S. Van Huffel

Canonical Correlation Analysis Applied to Remove Muscle Artifacts From the Electroencephalogram

The electroencephalogram (EEG) is often contaminated by muscle artifacts. In this paper, a new method for muscle artifact removal in EEG is presented, based on canonical correlation analysis (CCA) as a blind source separation (BSS) technique. This method is demonstrated on a synthetic data set. The method outperformed a low-pass filter with different cutoff frequencies and an independent component analysis (ICA)-based technique for muscle artifact removal. In addition, the method is applied on a real ictal EEG recording contaminated with muscle artifacts. The proposed method removed successfully the muscle artifact without altering the recorded underlying ictal activity

Simple ultrasound-based rules for the diagnosis of ovarian cancer

To derive simple and clinically useful ultrasound‐based rules for discriminating between benign and malignant adnexal masses.

Source Separation From Single-Channel Recordings by Combining Empirical-Mode Decomposition and Independent Component Analysis

In biomedical signal processing, it is often the case that many sources are mixed into the measured signal. The goal is usually to analyze one or several of them separately. In the case of multichannel measurements, several blind source separation techniques are available for decomposing the signal into its components [e.g., independent component analysis (ICA)]. However, only a few techniques have been reported for analyses of single-channel recordings. Examples are single-channel ICA (SCICA) and wavelet-ICA (WICA), which all have certain limitations. In this paper, we propose a new method for a single-channel signal decomposition. This method combines empirical-mode decomposition with ICA. We compare the separation performance of our algorithm with SCICA and WICA through simulations, and we show that our method outperforms the other two, especially for high noise-to-signal ratios. The performance of the new algorithm was also demonstrated in two real-life applications.

Brain tumor classification based on long echo proton MRS signals

There has been a growing research interest in brain tumor classification based on proton magnetic resonance spectroscopy (1H MRS) signals. Four research centers within the EU funded INTERPRET project have acquired a significant number of long echo 1H MRS signals for brain tumor classification. In this paper, we present an objective comparison of several classification techniques applied to the discrimination of four types of brain tumors: meningiomas, glioblastomas, astrocytomas grade II and metastases. Linear and non-linear classifiers are compared: linear discriminant analysis (LDA), support vector machines (SVM) and least squares SVM (LS-SVM) with a linear kernel as linear techniques and LS-SVM with a radial basis function (RBF) kernel as a non-linear technique. Kernel-based methods can perform well in processing high dimensional data. This motivates the inclusion of SVM and LS-SVM in this study. The analysis includes optimal input variable selection, (hyper-) parameter estimation, followed by performance evaluation. The classification performance is evaluated over 200 stratified random samplings of the dataset into training and test sets. Receiver operating characteristic (ROC) curve analysis measures the performance of binary classification, while for multiclass classification, we consider the accuracy as performance measure. Based on the complete magnitude spectra, automated binary classifiers are able to reach an area under the ROC curve (AUC) of more than 0.9 except for the hard case glioblastomas versus metastases. Although, based on the available long echo 1H MRS data, we did not find any statistically significant difference between the performances of LDA and the kernel-based methods, the latter have the strength that no dimensionality reduction is required to obtain such a high performance.

Which extrauterine pelvic masses are difficult to correctly classify as benign or malignant on the basis of ultrasound findings and is there a way of making a correct diagnosis

To determine which extrauterine pelvic masses are difficult to correctly classify as benign or malignant on the basis of ultrasound findings, and to determine if the use of logistic regression models for calculation of individual risk of malignancy would improve the diagnostic accuracy in difficult tumors.

Ovarian cancer prediction in adnexal masses using ultrasound‐based logistic regression models: a temporal and external validation study by the IOTA group

The aims of the study were to temporally and externally validate the diagnostic performance of two logistic regression models containing clinical and ultrasound variables in order to estimate the risk of malignancy in adnexal masses, and to compare the results with the subjective interpretation of ultrasound findings carried out by an experienced ultrasound examiner (‘subjective assessment’).

The MCA EXIN neuron for the minor component analysis

The minor component analysis (MCA) deals with the recovery of the eigenvector associated to the smallest eigenvalue of the autocorrelation matrix of the input data and is a very important tool for signal processing and data analysis. It is almost exclusively solved by linear neurons. This paper presents a linear neuron endowed with a novel learning law, called MCA EXINn and analyzes its features. The neural literature about MCA is very poor, in the sense that both a little theoretical basis is given (almost always focusing on the ODE asymptotic approximation) and only experiments on toy problems (at most four-dimensional problems) are presented, without any numerical analysis. This work addresses these problems and lays sound theoretical foundations for the neural MCA theory. In particular, it classifies the MCA neurons according to the Riemannian metric and justifies, from the analysis of the degeneracy of the error cost; the different behavior in approaching convergence. The cost landscape is studied and used as a basis for the analysis of the asymptotic behavior. All the phases of the dynamics of the MCA algorithms are investigated in detail and, together with the numerical analysis, lead to the identification of three possible kinds of divergence, here called sudden, dynamic, and numerical. The importance of the choice of low initial conditions is also explained. A lot of importance is given to the experimental part, where simulations on high-dimensional problems are,presented and analyzed. The orthogonal regression or total least squares (TLS) technique is also presented, together with a real-world application on the identification of the parameters of an electrical machine. It can be concluded that MCA EXIN is the best MCA neuron in terms of stability (no finite time divergence), speed, and accuracy.

Automated neonatal seizure detection mimicking a human observer reading EEG

OBJECTIVE The description and evaluation of a novel patient-independent seizure detection for the EEG of the newborn term infant. METHODS We identified characteristics of neonatal seizures by which a human observer is able to detect them. Neonatal seizures were divided into two types. For each type, a fully automated detection algorithm was developed based on the identified human observer characteristics. The first algorithm analyzes the correlation between high-energetic segments of the EEG. The second detects increases in low-frequency activity (<8 Hz) with high autocorrelation. RESULTS The complete algorithm was tested on multi-channel EEG recordings of 21 patients with and 5 patients without electrographic seizures, totaling 217 h of EEG. Sensitivity of the combined algorithms was found to be 88%, Positive Predictive Value (PPV) 75% and the false positive rate 0.66 per hour. CONCLUSIONS Our approach to separate neonatal seizures into two types yields a high sensitivity combined with a good PPV and much lower false positive rate than previously published algorithms. SIGNIFICANCE The proposed algorithm significantly improves neonatal seizure detection and monitoring.

Formulation and solution of structured total least norm problems for parameter estimation

The total least squares (TLS) method is a generalization of the least squares (LS) method for solving overdetermined sets of linear equations Ax/spl ap/b. The TLS method minimizes /spl par/[E|-r]/spl par//sub F/, where r=b-(A+E)x, so that (b-r)/spl isin/Range (A+E), given A/spl isin/C/sup m/spl times/n/, with m/spl ges/n and b/spl isin/C/sup m/spl times/1/. The most common TLS algorithm is based on the singular value decomposition (SVD) of [A/b]. However, the SVD-based methods may not be appropriate when the matrix A has a special structure since they do not preserve the structure. Previously, a new problem formulation known as structured total least norm (STLN), and the algorithm for computing the STLN solution, have been developed. The STLN method preserves the special structure of A or [A/b] and can minimize the error in the discrete L/sub p/ norm, where p=1, 2 or /spl infin/. In this paper, the STLN problem formulation is generalized for computing the solution of STLN problems with multiple right-hand sides AX/spl ap/B. It is shown that these problems can be converted to ordinary STLN problems with one right-hand side. In addition, the method is shown to converge to the optimal solution in certain model reduction problems. Furthermore, the application of the STLN method to various parameter estimation problems is studied in which the computed correction matrix applied to A or [A/B] keeps the same Toeplitz structure as the data matrix A of [A/B], respectively. In particular, the L/sub 2/ norm STLN method is compared with the LS and TLS methods in deconvolution, transfer function modeling, and linear prediction problems.

Influence of Mental Stress on Heart Rate and Heart Rate Variability

Stress is a huge problem in today’s society. Being able to measure stress, therefore, may help to address this problem. Although stress has a psychological origin, it affects several physiological processes in the human body: increased muscle tension in the neck, change in concentration of several hormones and a change in heart rate (HR) and heart rate variability (HRV). The brain innervates the heart by means of stimuli via the Autonomic Nervous System (ANS), which is divided into sympathetic and parasympathetic branches. The sympathetic activity leads to an increase in HR (e.g. during sports exercise), while parasympathetic activity induces a lower HR (e.g. during sleep). The two circuits are constantly interacting and this interaction is reflected in HRV. HRV, therefore, provides a measure to express the activity of the ANS, and may consequently provide a measure for stress. We therefore explored measures of HR and HRV with an imposed stressful situation. We recorded changes in HR and HRV in a group of 28 subjects at rest, and with a mental stressor. The results suggest that HR and HRV change with a mental task. HR and HRV recordings may have the potential, therefore, to measure stress levels and guide preventive measures to reduce stress related illnesses.