Stepwise Linear Discriminant Analysis Research Articles

Designing Of Efficient Model for Facial Feature Identification and Classification Using a Sparse Fingerprint Algorithm

By leveraging the geometrical and textural data of the retrieved local characteristics to align the probing partial face to gallery faces throughout this process, we present a novel partial face identification technique. Our fundamental assumption is that the cost function of our alignment approach ought to be reduced if the probing partial face patch and the gallery face picture are using the exact same individual. The similarity between the partial probing patch also the gallery photos across the identified facial feature points is also computed using a point-set distance metric that we provide. The usefulness of the suggested technique is demonstrated by experimental findings on four popular face datasets. This method focuses on choosing localised features from facial detection photos to distinguish between classes depending upon regression results, or partial F-test. The results demonstrate that traditional procedures are more resilient in terms of appropriate feature selection and categorization. The most noticeable characteristics were chosen by introducing a reliable method termed stepwise linear discriminant analysis, which concentrates on choosing the localised features from the face frames and classifying them relied upon regression values. A feature extraction procedure's goal is to extract the localised characteristics from faces that the earlier feature extraction methods were unable to analyse.

P300 Tabanlı Beyin Bilgisayar Arayüzü Sistemlerinde Uyaranlar Arası Sürenin ve Uyaran Yapısının Performansa Etkisi

Elektroensefalografi temelli beyin bilgisayar arayüzü sistemleri kas sistemini kullanamayan hastalar için dış dünya ile iletişimini mümkün kılmaktadır. EEG temelli beyin bilgisayar arayüzü sistemleri için çeşitli beyin sinyal aktiviteleri kullanılmaktadır. Olay odaklı potansiyellerden bir tanesi olan P300 sinyali beyin bilgisayar arayüzü sistemleri için elverişli bir beyin sinyalidir. P300 tabanlı bir beyin bilgsayar arayüzü için en önemli performans parametrelerinden birisi sınıflandırma doğruluk oranıdır. Bu çalışmada satır sütun temelli P300 heceletici yapısındaki değişiklikle daha yüksek doğruluk oranı ile elde edilmesi hedeflenmiştir. P300 heceleticisi matris yapısında ve uyaranların aralık süreleri üzerinde değişiklikler yapılmıştır. Bundan önceki bir çok çalışma P300 heceletici yapısındaki uyaran renk ve biçimlerindeki değişiklikler ile yapılmıştır. Uyaran yapısı ve uyaran aralık sürelerindeki değişikleri kıyaslayıcı P300 heceleticileri ile ilgili çalışmalar yeterli seviyede değildir. Bu çalışmada dört farklı yapıdaki satır sütun bazlı P300 heceletici kullanılarak deneyler yapılmıştır. Deneyler ile toplanan EEG kayıtları ön işlemden geçirildikten sonra adımsal doğrusal ayrışım analizi ile sınıflandırılmıştır. Sınıflandırma neticesinde bu çalışmada karşılaştırılan heceleticilerden; 4x4 satır sütın bazlı P300 heceleticinin 150 ms uyaran aralık süresine sahip olan yapıdaki formu, ortalama doğruluk oranı %84,76 ile en yüksek olarak tespit edilmiştir. En düşük performans ise; 6x6 satır sütın bazlı P300 heceleticinin 300 ms uyaranlar arası geçiş süresine sahip modunda %50,48 olarak gözlenmiştir. Bu çalışma, satır sütun bazlı P300 heceleticisinin uyaran matris yapısındaki değişikliği ve farklı uyaran aralık sürelerinde yapılan deneylerle yüksek doğruluk oranı ile elde edilebileceğini göstermiştir.

2 Stages-region-based P300 Speller in Brain–Computer Interface

ABSTRACTBrain–computer interface (BCI) applications present communication model without using peripheral nerves and neuromuscular systems. The P300 waves are used in BCI applications. Signal classification accuracy is a significant parameter for P300 BCI application. In this study, our goal is to investigate P300 speller structure for higher classification accuracy. There are a lot studies about P300 speller variations of stimulus model. This models mostly includes about row–column-based visual or audio stimuli model P300 spellers. But there is not enough study about region-based P300 spellers. This study contributes to region-based P300 spellers researches. Our new paradigm 2-stages region-based P300 speller has different amount of regions and stimulus position. During the experiment also, row–column-based P300 speller was used for comparing accuracy rates with our unique design 2-stages region-based P300 speller. The subject focused on the desired character stimulus. We used the stepwise linear discriminant analysis (SWLDA) method for classification that either included the desired P300 signal or not. According to SWLDA, the maximum mean classification accuracy value of the experiment was 83.33% with 2-stages region-based P300 speller. With this new paradigm, the classification accuracy was improved by 23.89% according to the most commonly used row–column-based P300 speller.

Urinary volatile metabolites of amygdala-kindled mice reveal novel biomarkers associated with temporal lobe epilepsy

Epilepsy is a chronic neurological disorder affecting mammals, including humans. Uncontrolled epilepsy is associated with poor quality of life, accidents, and sudden death. In particular, temporal lobe epilepsy (TLE) is the most common type of pharmacoresistant epilepsy, which easily gets out of control in human adults. The aim of this study was to profile urinary volatile organic compounds (VOCs) in a mouse model of TLE using solid-phase microextraction (SPME) gas chromatography mass spectrometry (GC-MS). Thirteen urinary VOCs exhibited differential abundance between epileptic and control mice, and the corresponding areas under the receiver operating characteristic (ROC) curve were greater than 0.8. Principal component analysis (PCA) based on these 13 VOCs separated epileptic from sham operated-mice, suggesting that all these 13 VOCs are epilepsy biomarkers. Promax rotation and dendrogram analysis concordantly separated the 13 VOCs into three groups. Stepwise linear discriminant analysis extracted methanethiol; disulfide, dimethyl; and 2-butanone as predictors. Based on known metabolic systems, the results suggest that TLE induced by amygdala stimulation could affect both endogenous metabolites and the gut flora. Future work will elucidate the physiological meaning of the VOCs as end-products of metabolic networks and assess the impact of the metabolic background involved in development of TLE.

Digital mapping of buried soil horizons using 2D and pseudo‐3D geoelectrical measurements in a ground moraine landscape

AbstractThe identification of buried soil horizons in agricultural landscapes helps to quantify sediment budgets and erosion‐related carbon dynamics. High‐resolution mapping of buried horizons using conventional soil surveys is destructive and time consuming. Geoelectrical sensors can offer a fast and non‐destructive alternative for determining horizon positions and properties. In this paper, we compare the suitability of several geoelectrical methods for measuring the depth to buried horizons (Apb, Ahb and Hab) in the hummocky ground moraine landscape of northeastern Germany. Soil profile descriptions were developed for 269 locations within a 6‐ha experimental field “CarboZALF‐D”. A stepwise linear discriminant analysis (LDA) estimated the lateral position of the buried horizons using electromagnetic induction data and terrain attributes. To predict the depth of a buried horizon, multiple linear regression (MLR) was used for both a 120‐m transect and a 0.2‐ha pseudo‐three‐dimensional (3D) area. At these scales, apparent electrical conductivity (ECa), electrical resistivity (ER) and terrain attributes were used as independent variables. The LDA accurately predicted Apb‐ and Ahb‐horizons (a correct classification of 93%). The LDA of the Hab‐horizon had a misclassification of 24%, which was probably related to the smaller test set and the higher depth of this horizon. The MLR predicted the depth of the Apb‐, Ahb‐ and Hab‐horizons with relative root mean square errors (RMSEs) of 7, 3 and 13%, respectively, in the pseudo‐3D area. MLR had a lower accuracy for the 2D transect compared to the pseudo‐3D area. Overall, the use of LDA and MLR has been an efficient methodological approach for predicting buried horizon positions.Highlights The suitability of geoelectrical measurements for digital modelling of diagnostic buried soil horizons was determined. LDA and MLR were used to detect multiple horizons with geoelectrical devices and terrain attributes. Geoelectrical variables were significant predictors of the position of the target soil horizons. The use of these tested digital technologies gives an opportunity to develop high‐resolution soil mapping procedures.

A two-step idle-state detection method for SSVEP BCI.

The role of detecting work/idle state in asynchronous Steady-state visual evoked potential (SSVEP) Brain-computer interface (BCI) or a self-paced SSVEP BCI has received increased attention in recent years. This study proposed a tree structure method which identifies the work/idle state based on the frequency recognition to detect work/idle state. Firstly, a frequency recognition estimated with task-related component analysis (TRCA). Then, the work/idle state is classified with step-wise linear discriminant analysis (SWLDA) using the data fusion of TRCA scores and power spectral density (PSD) as features. This method was evaluated by Electroencephalography (EEG) data from fourteen healthy participants with eight frequencies as work states and three idle state conditions. The averaged AUC of this method achieved 0.89 with data lengths of one second, which was significantly higher than that of the conventional power spectrum-based algorithm. The proposed method could identify the work/idle state fast and accurately, making the SSVEP BCI better suited for practical application.

Discrimination of Mineralized Rock Types in a Copper-Rich Volcanogenic Massive Sulfide Deposit Through Fast Independent Component and Factor Analysis

Multivariate methods are useful for simplifying the interpretation of variables in geochemical data and are widely used to uncover relationships between elements that are associated with geological and mineralization processes. Among these approaches, factor analysis (FA) is one of the most popular, whereas independent component analysis (ICA) has only been employed in a few cases. This study compared the effectiveness of these methods in distinguishing rock types and detecting mineralization signatures based on data on core samples obtained from the Nohkouhi copper deposit. The FA and ICA discriminated four known rock categories, namely barren rhyodacite, mineralized rhyodacite, barren black shale, and mineralized black shale. Stepwise linear discriminant analysis was used to compare the results of the FA and ICA and select components that effectively enable the discrimination of rock types and mineralization. First, rock types were distinguished with reference to the scores calculated via FA and ICA. The results showed that the FA and ICA achieved overall accuracy levels of 94% and 96% in rock-type discrimination, respectively. Second, each rock type was classified as either mineralized or barren. The FA exhibited classification accuracies of 76% for black shales and 70% for rhyodacites, whereas the ICA yielded classification accuracies of 79% and 88% for the two rock types, respectively.

Combination of an E-Nose and an E-Tongue for Adulteration Detection of Minced Mutton Mixed with Pork

An E-panel, comprising an electronic nose (E-nose) and an electronic tongue (E-tongue), was used to distinguish the organoleptic characteristics of minced mutton adulterated with different proportions of pork. Meanwhile, the normalization, stepwise linear discriminant analysis (step-LDA), and principle component analysis were employed to merge the data matrix of E-nose and E-tongue. The discrimination results were evaluated and compared by canonical discriminant analysis (CDA) and Bayesian discriminant analysis (BAD). It was shown that the capability of discrimination of the combined system (classification error 0%∼1.67%) was superior or equable to that obtained with the two instruments separately, and E-tongue system (classification error for E-tongue 0∼2.5%) obtained higher accuracy than E-nose (classification error 0.83%∼10.83% for E-nose). For the combined system, the combination of extracted data of 6 PCs of E-nose and 5 PCs of E-tongue was proved to be the most effective method. In order to predict the pork proportion in adulterated mutton, multiple linear regression (MLR), partial least square analysis (PLS), and backpropagation neural network (BPNN) regression models were used, and the results were compared, aiming at building effective predictive models. Good correlations were found between the signals obtained from E-tongue, E-nose, and fusion data of E-nose and E-tongue and proportions of pork in minced mutton with correlation coefficients higher than 0.90 in the calibration and validation data sets. And BPNN was proved to be the most effective method for the prediction of pork proportions with R2 higher than 0.97 both for the calibration and validation data set. These results indicated that integration of E-nose and E-tongue could be a useful tool for the detection of mutton adulteration.

Discrimination of the species and authenticity of Rhizoma Coptidis based on stable isotope and multielement fingerprinting and multivariate statistical analysis.

It is essential to be able to identify the source species and to determine the authenticity of traditional Chinese medicines (TCM) in order to prevent the use of false or inferior medicines. In this work, a stable and reliable method of discriminating among the three source species of Rhizoma Coptidis and checking the authenticity of Rhizoma Coptidis samples was established. The technique involved evaluating stable isotope ratios and the contents of multiple elements in samples along with the use of multivariate statistical techniques. The stable isotope ratios δ13C, δ15N, δ2H, and δ18O and the concentrations of various inorganic elements (Li, B, Na, Mg, Al, P, Si, K, Ca, Ti, Mn, Fe, Cu, Zn, Sr, and Ba) in authentic Rhizoma Coptidis samples from three source species (n = 56) and in counterfeit Rhizoma Coptidis samples (n = 39) were determined. The results showed that there were significant differences between the samples from different source species according to multivariate statistical analysis. The three species were clearly distinguished using hierarchical cluster analysis (HCA). Employing stepwise linear discriminant analysis (SLDA), a classification model for differentiating the three species was developed, and this model achieved 100% classification accuracy when applied to samples. In addition, authentic samples and counterfeit samples were successfully discriminated using stable isotope and multielement fingerprint analysis and orthogonal projections to latent structures discriminant analysis (OPLS-DA), and OPLS-DA models for checking the authenticity of Rhizoma Coptidis were established and verified. Therefore, stable isotope and multielement analysis combined with multivariate statistical analysis was shown to be a promising method of discriminating among the three source species of Rhizoma Coptidis and of establishing the authenticity of Rhizoma Coptidis samples. Graphical abstract.

Identification of geographical origin of Keemun black tea based on its volatile composition coupled with multivariate statistical analyses.

Keemun black tea (KBT) is one of the most popular tea beverages in China as a result of its unique flavor and potential health benefits. The geographical origin of KBT influences its quality and price. The present study aimed to apply a head-space solid phase microextraction approach and gas chromatography-mass spectrometry combined with chemometric analysis to profile the volatile compounds of KBT collected from five production areas. Thirty-one peaks were detected in 61 KBT samples. Hierarchical cluster analysis, principal component analysis (PCA), k-nearest neighbor (k-NN) and stepwise linear discriminant analysis (SLDA) were employed to visualize the volatile fractions. The results of unsupervised statistical tools were compared using a test for similarities and distinctions, which showed that different sources may be associated. A satisfying combination of average recognition (91.7%) and cross-validation prediction abilities (84.6%) was obtained for the PCA-k-NN. Among all of the statistical tools, SLDA provided promising results, with 100% recognition and 96.4% prediction ability. The results obtained in the present study indicate that the volatile compounds can be used as indicators to identify the geographical origin of KBT. © 2019 Society of Chemical Industry.

Mineralogical-chemical Alteration and Origin of Ignimbritic Stones Used in the Old Cathedral of Nostra Signora di Castro (Sardinia, Italy)

ABSTRACTThe pyroclastic rocks belonging to the Late Eocene-Miocene volcanic activity that occurred in Sardinia between 38 and 15 Ma ago were widely used as construction materials in several Romanesque churches of the easternmost Logudoro area, as well as in large parts of the Sardinia territory. In this work, the ancient Cathedral of Nostra Signora di Castro (twelfth century) was taken as a representative case study. There is no historical or archaeological evidence of ancient quarries. Based on the geochemical, petrographic, and volcanological data on several samples from an extensive field area (approximately 150 km2), a geographical zoning of the volcanics has been recognised. In the Oschiri sector, there are three different sub-zones, which can be identified with different volcanic rocks: less fractionated rocks (Differentation Index ∼70–78); intermediately fractionated rocks (D.I. ∼76–79); and more fractionated rocks (D.I. ∼77–82). To identify the origin of the ignimbrite rocks of the Church of Nostra Signora di Castro, two statistical methods were used: stepwise linear discriminant and canonical analysis. Moreover, to define the geochemical transformation processes induced by the alteration, a comparative study of concentrations of major and trace elements measured by XRF and SEM-EDX analyses on the surface portion and the innermost areas of the stone was made.

Contributions of olfactory and neuropsychological assessment to the diagnosis of first-episode schizophrenia.

First-episode schizophrenia and schizoaffective patients (SZ+) show olfactory impairments, but how these relate to cognitive dysfunction remains unclear. We examined the relationship between cognitive and olfactory dysfunction in SZ+ and the clinical utility of these measures in the assessment of SZ+ patients. First-episode SZ+ patients (n = 63) and controls (n = 63) were administered tests of odor identification and discrimination in addition to measures of manual dexterity, processing speed, attention and working memory, executive functioning, ideational fluency, and memory. We analyzed the relationships between olfactory and cognitive variables and conducted stepwise multiple regressions to identify which cognitive indices best predicted olfactory performance within the SZ+ group. Linear discriminant analysis was used to identify which measures best distinguished cases from controls. Among patients, odor discrimination correlated with perseverative errors and odor identification correlated with bilateral manual dexterity. Odor discrimination performance was best predicted by perseverative errors and letter fluency, whereas odor identification ability was best predicted by manual dexterity. Stepwise linear discriminant analysis revealed that manual dexterity, letter-guided word fluency, and odor discrimination best distinguished SZ+ from healthy adults. These findings indicate that manual dexterity, letter-guided word fluency, and odor discrimination may provide incremental information that strengthens a diagnosis of SZ+. Although odor discrimination tasks have received limited attention in schizophrenia studies, the extant data along with the present results indicate that odor discrimination tasks may have utility over odor identification measures as a neurodevelopmental risk marker. Additional studies examining odor discrimination as a predictor of SZ spectrum illness are warranted. (PsycINFO Database Record (c) 2019 APA, all rights reserved).

The use of cardiopulmonary exercise testing in identifying the presence of obstructive sleep apnea syndrome in patients with compatible symptomatology

The aim of this study was to investigate the relationship between cardiopulmonary exercise testing (CPET) and the presence of obstructive sleep apnea syndrome (OSAS) in order to provide an innovative tool to identify patients with OSAS. A prospective nested case control design was adopted. A consecutive population of male volunteers referred to a Sleep Unit was subjected to nocturnal polysomnography, full lung function testing and maximal CPET. A stepwise linear discriminant function analysis (DFA) was applied to construct a model which could identify individuals with moderate-to-severe OSAS from healthy controls. The total of 30 volunteers formed the OSAS and 24 the non-OSAS groups. Demographic and somatometric parameters were similar between groups. Patients presented with lower Expiratory Reserve Volume (ERV: 106.7 ± 28.3 vs. 123.9 ± 22.1, p < 0.001), Leg FatigueBorg scale (3.9 ± 1.1 vs. 6.1 ± 1.4, p < 0.001), VO2peak(25.0 ± 5.9 vs. 32.9 ± 7.2 ml/kg−1/min−1, p < 0.001), peak breathing frequency (31.0 ± 5.8 vs. 35.5 ± 7.3 1/min−1, p < 0.001) and peak heart rate (151.1 ± 17.7 vs. 171.2 ± 12.6 beats/min−1, p < 0.001) compared to controls, but higher peak end-tidal CO2 (PETCO2peak:38.6 ± 4.2 vs. 35.0 ± 4.9 mmHg, p = 0.043) and peak systolic (SBP:188.3 ± 21.9 vs. 173.1 ± 17.9 mmHg, p = 0.009) and diastolic (DBP: 91.3 ± 8.2 vs. 85.4 ± 8.2 mmHg, p = 0.011) blood pressure. Stepwise DFA indicated that ERV% of predicted (0.372), PETCO2peak (−0.376), SpO2resting (0.0667), Leg Fatigue Borg scale (0.564), HRpeak (0.530) and DBPpeak (−0.543) could separate the two groups, with an overall predictive accuracy of 96.3%. Selected CPET parameters (ERV% of predicted, PETCO2peak, SpO2resting, HRpeak, DBPpeak and Leg FatigueBorg Scale) are independently associated with OSAS presence and could discriminate patients with and without this disorder.

A New Paradigm for Region-Based P300 Speller in Brain Computer Interface

Electroencephalography-based brain computer interface systems could provide alternative communication methods for severely disabled people who cannot use their neuromuscular systems. The P300 signal is one of the event related potentials that are used for brain computer interface systems. The most important performance parameter of a P300 based brain computer interface system is information transfer rate that is calculated by using classification accuracy and P300 signal detection time. Moreover, P300 speller has a very critical role for classification accuracy and information transfer rate in a P300 based brain computer interface. Although most of studies are about row column based P300 speller in literature, region based P300 speller proved that has higher classification accuracy than row column based one. There are very few studies about region based P300 speller. This study aims to investigate methods for obtaining higher classification accuracy and information transfer rate with using region based P300 speller that constituted audio and visual stimulus. This is the first research that using audio and visual stimulus for a region based P300 speller in literature. Previous studies about region based P300 spellers focused on spellers with only visual stimulus types. Our new paradigm presents region based P300 spellers with only audio, only visual, and audio-visual stimuli. Audio-visual P300 speller structure is the newest model for region based spellers. The subject focused on the desired character stimulus. We used the stepwise linear discriminant analysis method for classification that either included the desired P300 signal or not. According to stepwise linear discriminant analysis, the mean classification accuracy value of the experiment was 90.31% with the audio-visual region based P300 speller. With this new paradigm, classification accuracy in the audio-visual P300 speller was improved 15.69% and 66,99% according to the visual only and audio only P300 speller that we used in the experiments, respectively.

Cultivar classification of cloudy apple juices from substandard fruits in China based on aroma profile analyzed by HS-SPME/GC-MS

Aroma profiles of cloudy apple juices from substandard apple with different cultivars were investigated by headspace solid phase micro-extraction (HS-SPME) combined with gas chromatography-mass spectrometry (GC-MS). 38 volatile compounds (including esters, alcohols, acids, aldehydes, terpenes, etc.) were identified. 30 Cloudy apple juices samples of 5 cultivars in China were characterized and classified by different chemometric analysis based on volatile compounds. Discrimination models were established based on volatile profiles of cloudy apple juices by chemometric techniques including principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA) and stepwise linear discriminant analysis (SLDA). Satisfactory discriminations were achieved by SLDA based on volatile compounds according to cultivar with prediction accuracy of 93.3%. The results suggested that SLDA could be used to discriminate substandard cloudy apple juices based on the volatile compounds. The obtained results could also provide data basis for selecting substandard apple fruits of different cultivars for juice production with better aroma profile.

Portable Drowsiness Detection through Use of a Prefrontal Single-Channel Electroencephalogram.

Drowsiness detection has been studied in the context of evaluating products, assessing driver alertness, and managing office environments. Drowsiness level can be readily detected through measurement of human brain activity. The electroencephalogram (EEG), a device whose application relies on adhering electrodes to the scalp, is the primary method used to monitor brain activity. The many electrodes and wires required to perform an EEG place considerable constraints on the movement of users, and the cost of the device limits its availability. For these reasons, conventional EEG devices are not used in practical studies and businesses. Many potential practical applications could benefit from the development of a wire-free, low-priced device; however, it remains to be elucidated whether portable EEG devices can be used to estimate human drowsiness levels and applied within practical research settings and businesses. In this study, we outline the development of a drowsiness detection system that makes use of a low-priced, prefrontal single-channel EEG device and evaluate its performance in an offline analysis and a practical experiment. Firstly, for the development of the system, we compared three feature extraction methods: power spectral density (PSD), autoregressive (AR) modeling, and multiscale entropy (MSE) for detecting characteristics of an EEG. In order to efficiently select a meaningful PSD, we utilized step-wise linear discriminant analysis (SWLDA). Time-averaging and robust-scaling were used to fit the data for pattern recognition. Pattern recognition was performed by a support vector machine (SVM) with a radial basis function (RBF) kernel. The optimal hyperparameters for the SVM were selected by the grind search method so as to increase drowsiness detection accuracy. To evaluate the performance of the detections, we calculated classification accuracy using the SVM through 10-fold cross-validation. Our model achieved a classification accuracy of 72.7% using the PSD with SWLDA and the SVM. Secondly, we conducted a practical study using the system and evaluated its performance in a practical situation. There was a significant difference (* p < 0.05) between the drowsiness-evoked task and concentration-needed task. Our results demonstrate the efficacy of our low-priced portable drowsiness detection system in quantifying drowsy states. We anticipate that our system will be useful to practical studies with aims as diverse as measurement of classroom mental engagement, evaluation of movies, and office environment evaluation.

Revisiting the drivers of at-a-station hydraulic geometry in stream reaches

Knowledge of temporal variations in river hydraulic characteristics (i.e., width, depth, velocity) across stream networks is a key element for catchment management because hydraulics influence physical habitats and biodiversity, water temperature, nutrient fluxes, sediment transport and associated ecosystem services. Reach scale at-a-station hydraulic geometry describes the variation of mean water depth and width with discharge within a stream reach. The ability to understand and predict the form of these hydraulic geometry relationships across river networks therefore represents a powerful tool for river managers. Although variations of these relationships between rivers have been widely documented, they are still largely unexplained. This study provides an analysis of a unique data set of hydraulic geometry collected in 492 stream reaches in France. Relationships between the hydraulic geometry parameters and many variables describing catchment- and reach-scale characteristics of reaches were analyzed using stepwise linear regression and discriminant analysis. Results show that width-discharge relationships across reaches are well predicted by longitudinal and lateral descriptors of reaches: the ratio of median to bankfull width, the median Froude number, the width to depth ratio, the active channel relative to the catchment area and the fluvial pattern (braided or not). Fluvial pattern, reach averaged active channel width and bankfull width can be accurately estimated using aerial imagery, making application of our models over river networks possible. Fewer relationships significantly explained depth-discharge relationships, despite a weak influence of reach Froude number, flow resistance (i.e., relative grain size, relative bedform submergence) and width to depth ratio. Unavailable descriptors of channel resistance would probably improve these models (e.g., describing bank stability, spill and form resistance). Despite the uncertainties of hydraulic geometry relationships, this study shows that predictions of hydraulic geometry exponents across river networks are possible, offering a large range of potential applications.

Establishing the Phenolic Composition of Olea europaea L. Leaves from Cultivars Grown in Morocco as a Crucial Step Towards Their Subsequent Exploitation.

In Morocco, the recovery of olive agro-industrial by-products as potential sources of high-added value substances has been underestimated so far. A comprehensive quantitative characterization of olive leaves’ bioactive compounds is crucial for any attempt to change this situation and to implement the valorization concept in emerging countries. Thus, the phenolic fraction of olive leaves of 11 varieties (‘Arbequina’, ‘Hojiblanca’, ‘Frantoio’, ‘Koroneiki’, ‘Lechín’, ‘Lucque’, ‘Manzanilla’, ‘Picholine de Languedoc’, ‘Picholine Marocaine’, ‘Picual’ and ‘Verdal’), cultivated in the Moroccan Meknès region, was investigated. Thirty eight phenolic or related compounds (including 16 secoiridoids, nine flavonoids in their aglycone form, seven flavonoids in glycosylated form, four simple phenols, one phenolic acid and one lignan) were determined in a total of 55 samples by using ultrasonic-assisted extraction and liquid chromatography coupled to electrospray ionization-ion trap mass spectrometry (LC-ESI-IT MS). Very remarkable quantitative differences were observed among the profiles of the studied cultivars. ‘Picholine Marocaine’ variety exhibited the highest total phenolic content (around 44 g/kg dry weight (DW)), and logically showed the highest concentration in terms of various individual compounds. In addition, chemometrics (principal components analysis (PCA) and stepwise-linear discriminant analysis (s-LDA)) were applied to the quantitative phenolic compound data, allowing good discrimination of the selected samples according to their varietal origin.

Single-Trial Classification of Disfluent Brain States in Adults Who Stutter.

Normal human speech requires precise coordination between motor planning and sensory processing. Speech disfluencies are common when children learn to talk, but usually abate with time. About 5% of children experience stuttering. For most, this resolves within a year. However, for approximately 1% of the world population, stuttering continues into adulthood, which is termed 'persistent developmental stuttering'. Most stuttering events occur at the beginning of an utterance. So, in principle, brain activity before speaking should differ between fluent and stuttered speech. Here we present a method for classifying brain network states associated with fluent vs. stuttered speech on a single trial basis. Brain activity was recorded with EEG before people who stutter read aloud pseudo-word pairs. Offline independent component analysis (ICA) was used to identify the independent neural sources that underlie speech preparation. A time window selection algorithm extracted spectral power and coherence data from salient windows specific to each neural source. A stepwise linear discriminant analysis (sLDA) algorithm predicted fluent vs. stuttered speech for 81% of trials in two subjects. These results support the feasibility of developing a brain-computer interface (BCI) system to detect stuttering before it occurs, with potential for therapeutic application.

Automatic Detection of Atrial Fibrillation and Other Arrhythmias in ECG Recordings Acquired by a Smartphone Device

Atrial fibrillation (AF) is the most common cardiac disease and is associated with other cardiac complications. Few attempts have been made for discriminating AF from other arrhythmias and noise. The aim of this study is to present a novel approach for such a classification in short ECG recordings acquired using a smartphone device. The implemented algorithm was tested on the Physionet Computing in Cardiology Challenge 2017 Database and, for the purpose of comparison, on the MIT-BH AF database. After feature extraction, the stepwise linear discriminant analysis for feature selection was used. The Least Square Support Vector Machine classifier was trained and cross-validated on the available dataset of the Challenge 2017. The best performance was obtained with a total of 30 features. The algorithm produced the following performance: F1 Normal rhythm = 0.92; F1 AF rhythm: 0.82; F1 Other rhythm = 0.75; Global F1 = 0.83, obtaining the third best result in the follow-up phase of the Physionet Challenge. On the MIT-BH ADF database the algorithm gave the following performance: F1 Normal rhythm = 0.98; F1 AF rhythm: 0.99; Global F1 = 0.98. Since the algorithm reliably detect AF and other rhythms in smartphone ECG recordings, it could be applied for personal health monitoring systems.