Catálogo de publicaciones - libros

The paper gives a brief explanation of a new International project EMITEL and its associated multilingual e-Dictionary. The project is developing the first web-based e-Encyclopedia in the profession. EMITEL will address the lifelong learning of a wide range of specialists and will be available free on Internet. The project advanced work-inprogress - the e-Dictionary is already functioning at www.emitdictionary.co.uk

We suggest in this review paper that control of assistive systems for individuals with disability caused by injury or disease of central nervous system has to be approached with rather sophisticated methods that are capable to deal with high redundancy, nonlinearities, time variations, adaptation to the environment, and perturbations. The use of three levels that provide interaction with user, coordination of multi joint activity, and biological actuators is likely to be the solution for future electrical stimulation assistive systems. This is especially important for therapeutic assistive systems that must mimic life-like movement. The top control level needs to be discrete and secure the recognition of intended movement and possibly some kind of feedback, the middle control level needs to be discrete and provide multi joint coordination that is based on temporal and spatial synergistic model of the movement. The lowest control level needs to be model-based in order to match the specifics of the musculo-skeletal system. The hierarchical hybrid control is inherently predictive adaptive controller that, if properly designed, could results with effective generation of segment movements that lead to life like function (e.g., walking, standing, manipulation, grasping, etc.).

The translation “from bench to bedside” of a scientific discovery, proof of principle or simple idea, that originated within academia, into a successful industrial product is a complex, long and costly process. Many factors need to be accounted for and careful planning and protection of intellectual property are essential to retain the value of the idea. This analysis, based on more than twenty-five years of experience in the biomedical field at IGEA, is presented to outline the different steps involved in such an endeavour. Critical factors defining the different phases, from initial evaluation of the idea to marketing and post-marketing monitoring, are described focusing on development processes.

In this paper we investigate the influence of vagal blockage on heart rate variability complexity measures. Nine conscious rats are injected with methyl-scopolamine brobide (50 μg/kg s.c.). We analyze 10 minute segments of beat-to-beat intervals before and after injection by standard time and frequency domain methods, compression entropy, sample entropy, Poincaré plot, detrended fluctuation analysis and symbolic dynamics. All parameter domains show changes in heart rate variability after vagal blockade, indicating a decrease in heart rate complexity. In conclusion, vagal modulation plays an important role in the generation of heart complexity in rats or, in other words, heart rate complexity measures are sensitive to vagal heart rate modulation.

The aim of this study is to compare different methods for the classification type problems specifically in predicting Atrial Fibrillation (AF) after Coronary Artery Bypass Grafting (CABG). The prediction/classification model tends to predict a categorical dependent variable (which determines the belonging of patient to a group of patients that have or to a group of patients that have not developed AF), by one or more continuous and/or categorical predictor variables derived from the patients' history, electrocardiograms and in particular from the P wave. We have obtained the parameters from continuously recorded ECG after the surgery.

In this paper, we present a simulation study of the effects of Sotalol, a known anti-arrhythmic drug, on the rapid delayed rectifier potassium current (). The current is encoded by the (HERG), which plays a major role in repolarization in mammalian ventricles. HERG is also the target of class III anti-arrhythmic drugs, such as Sotalol. Due to its unique structure and electrophysiological qualities, non-cardiac drugs readily bind with residues inside HERG's intracellular cavity. A novel Markov model was developed to model Sotalol's interaction with HERG. The model was validated using experimental data from HERG expressed in Human Embryonic Kidney (HEK) cells and integrated into the ten Tusscher (2006) human ventricular cell model. The simulation results show that an increase in Sotalol concentration decreases the overall conductance of over time, resulting in prolongation of the action potential duration. This effect is larger in mid-myocardial than in endocardial and epicardial cells. Therefore, Sotalol-induced effects on cardiac repolarization may result in enhanced transmural dispersion of repolarization in the ventricles, and also in changes in the T wave.

We present a Field Programmable Gate Arraybased system for single-lead electrocardiogram signal processing which performs beat detection and classification to normal and ventricular beats. Geometrical properties of a phase-space portrait of an ECG signal are used for QRS complex detection, while classification is done with a modified classification algorithm that is a part of the Open Source ECG Analysis Software. The chosen Field Programmable Gate Array has an embedded PowerPC processor and is very suitable for mixed hardware and software designs. Beat detection is implemented in hardware and the classification is executed on the embedded PowerPC 405 core. The algorithm was developed on the MITBIH Arrhythmia Database resampled to 250 samples per second. Sensitivity of 99.80% and positive predictivity of 99.84% was achieved for QRS complex detection and sensitivity of 92.59% and positive predictivity of 95.55% was achieved for identification of premature ventricular complexes. A comparison of processing speed between a personal computer and the embedded system shows that a personal computer running at 18-times faster clock speed processes data only six times faster.

Proposed methods for feature extraction and selection stages of biomedical pattern recognition system are presented. Time-Frequency signal analysis based on adaptive wavelet transform and Principle Component Algorithm (PCA) algorithm is used for extracting and selecting from original data the input features that are most predictive for a given outcome. From the discrete fast wavelet transform coefficients optimal feature set based on energy and entropy of wavelet components is created. Then PCA is used to shrink this feature group by creating the most representative parameter subset for given problem, which is the input for last neural classifier stage. System was positively verified on the set of clinically classified ECG signals for control and atrial fibrillation (AF) disease patients taken from MITBIH data base. The measures of specificity and sensitivity computed for the set of 20 AF and 20 patients from control group divided into learning and verifying subsets were used to evaluate presented pattern recognition structure. Different types of wavelet basic function for feature extraction stage as well as supervised (Multilayer Perceptron) and unsupervised (Self Organizating Maps) neural network classification units were tested to find the best system structure.