Biomedical subjects
Publications and source records attributed to J Vandewalle.
In this paper, we propose the emerging technique of independent component analysis, also known as blind source separation, as an interesting tool for the extraction of the antepartum fetal electrocardiogram from multilead cutaneous potential recordings. The technique is illustrated by means of a real-life example.
Explore the source record for details and available documents.
OBJECTIVE: The aim of this study was to generate and evaluate artificial neural network (ANN) models from simple clinical and ultrasound-derived criteria to predict whether or not an adnexal mass will have histological evidence of malignancy. DESIGN: The data were collected prospectively from 173 consecutive patients who were scheduled to undergo surgical investigations at the University Hospitals, Leuven, between August 1994 and August 1996. The outcome measure was the histological classification of excised tissues as malignant (including borderline) or benign. METHODS: Age, menopausal status and serum CA 125 levels and sonographic features of the adnexal mass were encoded as variables. The ANNs were trained on a randomly selected set of 116 patient records and tested on the remainder (n = 57). The performance of each model was evaluated using receiver operating characteristic (ROC) curves and compared with corresponding data from an established risk of malignancy index (RMI) and a logistic regression model. RESULTS: There were 124 benign masses, five of borderline malignancy and 44 invasive cancers (of which 29% were metastatic); 37% of patients with a malignant or borderline tumor had stage I disease. The best ANN gave an area under the ROC curve of 0.979 for the whole dataset, a sensitivity of 95.9% and specificity of 93.5%. The corresponding values for the RMI were 0.882, 67.3% and 91.1%, and for the logistic regression model 0.956, 95.9% and 85.5%, respectively. CONCLUSION: An ANN can be trained to provide clinically accurate information, on whether or not an adnexal mass is malignant, from the patient's menopausal status, serum CA 125 levels, and some simple ultrasonographic criteria.
A neural network approach is proposed for the automated classification of the normal and abnormal EGG. Two learning algorithms, the quasi-Newton and the scaled conjugate gradient method for the multilayer feedforward neural networks (MFNN), are introduced and compared with the error backpropagation algorithm. The configurations of the MFNN are determined by experiment. The raw EGG data, its power spectral data, and its autoregressive moving average (ARMA) modelling parameters are used as the input to the MFNN and compared with each other. Three indexes (the percent correct, sum-squared error and complexity per iteration) are used to evaluate the performance of each learning algorithm. The results show that the scaled conjugate gradient algorithm performs best, in that it is robust and provides a super-linear convergence rate. The power spectral representation and the ARMA modelling parameters of the EGG are found to be better types of the input to the network for this specific application, both yielding a percent correctness of 95% on the test set. Although the results are focused on the classification of the EGG, this paper should provide useful information for the classification of other biomedical signals.
An algorithm has been developed for the offline analysis of prolonged manometric recordings in the upper small intestine of humans. Sample data are acquired in the human duodenum and jejunum six solid-state strain-gauge transducers mounted on a silicon catheter that is connected to a portable digital recording device. The data are sampled at 4 Hz and filtered. For accurate calculations, the filtered signals are converted to cubic B-spline functions of order four. Based on an exponential weighted moving average, a base-line is calculated from the signal. Contractions are recognised on the basis of thresholds for minimum amplitude and duration. The developed algorithm calculates properties of these contractions, such as amplitude, duration, area and a motility index. In addition, the program automatically recognises normal motor patterns of the fasted human small intestine, such as the migrating motor complex, and aids in the identification of the postprandial motor pattern. Motor patterns are defined in terms of properties such as contraction frequency and propagation. In a validation procedure using conventional manual analysis, the program correctly identifies the number of individual contractions with a 98% confidence interval and also correctly recognises 96% of phase 3 motor activity.
In this paper, we present a new model of discrete neural associative memories and its design rule. The most important feature of this new model is that a static mapping instead of the dynamic convergent process is used to retrieve the stored messages. The new model features a two-layer structure, with feedforward connections only and uses two kinds of neurons which implement different output functions. Another important feature is that this new model employs an extremely simple weight setup rule and all the resulted weights can only assume two different values, -1 and +1, which facilitates the VLSI implementation. Compared to the famous discrete Hopfield model designed with the well-known Hebbian rule or any other rule, the new model can guarantee all the given patterns to be stored as fixed points. Moreover, each fixed point is surrounded by an attraction basin (which is a ball in the Hamming distance sense) with the maximal possible radius. The performances of the new model are compared through some illustrative examples with those of the Hopfield associative memory designed using different methods.
Using the property of universal approximation of multilayer perceptron neural network, a class of discrete nonlinear dynamical systems are modeled by a perceptron with two hidden layers. A backpropagation algorithm is then used to train the model to identify the nonlinear systems to a desired level of accuracy. Based on the identified model, a one-step-ahead predictive control scheme is proposed in which the future control inputs are obtained through some nonlinear optimization process. Making use of the online learning properties of neural networks, the predictive control scheme is further developed into an adaptive one which is robust to the incompleteness of identification. Simulation results show that this neural control scheme works well even for some very complicated nonlinear systems.
This paper tries to demonstrate how a heuristic neural control approach can be used to solve a complex nonlinear control problem. The control task is to swing up a pendulum mounted on a cart from its stable position (vertically down) to the zero state (up right) and keep it there by applying a sequence of two opposing constant forces of equal magnitude to the mass center of the cart. In addition, the displacement of the cart itself is confined to within a preset limit during the swinging up action and it will eventually be brought to the origin of the track. This is truly a nontrivial nonlinear regulation problem and is considerably difficult compared to the pendulum balancing problem (and its variations) widely adopted as a benchmarking test system for neural controllers. Through the solution of this specific control problem, we try to illustrate a heuristic neural control approach with task decomposition, control rule extraction and neural net rule implementation as its basic elements. Specializing to the pendulum problem, the global control task is decomposed into subtasks namely pendulum positioning and cart positioning. Accordingly, three separate neural subcontrollers are designed to cater to the subtasks and their coordination, i.e., pendulum subcontroller (PSC), cart subcontroller (CSC) and the switching subcontroller (SSC). Each of the subcontrollers is designed based on the rules and guidelines obtained from the experiences of a human operator. The simulation result is included to show the actual performance of the controller.
The mathematical model for the penicillin G fed-batch fermentation proposed by Heijnen et al. (1979) is compared with the model of Bajpai & Reuss (1980). Although the general structure of these models is similar, the difference in metabolic assumptions and specific growth and production kinetics results in a completely different behaviour towards product optimization. A detailed analysis of both models reveals some physical and biochemical shortcomings. It is shown that it is impossible to make a reliable estimation of the model parameters, only using experimental data of simple constant glucose feed rate fermentations with low initial substrate amount. However, it is demonstrated that some model parameters might be key factors in concluding whether or not altering the substrate feeding strategy has an important influence on the final amount of product. It is illustrated that feeding strategy optimization studies can be a tool in designing experiments for parameter estimation purposes.
Many sigmoidal functions to describe a bacterial growth curve as an explicit function of time have been reported in the literature. Furthermore, several expressions have been proposed to model the influence of temperature on the main characteristics of this growth curve: maximum specific growth rate, lag time, and asymptotic level. However, as the predictive value of such explicit models is most often guaranteed only at a constant temperature within the temperature range of microbial growth, they are less appropriate in optimization studies of a whole production and distribution chain. In this paper a dynamic mathematical model--a first-order differential equation--has been derived, describing the bacterial population as a function of both time and temperature. Furthermore, the inactivation of the population at temperatures above the maximum temperature for growth has been incorporated. In the special case of a constant temperature, the solution coincides exactly with the corresponding Gompertz model, which has been validated in several recent reports. However, the main advantage of this dynamic model is its ability to deal with time-varying temperatures, over the whole temperature range of growth and inactivation. As such, it is an essential building block in (time-saving) simulation studies to design, e.g., optimal temperature-time profiles with respect to microbial safety of a production and distribution chain of chilled foods.
The recording of the human gastric myoelectrical activity by means of cutaneous electrodes is called electrogastrography (EGG). It provides a noninvasive method of studying electrogastric behaviour. The normal frequency of the gastric signal is about 0.05 Hz. However, sudden changes of its frequency have been observed and are generally considered to be related to gastric motility disorders. Thus, spectral analysis, especially online spectral analysis, can serve as a valuable tool for practical purposes. The paper presents a new method of the adaptive spectral analysis of cutaneous electrogastric signals using autoregressive moving average (ARMA) modelling. It is based on an adaptive ARMA filter and provides both time and frequency information of the signal. Its performance is investigated in comparison with the conventional FFT-based periodogram method. Its properties in tracking time-varying instantaneous frequencies are shown. Its applications to the running spectral analysis of cutaneous electrogastric signals are presented. The proposed adaptive ARMA spectral analysis method is easy to implement and is efficient in computations. The results presented in the paper show that this new method provides a better performance and is very useful for the online monitoring of cutaneous electrogastric signals.
The paper presents and compares three methods making use of the singular value decomposition (SVD) of a matrix to extract the foetal electrocardiogram (FECG) from cutaneously recorded electrode signals. The first method constructs a set of orthogonal foetal signals (the so-called principal foetal signals) from the recordings, but needs electrode positions far from the foetal heart, in addition to the abdominal electrodes that pick up a mixture of maternal and foetal electrocardiogram. An online adaptive algorithm has been developed such that a real-time implementation becomes feasible. The second method is a new online approach to a technique presented by van Oosterom. Although this method has some important drawbacks and is suboptimal as far as foetal signal-to-noise ratio is concerned, it is still very useful when only a foetal trigger is required, as the signal obtained is not a complete FECG. Finally, a third method is proposed, based on the generalised SVD and interpreted with the new concept of oriented signal-to-signal ratio. An online version is also presented for this method and some results are shown.
The electrogastric signal can be measured cutaneously on the abdomen. This is attractive because it is harmless to patients or volunteers. However, the poor quality of the cutaneous measurements necessitates signal enhancements. Hence, in this paper, an adaptive multichannel signal enhancing system is proposed. The mu-vector least mean square (LMS) algorithm is applied to adjust the weights of the adaptive filters in the system. The detailed description and the performance analysis of the system is given in the paper. Applying the proposed system, the respiratory artifact, the electrode-skin noise, some of motion artifacts, and the electrocardiography (ECG) can be efficiently reduced while the characteristics of the relevant gastric signal is less affected.
Electrogastric signals have been successfully measured both intraluminally and cutaneously. Although it has been claimed by several researchers that the propagation direction of the electrogastric activities cannot be observed from cutaneous recordings, it is the aim of the paper to show that it is feasible. The reason why the propagation direction has never been observed from cutaneous recordings is that the reported methods for the abdominal measurements are not adequate. In the paper it is pointed out that the stomach should be localised before the measurement and the electrodes should be attached along the longitudinal axis of the stomach.
An algorithm is presented for the processing of smooth biological signals with an important implicit data reduction. It is a solid basis for an automatic analysis of these signals. The algorithm is based on the approximation of the signals with a linear combination of cubic B-splines. Depending on the parameters, a data reduction of between 8:1 and 12:1 can be achieved. The most important differences from conventional biological signal processing methods are the representation of the signal by the coefficients of the linear combination instead of a series of samples and the analysis by calculations of the time of the appearance of patterns instead of iterative searching. The approximation and analysis of long-lasting multichannel signals can be performed online with a modern microprocessor. Approximating the signal with a linear combination of cubic B-splines with equally spaced knots, according to the linear least-squares criterion gives the desired data reduction and an elegant way to perform an automatic analysis. A window calculation scheme makes it possible to handle very long signals online.
Electrogastric measurements are useful for medical research and in clinical diagnosis. The measurements, however, contain very heavy respiratory artefact. Existing conventional frequency-domain filters cannot be used because of the possible overlap of the frequencies of the gastric signals and respiratory artefact. In the paper, the methods of measuring cutaneous and intraluminal gastric signals and reference respiratory signals are described. An adaptive cancellation technique is developed, which is simple and easy to implement for online processing. It is proved by experiments to be very efficient, i.e. the respiratory artefact can be completely cancelled, while the gastric signal component is not affected. Other possible applications in biomedical signal analysis are also discussed.
An overview is given of the FEMME-project (Fetal Electrocardiogram Measuring Method and Equipment). The project started in 1981 and is, at the moment, close to producing a prototype personal computer-based system. This records simultaneously a number of cutaneous potential signals and derives from this set one or more maternal electrocardiogram-free fetal heart signals, by combining linearly the recorded signals. An on-line adaptive algorithm based on the Singular Value Decomposition (SVD) has been designed to compute the coefficients in these linear combinations. This algorithm will be implemented on a DSP board that can be plugged into the real-time recording system. The system will be very useful in studies of the fetal electrocardiogram during pregnancy, but also in all other studies such as fetal heart rate variability, fetal movements, etc., where a precise trigger of the electrical signal from the fetal heart is required.
Proton NMR longitudinal relaxation times (T1; 10.7 MHz; 37 degrees C) were measured in the kidneys and blood serum of mice inoculated with P388 leukemia, and/or treated with the chemotherapeutic drug cis-diamminedichloroplatinum(II) (cis-Pt). In parallel, serum total protein content, urea and creatinine levels were determined and protein fractions were separated electrophoretically. Serum T1 was found to be 1518 +/- 73 ms (1 SD) in control mice, 1670 +/- 69 ms in leukemic mice, and 1380 +/- 71 ms in the healthy and the leukemic cis-Pt treated mice. The T1 increase in leukemic serum and T1 decrease in the serum of cis-Pt injected mice are attributed to decreased and increased protein contents respectively. A detailed analysis in terms of electrophoretic fractions of serum proteins reveals that the serum relaxation rate 1/T1 is a multilinear function of the mass concentrations of the main serum protein fractions, explaining all serum T1 effects. This makes T1 a non-specific blood parameter. The kidney T1 was found to be 311 +/- 12 ms in normal mice and 334 +/- 20 ms in leukemic mice. A dramatic T1 increase is observed when the mice are injected with cis-Pt; the values are 400 +/- 38 ms and 407 +/- 39 ms for healthy and leukemic mice, respectively. This effect is related to the nephrotoxicity of the drug, as evidenced by serum urea and creatinine levels and protein content being higher than normal.