6.2 ± 0.4 km h(-1), respectively). The stronger predictor variables of pacing were the randomness in distance to team centroid and the distances covered above 13 km h(-1). The results also changed according to game status and team unbalance. The strongest predictor variables were the distance covered below 6.9 km h(-1), distance and randomness to team centroid, with higher values when winning in superiority conditions.

CONCLUSIONS

Practice task design manipulating game pace, status and team unbalance significantly influenced the emergent behavioural dynamics. Collective positioning variables were more accurate in discriminating these constraints and, therefore, need to be considered when planning and monitoring performance.

Volume of interest-based [18F]fluorodeoxyglucose PET discriminates MCI converting to Alzheimer's disease from healthy controls. A European Alzheimer's Disease Consortium (EADC) study

An emerging issue in neuroimaging is to assess the diagnostic reliability of PET and its application in clinical practice. We aimed at assessing the accuracy of brain FDG-PET in discriminating patients with MCI due to Alzheimer's disease and healthy controls. Sixty-two patients with amnestic MCI and 109 healthy subjects recruited in five centers of the European AD Consortium were enrolled. Group analysis was performed by SPM8 to confirm metabolic differences. Discriminant analyses were then carried out using the mean FDG uptake values normalized to the cerebellum computed in 45 anatomical volumes of interest (VOIs) in each hemisphere (90 VOIs) as defined in the Automated Anatomical Labeling (AAL) Atlas and on 12 meta-VOIs, bilaterally, obtained merging VOIs with similar anatomo-functional characteristics. Further, asymmetry indexes were calculated for both datasets. Accuracy of discrimination by a Support Vector Machine (SVM) and the AAL VOIs was tested against a validated method (PALZ). At the voxel level SMP8 showed a relative hypometabolism in the bilateral precuneus, and posterior cingulate, temporo-parietal and frontal cortices. Discriminant analysis classified subjects with an accuracy ranging between .91 and .83 as a function of data organization. The best values were obtained from a subset of 6 meta-VOIs plus 6 asymmetry values reaching an area under the ROC curve of .947, significantly larger than the one obtained by the PALZ score. High accuracy in discriminating MCI converters from healthy controls was reached by a non-linear classifier based on SVM applied on predefined anatomo-functional regions and inter-hemispheric asymmetries. Data pre-processing was automated and simplified by an in-house created Matlab-based script encouraging its routine clinical use. Further validation toward nonconverter MCI patients with adequately long follow-up is needed.

•

¹⁸F-FDG-PET/CT analysis of metabolic differences between MCI converting to AD and HC

•

Large and very well controlled cohorts from EADC-Consortium were investigated.

•

Data were analyzed by a friendly-to-use Matlab-based script and Support Vector Machine.

•

Excellent discrimination between MCI and HC (sensitivity 92%; specificity 91%)

•

Highest accuracy reported so far in MCI and promising implementation in clinical routine

Using global maps to predict the risk of dengue in Europe

This article attempts to quantify the risk to Europe of dengue, following the arrival and spread there of one of dengue's vector species Aedes (Stegomyia) albopictus. A global risk map for dengue is presented, based on a global database of the occurrence of this disease, derived from electronic literature searches. Remotely sensed satellite data (from NASA's MODIS series), interpolated meteorological data, predicted distribution maps of dengue's two main vector species, Aedes aegypti and Aedes albopictus, a digital elevation surface and human population density data were all used as potential predictor variables in a non-linear discriminant analysis modelling framework. One hundred bootstrap models were produced by randomly sub-sampling three different training sets for dengue fever, severe dengue (i.e. dengue haemorrhagic fever, DHF) and all-dengue, and output predictions were averaged to produce a single global risk map for each type of dengue. This paper concentrates on the all-dengue models. Key predictor variables were various thermal data layers, including both day- and night-time Land Surface Temperature, human population density, and a variety of rainfall variables. The relative importance of each may be shown visually using rainbow files and quantitatively using a ranking system. Vegetation Index variables (a common proxy for humidity or saturation deficit) were rarely chosen in the models. The kappa index of agreement indicated an excellent (dengue haemorrhagic fever, Cohen's kappa=0.79 ± 0.028, AUC=0.96 ± 0.007) or good fit of the top ten models in each series to the data (Cohen's kappa=0.73 ± 0.018, AUC=0.94 ± 0.007 for dengue fever and 0.74 ± 0.017, AUC=0.95 ± 0.005 for all dengue). The global risk map predicts widespread dengue risk in SE Asia and India, in Central America and parts of coastal South America, but in relatively few regions of Africa. In many cases these are less extensive predictions than those of other published dengue risk maps and arise because of the key importance of high human population density for the all-dengue risk maps produced here. Three published dengue risk maps are compared using the Fleiss kappa index, and are shown to have only fair agreement globally (Fleiss kappa=0.377). Regionally the maps show greater (but still only moderate) agreement in SE Asia (Fleiss kappa=0.566), fair agreement in the Americas (Fleiss kappa=0.325) and only slight agreement in Africa (Fleiss kappa=0.095). The global dengue risk maps show that very few areas of rural Europe are presently suitable for dengue, but several major cities appear to be at some degree of risk, probably due to a combination of thermal conditions and high human population density, the top two variables in many models. Mahalanobis distance images were produced of Europe and the southern United States showing the distance in environmental rather than geographical space of each site from any site where dengue currently occurs. Parts of Europe are quite similar in Mahalanobis distance terms to parts of the southern United States, where dengue occurred in the recent past and which remain environmentally suitable for it. High standards of living rather than a changed environmental suitability keep dengue out of the USA. The threat of dengue to Europe at present is considered to be low but sufficiently uncertain to warrant monitoring in those areas of greatest predicted environmental suitability, especially in northern Italy and parts of Austria, Slovenia and Croatia, Bosnia and Herzegovina, Serbia and Montenegro, Albania, Greece, south-eastern France, Germany and Switzerland, and in smaller regions elsewhere.

Comparing physicochemical properties of banana pulp and peel flours prepared from green and ripe fruits

Banana pulp and peel flour prepared from green and ripe Cavendish banana were assessed for physicochemical properties such as pH, total soluble solids (TSS), water holding capacity (WHC) and oil holding capacity (OHC) at 40, 60 and 80°C, colour values L^∗, a^∗ and b^∗, back extrusion force (BEF) and viscosity. Data obtained were analysed by MANOVA, discriminant analysis and cluster analysis. All statistical analyses showed that physicochemical properties of flour prepared from pulp and peel, and green and ripe banana were different from each other. TSS, WHC40, WHC60 and BEF can be used to discriminate between peel and pulp flour, whilst TSS and viscosity can be used to discriminate between flour prepared from green and ripe banana.

Major depressive disorder discrimination using vocal acoustic features

BACKGROUND

The voice carries various information produced by vibrations of the vocal cords and the vocal tract. Though many studies have reported a relationship between vocal acoustic features and depression, including mel-frequency cepstrum coefficients (MFCCs) which applied to speech recognition, there have been few studies in which acoustic features allowed discrimination of patients with depressive disorder. Vocal acoustic features as biomarker of depression could make differential diagnosis of patients with depressive state. In order to achieve differential diagnosis of depression, in this preliminary study, we examined whether vocal acoustic features could allow discrimination between depressive patients and healthy controls.

METHODS

Subjects were 36 patients who met the criteria for major depressive disorder and 36 healthy controls with no current or past psychiatric disorders. Voices of reading out digits before and after verbal fluency task were recorded. Voices were analyzed using OpenSMILE. The extracted acoustic features, including MFCCs, were used for group comparison and discriminant analysis between patients and controls.

RESULTS

The second dimension of MFCC (MFCC 2) was significantly different between groups and allowed the discrimination between patients and controls with a sensitivity of 77.8% and a specificity of 86.1%. The difference in MFCC 2 between the two groups reflected an energy difference of frequency around 2000-3000Hz.

CONCLUSIONS

The MFCC 2 was significantly different between depressive patients and controls. This feature could be a useful biomarker to detect major depressive disorder.

LIMITATIONS

Sample size was relatively small. Psychotropics could have a confounding effect on voice.

Gazefinder as a clinical supplementary tool for discriminating between autism spectrum disorder and typical development in male adolescents and adults

Background

Gaze abnormality is a diagnostic criterion for autism spectrum disorder (ASD). However, few easy-to-use clinical tools exist to evaluate the unique eye-gaze patterns of ASD. Recently, we developed Gazefinder, an all-in-one eye-tracking system for early detection of ASD in toddlers. Because abnormal gaze patterns have been documented in various ASD age groups, we predicted that Gazefinder might also detect gaze abnormality in adolescents and adults. In this study, we tested whether Gazefinder could identify unique gaze patterns in adolescents and adults with ASD.

Methods

We measured the percentage of eye fixation time allocated to particular objects depicted in movies (i.e., eyes and mouth in human face movies, upright and inverted biological motion in movies that presented these stimuli simultaneously, and people and geometry in movies that presented these stimuli simultaneously) by male adolescents and adults with ASD (N = 26) and age-matched males with typical development (TD; N = 35). We compared these percentages between the two groups (ASD and TD) and with scores on the social responsiveness scale (SRS). Further, we conducted discriminant analyses to determine if fixation times allocated to particular objects could be used to discriminate between individuals with and without ASD.

Results

Compared with the TD group, the ASD group showed significantly less fixation time at locations of salient social information (i.e., eyes in the movie of human faces without lip movement and people in the movie of people and geometry), while there were no significant groupwise differences in the responses to movies of human faces with lip movement or biological motion. In a within-group correlation analysis, a few of the fixation-time items correlated with SRS, although most of them did not. No items significantly correlated with SRS in both ASD and TD groups. The percentage fixation times to eyes and people, which exhibited large effect sizes for the group difference, could differentiate ASD and TD with a sensitivity of 81.0 % and a specificity of 80.0 %.

Conclusions

These findings suggest that Gazefinder is potentially a valuable and easy-to-use tool for objectively measuring unique gaze patterns and discriminating between ASD and TD in male adolescents and adults.

Electronic supplementary material

The online version of this article (doi:10.1186/s13229-016-0083-y) contains supplementary material, which is available to authorized users.

Surface-enhanced Raman scattering method for the identification of methicillin-resistant Staphylococcus aureus using positively charged silver nanoparticles

The article describes a SERS-based method for diagnosis of bacterial infections. Positively charged silver nanoparticles (AgNPs⁺) were employed for identification of methicillin-resistant Staphylococcus aureus (MRSA). It is found that AgNPs⁺ undergo self-assembly on the surface of bacteria via electrostatic aggregation. The assembled AgNPs⁺ are excellent SERS substrates. To prove the capability of SERS to differentiate between S. aureus and other microorganisms, six standard strains including S. aureus 29213, S. aureus 25923, C. albicans, B. cereus, E. coli, and P. aeruginosa were tested. To further demonstrate its applicability for the identification of MRSA in clinical samples, 52 methicillin-sensitive S. aureus (MSSA) isolates and 215 MRSA isolates were detected by SERS. The total measurement time (include incubation) is 45 min when using a 3 μL sample. The method gives a strongly enhanced Raman signal (at 730 cm^-1 and 1325 cm^-1) with good reproducibility and repeatability. It was successfully applied to the discrimination of the six strain microorganisms. The typical Raman peaks of S. aureus at 730, 1154, 1325, and 1457 cm^-1 were observed, which were assigned to the bacterial cell wall components (730 cm^-1- adenine, glycosidic ring mode, 1154 cm^-1- unsaturated fatty acid, 1325 cm^-1- adenine, polyadenine, and 1457 cm^-1 for -COO- stretching). S. aureus was completely separated from other species by partial least squares discriminant analysis (PLS-DA). Moreover, 52 MSSA isolates and 215 MRSA isolates from clinical samples were identified by PLS-DA. The accuracy was almost 100% when compared to the standard broth microdilution method. A classification based on latent structure discriminant analysis provided spectral variability directly. Conceivably, the method offers a potent tool for the identification of bacteria and antibiotics resistance, and for studies on antibiotic-resistance in general. Graphical abstract Schematic of the surface-enhanced Raman scattering (SERS) measurements on Staphylococcus aureus (S. aureus) using positively charged silver nanoparticles (AgNPs⁺). AgNPs⁺ are adsorbed on the bacterial cell wall by electrostatic attraction. SERS spectra were analyzed by PLS-DA for the identification of Staphylococcus aureus (MRSA) and methicillin-resistant Staphylococcus aureus (MSSA). MRSA isolates were divided into four groups, including R1, R2, R3, and R4. MSSA just includes group S.

Delirium diagnosis defined by cluster analysis of symptoms versus diagnosis by DSM and ICD criteria: diagnostic accuracy study

BACKGROUND

Information on validity and reliability of delirium criteria is necessary for clinicians, researchers, and further developments of DSM or ICD. We compare four DSM and ICD delirium diagnostic criteria versions, which were developed by consensus of experts, with a phenomenology-based natural diagnosis delineated using cluster analysis of delirium features in a sample with a high prevalence of dementia. We also measured inter-rater reliability of each system when applied by two evaluators from distinct disciplines.

METHODS

Cross-sectional analysis of 200 consecutive patients admitted to a skilled nursing facility, independently assessed within 24-48 h after admission with the Delirium Rating Scale-Revised-98 (DRS-R98) and for DSM-III-R, DSM-IV, DSM-5, and ICD-10 criteria for delirium. Cluster analysis (CA) delineated natural delirium and nondelirium reference groups using DRS-R98 items and then diagnostic systems' performance were evaluated against the CA-defined groups using logistic regression and crosstabs for discriminant analysis (sensitivity, specificity, percentage of subjects correctly classified by each diagnostic system and their individual criteria, and performance for each system when excluding each individual criterion are reported). Kappa Index (K) was used to report inter-rater reliability for delirium diagnostic systems and their individual criteria.

RESULTS

117 (58.5 %) patients had preexisting dementia according to the Informant Questionnaire on Cognitive Decline in the Elderly. CA delineated 49 delirium subjects and 151 nondelirium. Against these CA groups, delirium diagnosis accuracy was highest using DSM-III-R (87.5 %) followed closely by DSM-IV (86.0 %), ICD-10 (85.5 %) and DSM-5 (84.5 %). ICD-10 had the highest specificity (96.0 %) but lowest sensitivity (53.1 %). DSM-III-R had the best sensitivity (81.6 %) and the best sensitivity-specificity balance. DSM-5 had the highest inter-rater reliability (K =0.73) while DSM-III-R criteria were the least reliable.

CONCLUSIONS

Using our CA-defined, phenomenologically-based delirium designations as the reference standard, we found performance discordance among four diagnostic systems when tested in subjects where comorbid dementia was prevalent. The most complex diagnostic systems have higher accuracy and the newer DSM-5 have higher reliability. Our novel phenomenological approach to designing a delirium reference standard may be preferred to guide revisions of diagnostic systems in the future.

Identification and classification of textile fibres using ATR-FT-IR spectroscopy with chemometric methods

The possibility of classification of single- and two-component textile materials using ATR-FT-IR spectra and chemometric methods, principal component analysis (PCA) and discriminant analysis, was assessed. Altogether 89 textile samples belonging to 26 different types (11 one- and 15 two-component textiles) were investigated. It was found that PCA classification using only two or three principal components (PCs) enables identifying different one- and two-component textiles, although with two important limitations: it was not always possible to distinguish between the cellulose-based fibres (cotton, linen and in some cases viscose) and it was only partly possible to distinguish between silk and wool. The statistical discriminant analysis can use as many PCs as there are sample classes and due to that can discriminate between single-component fibres, including viscose from linen and cotton as well as silk from wool. Besides that, in both of these cases, involving optical microscopy as an additional technique enabled unequivocal identification of the fibres. The possibilities of semi-quantitative analysis of mixed fibres (cotton-polyester, wool-polyester and wool-polyamide) with PCA were investigated and it was found that approximate quantitative composition is obtainable if for the mixed fibre sample a number of spectra are averaged in order to minimize the effect of structural inhomogeneity. For approximate content determination 25 spectra of selected two-component samples were registered for calibration and the averaged spectrum for each sample was computed. Due to the structural inhomogeneity of mixed textiles, obtaining accurate quantitative composition from real samples is not possible with ATR-FT-IR. The main problems with ATR-FT-IR-PCA classification are (1) difficulties in getting high quality spectra from some textiles (e.g. polyacrylic), (2) inhomogeneity of the textile fibres in the case of two-component fibres and (3) intrinsic similarity between the spectra of some fibres (e.g. cotton and linen). In order to test the homogeneity of mixed fibres, microscopic and IR-microspectroscopic analysis was carried out.

A simple voltammetric electronic tongue for the analysis of coffee adulterations

This work presents a simple and low-cost analytical approach to detect adulterations in ground roasted coffee by using voltammetry and chemometrics. The voltammogram of a coffee extract (prepared as simulating a home-made coffee cup) obtained with a single working electrode is submitted to pattern recognition analysis preceded by variable selection to detect the addition of coffee husks and sticks (adulterated/unadulterated), or evaluate the shelf-life condition (expired/unexpired). Two pattern recognition methods were tested: linear discriminant analysis (LDA) with variable selection by successive projections algorithm (SPA), or genetic algorithm (GA); and partial least squares discriminant analysis (PLS-DA). Both LDA models presented satisfactory results. The voltammograms were also evaluated for the quantitative determination of the percentage of impurities in ground roasted coffees. PLS and multivariate linear regression (MLR) preceded by variable selection with SPA or GA were evaluated. An excellent predictive power (RMSEP = 0.05%) was obtained with MLR aided by GA.

Rapid detection of NBOME's and other NPS on blotter papers by direct ATR-FTIR spectrometry

Blotter paper is among the most common forms of consumption of new psychotropic substances (NPS), formerly referred as designer drugs. In many cases, users are misled to believe they are taking LSD when, in fact, they are taking newer and less known drugs like the NBOMEs or other substituted phenethylamines. We report our findings in quick testing of blotter papers for illicit substances like NBOMEs and other NPS by taking ATR-FTIR spectra directly from blotters seized on the streets, without any sample preparation. Both sides (front and back) of each blotter were tested. Collected data were analyzed by single- and multi-component spectral matching and submitted to chemometric discriminant analysis. Our results showed that, on 66.7% of the cases analyzed, seized blotters contained one or more types of NBOMEs, confirming the growing presence of this novel substances on the market. Matching IR signals were detected on both or just one side of the blotters and showed variable strength. Although no quantitative analysis was made, detection of these substances by the proposed approach serves as indication of variable and possibly higher dosages per blotter when compared to LSD, which showed to be below the detection limit of the applied method. Blotters containing a mescaline-like compound, later confirmed by GC-MS and LC-MS to be MAL (methallylescaline), a substance very similar to mescaline, were detected among the samples tested. Validity of direct ATR-FTIR testing was confirmed by checking the obtained results against independent GC-MS or LC-MS results for the same cases/samples.

Fast gradient HPLC/MS separation of phenolics in green tea to monitor their degradation

The degradation of catechins and other phenolics in green tea infusions were monitored using fast HPLC/MS separation. The final separation was performed within 2.5min using Ascentis Express C18 column (50mm×2.1mm i.d.) packed with 2μm porous shell particles. Degradation was studied in relation to the temperature of water (70, 80, 90°C) and the standing time of the infusion (up to 6h). Along with chromatographic separation, the antioxidant properties of the infusions were monitored using two spectrophotometric methods. During staying of green tea infusion, the degradation of some catechins probably to gallic acid was observed. Finally, the influence of tea bag storage on antioxidant properties of green tea was evaluated. Rapid degradation of antioxidants after 3weeks was observed. The principal component analysis, factor analysis and discriminant analysis were used for the statistical evaluation of obtained experimental data.

Surface-enhanced Raman spectroscopy for the identification of tigecycline-resistant E. coli strains

Tigecycline (TGC) is recognised as last resort of drugs against several antibiotic-resistant bacteria. Bacterial resistance to tigecycline due to presence of plasmid-mediated mobile TGC resistance genes (tet X3/X4) has broken another defense line. Therefore, rapid and reproducible detection of tigecycline-resistant E. coli (TREC) is required. The current study is designed for the identification and differentiation of TREC from tigecycline-sensitive E. coli (TSEC) by employing SERS by using Ag NPs as a SERS substrate. The SERS spectral fingerprints of E. coli strains associated directly or indirectly with the development of resistance against tigecycline have been distinguished by comparing SERS spectral data of TSEC strains with each TREC strain. Moreover, the statistical analysis including Principal Component Analysis (PCA), Hierarchical Cluster Analysis (HCA) and Partial Least Squares-Discriminant Analysis (PLS-DA) were employed to check the diagnostic potential of SERS for the differentiation among TREC and TSEC strains. The qualitative identification and differentiation between resistant and sensitive strains and among individual strains have been efficiently done by performing both PCA and HCA. The successful discrimination among TREC and TSEC at the strain level is performed by PLS-DA with 98% area under ROC curve, 100% sensitivity, 98.7% specificity and 100% accuracy.

Fingerprint and authenticity roasted coffees by 1H-NMR: the Brazilian coffee case

With globalization, it has become necessary to adopt policies to regulate the coffee market, addressing problems including the authenticity and traceability of products. It is therefore important to establish methodologies that can help to safeguard the interests of producer countries and add value to products. For this purpose, the use of NMR combined with multivariate statistical procedures can be an attractive option. The aim of this study was to develop a fast and effective technique, using 1H NMR coupled with multivariate statistics, to create a fingerprint of roasted coffees, distinguishing them according to the main Brazilian producer regions. Several compounds suitable for differentiating roasted coffees were identified in the fingerprint. Discriminant analysis revealed good distinction among the samples. The compounds catechol, trigonelline, caffeine, and n-methylpyridine were most important for the differentiation. The findings should assist coffee-producing countries in adopting measures to protect their markets and to add value to coffee products.

A new kernel discriminant analysis framework for electronic nose recognition

Electronic nose (e-Nose) technology based on metal oxide semiconductor gas sensor array is widely studied for detection of gas components. This paper proposes a new discriminant analysis framework (NDA) for dimension reduction and e-Nose recognition. In a NDA, the between-class and the within-class Laplacian scatter matrix are designed from sample to sample, respectively, to characterize the between-class separability and the within-class compactness by seeking for discriminant matrix to simultaneously maximize the between-class Laplacian scatter and minimize the within-class Laplacian scatter. In terms of the linear separability in high dimensional kernel mapping space and the dimension reduction of principal component analysis (PCA), an effective kernel PCA plus NDA method (KNDA) is proposed for rapid detection of gas mixture components by an e-Nose. The NDA framework is derived in this paper as well as the specific implementations of the proposed KNDA method in training and recognition process. The KNDA is examined on the e-Nose datasets of six kinds of gas components, and compared with state of the art e-Nose classification methods. Experimental results demonstrate that the proposed KNDA method shows the best performance with average recognition rate and total recognition rate as 94.14% and 95.06% which leads to a promising feature extraction and multi-class recognition in e-Nose.

Development of a FTIR-ATR based model for typing clinically relevant Acinetobacter baumannii clones belonging to ST98, ST103, ST208 and ST218

In this work we developed and validated a FTIR-ATR (Fourier transform infrared spectroscopy with attenuated total reflectance) based model for typing Acinetobacter baumannii clinical isolates belonging to ST98, ST208 and ST218 included into the worldwide spread clonal complex (CC) 92 and ST103. FTIR-ATR spectra of seventy-seven previously characterized isolates (Multi Locus Sequence Type-MLST, Pulsed-Field Gel Electrophoresis-PFGE and carbapenem-hydrolyzing class D β-lactamase-CHDL) were acquired and modeled by partial least squares discriminant analysis (PLSDA). The model was tested and successfully validated with a diverse collection of isolates (n=148) recovered from different countries and periods of time belonging to modeled and non-modeled STs.

Organochlorine concentrations (PCBs, DDTs, HCHs, HCB and MIREX) in delphinids stranded at the northeastern Brazil

Organochlorine compounds are highly persistent in the environment, causing bioaccumulation and biomagnification through the marine food chain. To verify the bioaccumulation pattern of DDT and its metabolites, as well as PCBs, Mirex, HCHs and HCB, samples of subcutaneous adipose tissue from 25 Guiana dolphins (Sotalia guianensis), two Fraser's dolphins (Lagenodelphis hosei), two spinner dolphins (Stenella longirostris), one spotted dolphin (Stenella frontalis), and one striped dolphin (Stenella coeruleoalba) were analyzed. The collected specimens were found stranded in northeastern Brazil between 2005 and 2011. The concentrations were reported in μg·g(-1) lip, varying from 0.02 to 17.3 for ΣPCB, from 0.003 to 5.19 for ΣDDT, from 0.005 to 0.16 for ΣHCH, from 0.002 to 0.16 for HCB, and from 0.02 to 2.38 for Mirex. The PCBs dominated the bioaccumulation pattern for both the Guiana and spinner dolphins; for the other species, DDT was the major compound. A discriminant function analysis revealed the differences in the organochlorine accumulation patterns between the Guiana dolphin (S. guianensis) and the continental shelf/oceanic dolphins (S. frontalis and S. longirostris). The results also show that the concentrations of organochlorine compounds found in the delphinids in northeastern Brazil were lower than the delphinids found in other regions of Brazil, as well as other locations worldwide with intense agroindustrial development and/or a larger population.

Seizure prediction in patients with mesial temporal lobe epilepsy using EEG measures of state similarity

OBJECTIVES

In patients with intractable epilepsy, predicting seizures above chance and with clinically acceptable performance has yet to be demonstrated. In this study, an intracranial EEG-based seizure prediction method using measures of similarity with a reference state is proposed.

METHODS

1565 h of continuous intracranial EEG data from 17 patients with mesial temporal lobe epilepsy were investigated. The recordings included 175 seizures. In each patient the data was split into a training set and a testing set. EEG segments were analyzed using continuous wavelet transform. During training, a reference state was defined in the immediate preictal data and used to derive three features quantifying the discrimination between preictal and interictal states. A classifier was then trained in the feature space. Its performance was assessed using testing set and compared with a random predictor for statistical validation.

RESULTS

Better than random prediction performance was achieved in 7 patients. The sensitivity was higher than 85%, the warning rate was less than 0.35/h and the proportion of time under warning was less than 30%.

CONCLUSION

Seizures are predicted above chance in 41% of patients using measures of state similarity.

SIGNIFICANCE

Sensitivity and specificity levels are potentially interesting for closed-loop seizure control applications.

Biochemical characterization of pathogenic bacterial species using Raman spectroscopy and discrimination model based on selected spectral features

This study aimed to evaluate the differences in the Raman spectra of nine clinical species of bacteria isolated from infections (three Gram-positive and six Gram-negative species), correlating the spectra with the chemical composition of each species and to develop a classification model through discriminant analysis to categorize each bacterial strain using the peaks with the most significant differences. Bacteria were cultured in Mueller Hinton agar and a sample of biomass was harvested and placed in an aluminum sample holder. A total of 475 spectra from 115 different strains were obtained through a dispersive Raman spectrometer (830 nm) with exposure time of 50 s. The intensities of the peaks were evaluated by one-way analysis of variance (ANOVA) and the peaks with significant differences were related to the differences in the biochemical composition of the strains. Discriminant analysis based on quadratic distance applied to the peaks with the most significant differences and partial least squares applied to the whole spectrum showed 89.5% and 90.1% of global accuracy, respectively, for classification of the spectra in all the groups. Raman spectroscopy could be a promising technique to identify spectral differences related to the biochemical content of pathogenic microorganisms and to provide a faster diagnosis of infectious diseases.

Evaluation of surface water quality by using Canadian Council of Ministers of the Environment Water Quality Index (CCME WQI) method and discriminant analysis method: a case study Coruh River Basin

In this study, the water quality of the Coruh River Basin, which is located in the Eastern Black Sea Region of Turkey, was evaluated. The water quality data measurement results obtained by the State Hydraulic Works 26th Regional Directorate from four different sites over a course of 4 years between the years 2011 and 2014 in the Coruh River Basin were used as the data. In this study, the water quality was evaluated by using the Canadian Council of Ministers of the Environmental Water Quality Index (CCME WQI) method and discriminant analysis (DA). The water quality of the Coruh River Basin was calculated as 30.4 and 71.35 by using the CCME WQI and classified as "poor," "marginal," and "fair". These values show that the water of the Coruh River Basin is degraded and under threat and its overall quality is not close to natural or desired levels. The monitoring sites were divided into two groups by the cluster analysis (CA). DA is a multivariate analysis technique used to divide individuals or objects into different groups and assign them into predetermined groups. As a result of DA, calcium (Ca) and sulfate (SO₄) were determined to be significant parameters in the determination of the water quality of the Coruh River Basin. The success of DA depends on the percentage of correct classification. As a result of the analysis, 23% of the parameters in the first measurement point, 69.2% of the parameters in the second and third measurement points, and 76.9% of the parameters in the fourth measurement point were classified correctly. Since the second measurement point is the discharge point of a copper mine, it can be said that the water quality parameters measured may provide accurate results in detecting pollution at this point.

A pilot study to determine whether combinations of objectively measured activity parameters can be used to differentiate between mixed states, mania, and bipolar depression

BACKGROUND

Until recently, actigraphy studies in bipolar disorders focused on sleep rather than daytime activity in mania or depression, and have failed to analyse mixed episodes separately. Furthermore, even those studies that assessed activity parameters reported only mean levels rather than complexity or predictability of activity. We identified cases presenting in one of three acute phases of bipolar disorder and examined whether the application of non-linear dynamic models to the description of objectively measured activity can be used to predict case classification.

METHODS

The sample comprised 34 adults who were hospitalized with an acute episode of mania (n = 16), bipolar depression (n = 12), or a mixed state (n = 6), who agreed to wear an actiwatch for a continuous period of 24 h. Mean level, variability, regularity, entropy, and predictability of activity were recorded for a defined 64-min active morning and active evening period. Discriminant function analysis was used to determine the combination of variables that best classified cases based on phase of illness.

RESULTS

The model identified two discriminant functions: the first was statistically significant and correlated with intra-individual fluctuation in activity and regularity of activity (sample entropy) in the active morning period; the second correlated with several measures of activity from the evening period (e.g. Fourier analysis, autocorrelation, sample entropy). A classification table generated from both functions correctly classified 79% of all cases based on phase of illness (χ ² = 36.21; df 4; p = 0.001). However, 42% of bipolar depression cases were misclassified as being in manic phase.

CONCLUSIONS

The findings should be treated with caution as this was a small-scale pilot study and we did not control for prescribed treatments, medication adherence, etc. However, the insights gained should encourage more widespread adoption of statistical approaches to the classification of cases alongside the application of more sophisticated modelling of activity patterns. The difficulty of accurately classifying cases of bipolar depression requires further research, as it is unclear whether the lower prediction rate reflects weaknesses in a model based only on actigraphy data, or if it reflects clinical reality i.e. the possibility that there may be more than one subtype of bipolar depression.

Semiparametric Bayesian classification with longitudinal markers

We analyse data from a study involving 173 pregnant women. The data are observed values of the β human chorionic gonadotropin hormone measured during the first 80 days of gestational age, including from one up to six longitudinal responses for each woman. The main objective in this study is to predict normal versus abnormal pregnancy outcomes from data that are available at the early stages of pregnancy. We achieve the desired classification with a semiparametric hierarchical model. Specifically, we consider a Dirichlet process mixture prior for the distribution of the random effects in each group. The unknown random-effects distributions are allowed to vary across groups but are made dependent by using a design vector to select different features of a single underlying random probability measure. The resulting model is an extension of the dependent Dirichlet process model, with an additional probability model for group classification. The model is shown to perform better than an alternative model which is based on independent Dirichlet processes for the groups. Relevant posterior distributions are summarized by using Markov chain Monte Carlo methods.

Discrimination between schizophrenia and major depressive disorder by magnetic resonance imaging of the female brain

BACKGROUND

Although schizophrenia and major depressive disorder (MDD) differ on a variety of neuroanatomical measures, a diagnostic tool to discriminate these disorders has not yet been established. We tried to identify structural changes of the brain that best discriminate between schizophrenia and MDD on the basis of gray matter volume, ventricle volume, and diffusion tensor imaging (DTI).

METHOD

The first exploration sample consisted of 25 female patients with schizophrenia and 25 females with MDD. Regional brain volumes and fractional anisotropy (FA) values were entered into a discriminant analysis. The second validation sample consisted of 18 female schizophrenia and 16 female MDD patients.

RESULTS

The stepwise discriminant analysis resulted in correct classification rates of 0.80 in the schizophrenic group and 0.76 in MDD. In the second validation sample, the obtained model yielded correct classification rates of 0.72 in the schizophrenia group and 0.88 in the MDD group.

CONCLUSION

Our results suggest that schizophrenia and MDD have differential structural changes in the examined brain regions and that the obtained discriminant score may be useful to discriminate the two disorders.

Effect of finishing diet and duration on the sensory quality and volatile profile of lamb meat

Animal production factors can affect the sensory quality of lamb meat. The study investigated the effect of diet composition and duration of consumption on the proximate analysis, volatile profile and sensory quality of lamb meat. Ninety-nine male Texel × Scottish Blackface lambs were raised at pasture for 10 months before being assigned in groups of 11 to one of the following treatments: 100% Silage (S) for 36 (S36), 54 (S54) or 72 (S72) days; 50% Silage - 50% Concentrate (SC) for 36 (SC36), 54 (SC54) or 72 (SC72) days; 100% Concentrate (C) for 36 (C36) or 54 (C54) or 72 (C72) days. A trained sensory panel found Intensity of Lamb Aroma, Dry Aftertaste and Astringent Aftertaste to be higher in meat from lambs on the concentrate diet. Discriminant analysis showed that the volatile profile enabled discrimination of lamb based on dietary treatment but the volatile differences were insufficient to impact highly on sensory quality. Muscle from animals in the S54 group had higher Manure/Faecal Aroma and Woolly Aroma than the SC54 and C54 groups, possibly related to higher levels of indole and skatole. Further research is required to establish if these small differences would influence consumer acceptability.