Comparison of t -test ranking with PCA and SEPCOR feature selection for wake and stage 1 sleep pattern recognition in multichannel electroencephalograms

Upper-Limb Motion Recognition Based on Hybrid Feature Selection: Algorithm Development and Validation

Background

For rehabilitation training systems, it is essential to automatically record and recognize exercises, especially when more than one type of exercise is performed without a predefined sequence. Most motion recognition methods are based on feature engineering and machine learning algorithms. Time-domain and frequency-domain features are extracted from original time series data collected by sensor nodes. For high-dimensional data, feature selection plays an important role in improving the performance of motion recognition. Existing feature selection methods can be categorized into filter and wrapper methods. Wrapper methods usually achieve better performance than filter methods; however, in most cases, they are computationally intensive, and the feature subset obtained is usually optimized only for the specific learning algorithm.

Objective

This study aimed to provide a feature selection method for motion recognition of upper-limb exercises and improve the recognition performance.

Methods

Motion data from 5 types of upper-limb exercises performed by 21 participants were collected by a customized inertial measurement unit (IMU) node. A total of 60 time-domain and frequency-domain features were extracted from the original sensor data. A hybrid feature selection method by combining filter and wrapper methods (FESCOM) was proposed to eliminate irrelevant features for motion recognition of upper-limb exercises. In the filter stage, candidate features were first selected from the original feature set according to the significance for motion recognition. In the wrapper stage, k-nearest neighbors (kNN), Naïve Bayes (NB), and random forest (RF) were evaluated as the wrapping components to further refine the features from the candidate feature set. The performance of the proposed FESCOM method was verified using experiments on motion recognition of upper-limb exercises and compared with the traditional wrapper method.

Results

Using kNN, NB, and RF as the wrapping components, the classification error rates of the proposed FESCOM method were 1.7%, 8.9%, and 7.4%, respectively, and the feature selection time in each iteration was 13 seconds, 71 seconds, and 541 seconds, respectively.

Conclusions

The experimental results demonstrated that, in the case of 5 motion types performed by 21 healthy participants, the proposed FESCOM method using kNN and NB as the wrapping components achieved better recognition performance than the traditional wrapper method. The FESCOM method dramatically reduces the search time in the feature selection process. The results also demonstrated that the optimal number of features depends on the classifier. This approach serves to improve feature selection and classification algorithm selection for upper-limb motion recognition based on wearable sensor data, which can be extended to motion recognition of more motion types and participants.

Liquid Cathode Glow Discharge as an Excitation Source for the Analysis of Complex Water Samples with Atomic Emission Spectrometry

A liquid cathode glow discharge (LCGD) was developed as a low-power and miniaturized excitation source of atomic emission spectrometry (AES) for the determination of K, Na, Ca, and Mg in water samples from rivers and lakes. The discharge stability and parameter influencing the analytical performance of LCGD-AES were systematically examined. Moreover, the measurement results of water samples using LCGD-AES were verified by ion chromatography (IC). The results showed that the optimized operating parameters are a 660 V discharge voltage, pH = 1.0 HNO3 as the supporting electrolyte, and a 4.0 mL min-1 solution flow rate. High concentrations of some metals may interfere with the detection of Ca and Mg. Low-molecular-weight organic substances do not have a remarkable enhancement on signal intensity. With the addition of 0.5% cetyltrimethylammonium chloride (CTAC), the emission intensity of elements can enhance significantly. However, it is not used to further evaluate the analytical performance due to instability of plasma after adding CTAC. The maximum power of LCGD is 52 W. The limits of detection and precision (RSD, in 1 mg L-1) of K, Na, Ca, and Mg are 0.20, 0.02, 0.01, and 0.01 mg L-1 and 0.9, 1.5, 0.6, and 1.2%, respectively. The measurement results of K, Na, Ca, and Mg in water samples by LCGD-AES are basically in agreement with the reference values measured by IC. The recovery of samples ranged from 84 to 113% except for Na, suggesting that the measurement results have high accuracy and reliability. All the results indicated that the LCGD-AES can provide an alternative analytical method for in situ, real-time, on-line determination of K, Na, Ca, and Mg in water samples from rivers and lakes.

Analysis of statistical coefficients and autoregressive parameters over intrinsic mode functions (IMFs) for epileptic seizure detection

Epilepsy is a persistent neurological disorder impacting over 50 million people around the world. It is characterized by repeated seizures defined as brief episodes of involuntary movement that might entail the human body. Electroencephalography (EEG) signals are usually used for the detection of epileptic seizures. This paper introduces a new feature extraction method for the classification of seizure and seizure-free EEG time segments. The proposed method relies on the empirical mode decomposition (EMD), statistics and autoregressive (AR) parameters. The EMD method decomposes an EEG time segment into a finite set of intrinsic mode functions (IMFs) from which statistical coefficients and autoregressive parameters are computed. Nevertheless, the calculated features could be of high dimension as the number of IMFs increases, the Student's t-test and the Mann-Whitney U test were thus employed for features ranking in order to withdraw lower significant features. The obtained features have been used for the classification of seizure and seizure-free EEG signals by the application of a feed-forward multilayer perceptron neural network (MLPNN) classifier. Experimental results carried out on the EEG database provided by the University of Bonn, Germany, demonstrated the effectiveness of the proposed method which performance assessed by the classification accuracy (CA) is compared to other existing performances reported in the literature.

Analysis of Streamflow Complexity Based on Entropies in the Weihe River Basin, China

The study on the complexity of streamflow has guiding significance for hydrologic simulation, hydrologic prediction, water resources planning and management. Utilizing monthly streamflow data from four hydrologic control stations in the mainstream of the Weihe River in China, the methods of approximate entropy, sample entropy, two-dimensional entropy and fuzzy entropy are introduced into hydrology research to investigate the spatial distribution and dynamic change in streamflow complexity. The results indicate that the complexity of the streamflow has spatial differences in the Weihe River watershed, exhibiting an increasing tendency along the Weihe mainstream, except at the Linjiacun station, which may be attributed to the elevated anthropogenic influence. Employing sliding entropies, the variation points of the streamflow time series at the Weijiabu station were identified in 1968, 1993 and 2003, and those at the Linjiacun station, Xianyang station and Huaxian station occurred in 1971, 1993 and 2003. In the verification of the above points, the minimum value of t-test is 3.7514, and that of Brown-Forsythe is 7.0307, far exceeding the significance level of 95%. Also, the cumulative anomaly can detect two variation points. The t-test, Brown-Forsythe test and cumulative anomaly test strengthen the conclusion regarding the availability of entropies for identifying the streamflow variability. The results lead us to conclude that four entropies have good application effects in the complexity analysis of the streamflow time series. Moreover, two-dimensional entropy and fuzzy entropy, which have been rarely used in hydrology research before, demonstrate better continuity and relative consistency, are more suitable for short and noisy hydrologic time series and more effectively identify the streamflow complexity. The results could be very useful in identifying variation points in the streamflow time series.

Data mining based approach to study the effect of consumption of caffeinated coffee on the generation of the steady-state visual evoked potential signals

The steady-state visual evoked potentials (SSVEP), are elicited at the parieto-occipital region of the cortex when a light source (3.5-75 Hz), flickering at a constant frequency, stimulates the retinal cells. In the last few decades, researchers have reported that caffeine enhances the vigilance and the executive control of visual attention. However, no study has investigated the effect of caffeinated coffee on the SSVEP response, which is used for controlling the brain-computer interface (BCI) devices for rehabilitative applications. The current work proposes a data mining-based approach to gain insight into the alterations in the SSVEP signals after the consumption of caffeinated coffee. Recurrence quantification analysis (RQA) of the electroencephalogram (EEG) signals was employed for this purpose. The EEG signals were acquired at seven frequencies of photic stimuli. The stimuli frequencies were chosen such that they were distributed throughout the EEG frequency bands. The prominent SSVEP signals were identified using the Canonical Correlation Analysis (CCA) method. Several statistical features were extracted from the recurrence plot of the SSVEP signals. Statistical analyses using the t-test and decision tree-based methods helped to select the most relevant features, which were then classified using Automated Neural Network (ANN). The relevant features could be classified with a maximum accuracy of 97%. This supports our hypothesis that the consumption of caffeinated coffee can alter the SSVEP response. In conclusion, utmost care should be taken in selecting the features for designing BCI devices.

Risk assessment of soil heavy metals associated with land use variations in the riparian zones of a typical urban river gradient

Urbanization-induced land use changes in riparian area alter soil and water regimes in complex ways, which may also affect the migration and transformation of soil heavy metals and increase the risk of release. In this study, soil samples from the riparian zone of Beiyun River, which located in the rapidly urbanized Beijing metropolis, were collected and analyzed for heavy metals (As, Cd, Cr, Cu, Mn, Ni, Pb, and Zn). Then their zoning distribution pattern along river (section 1 to section 4 from upper to low reaches) and the correlation of heavy metals between riparian soils and riverine sediments were investigated. Results showed that the average soil heavy metal concentrations of Cd, Cr, Cu and Zn in riparian zone were approximately 2.2, 1.7, 1.9 and 2.0 times higher than the background values. Sectionally, the concentrations of Cd, Ni, Pb and Zn displayed a decreasing order with section 2 > section 3 > section 4 > section 1, while the highest values of Cr and Cu were found in section 3. The concentrations of all heavy metals except Cr in artificial garden land were higher than those in other land use types, and the concentrations of Cr among five land use types were in the order of grass land > farmland > artificial garden land > forest land > forest-grass land. Generally, most of the heavy metals in the riverine sediments had higher contents than those in riparian zones, especially Cu and Zn. There was a decreasing order for the average geo-accumulation index (I_geo) of measured heavy metals in the soils of riparian zone: Zn (0.15) > Cr (0.08) > Cu (0.07) > Cd (-0.08) > As (-0.57) > Pb (-0.67) > Mn (-0.75) > Ni (-0.86), whereas they had different "high-low" orders in different land use types. The I_geo index indicated most regions of riparian zone were moderately polluted with Cd, Cr, Cu and Zn, especially in grass land and forest land. Also, Cd, Cr and Zn in riparian zone have positive relationships with the concentrations in riverine sediments. Health risk assessment showed that the contribution of ingestion HQ to HI was the highest among the three exposure pathways (ingestion, inhalation and dermal contact), and children had higher non-carcinogenic risk and carcinogenic risk index than adult. Our findings suggest that land use and soil in riparian zone should be protected and managed scientifically to control the riverine pollution and ensure human health.

Improving Generalization Based on l1-Norm Regularization for EEG-Based Motor Imagery Classification

Multichannel electroencephalography (EEG) is widely used in typical brain-computer interface (BCI) systems. In general, a number of parameters are essential for a EEG classification algorithm due to redundant features involved in EEG signals. However, the generalization of the EEG method is often adversely affected by the model complexity, considerably coherent with its number of undetermined parameters, further leading to heavy overfitting. To decrease the complexity and improve the generalization of EEG method, we present a novel l₁-norm-based approach to combine the decision value obtained from each EEG channel directly. By extracting the information from different channels on independent frequency bands (FB) with l1-norm regularization, the method proposed fits the training data with much less parameters compared to common spatial pattern (CSP) methods in order to reduce overfitting. Moreover, an effective and efficient solution to minimize the optimization object is proposed. The experimental results on dataset IVa of BCI competition III and dataset I of BCI competition IV show that, the proposed method contributes to high classification accuracy and increases generalization performance for the classification of MI EEG. As the training set ratio decreases from 80 to 20%, the average classification accuracy on the two datasets changes from 85.86 and 86.13% to 84.81 and 76.59%, respectively. The classification performance and generalization of the proposed method contribute to the practical application of MI based BCI systems.

Spatial Distribution of Metals and Associated Risks in Surface Sediments Along a Typical Urban River Gradient in the Beijing Region

Surface sediments from Beiyun River located in the rapidly urbanized Beijing metropolis were collected and analyzed for heavy metals (As, Cd, Cr, Cu, Mn, Ni, Pb, and Zn) to investigate their spatial distribution pattern, ecotoxicology and source identification. Results indicated the average heavy metal concentrations of Cd, Cr, Cu, Pb, and Zn were approximately 4, 2, 3, 2, and 4 times higher than their background values. Spatially, we found the concentrations of heavy metals made significant change in four sections along urbanized river gradients. The contents in midstream of urban region and farmland region (section 2 and section 3) were greater than those in upstream (section 1) and downstream (section 4). However, one-way analysis of variance for spatial analysis suggested there were no significant differences between mainstream and tributaries. The geo-accumulation index (I _geo) used to assess the sediment quality exhibited there was a decreasing order for the average I _geo of measured heavy metals: Zn (0.82) > Cd (0.53) > Cu (0.50) > Cr (- 0.08) > Pb (- 0.45) > Ni (- 0.96) > Mn (- 0.97) > As (- 1.01), whereas they had different "high-low" orders at different sampling transects. Ecological risk index values showed that section 2 and section 3 revealed a high and moderate ecological risk, respectively. Furthermore, principal component analysis indicated the first principle component explained 64.73% of total variance with the main pollutants of As, Cr, Ni, Zn, and Cu which were probably controlled by the mixed sources covering natural factors and anthropogenic input.

High-Sensitivity Determination of K, Ca, Na, and Mg in Salt Mines Samples by Atomic Emission Spectrometry with a Miniaturized Liquid Cathode Glow Discharge

An atomic emission spectrometer (AES) based on a novel atmospheric pressure liquid cathode glow discharge (LCGD) as one of the most promising miniaturized excitation sources has been developed, in which the glow discharge is produced between a needle-like Pt anode and the electrolyte (as cathode) overflowing from a quartz capillary. Lower energy consumption (<50 W) and higher excitation efficiency can be realized by point discharge of the needle-like Pt. The miniaturized LCGD seems particularly well suited to rapid and high-sensitivity determination of K, Ca, Na, and Mg in salt mines samples. The optimized analytical conditions of LCGD-AES were pH = 1 with HNO₃ as electrolyte, 650 V discharge voltage, and 3 mL min^-1 solution flow rate. The limits of detections (LODs) of K, Ca, Na, and Mg were 0.390, 0.054, 0.048, and 0.032 mg L^-1, respectively. Measurement results of the LCGD-AES are in good agreement with the comparison value obtained by inductively coupled plasma (ICP) and ion chromatography (IC). All results suggested that the developed portable analytical instrument can be used for on-site and real-time monitoring of metal elements in field with further improvement.