Kai Noponen

Invariant trajectory classification of dynamical systems with a case study on ECG

An invariant pattern recognition framework for classification of phase space trajectories of nonlinear dynamical systems is presented. Using statistical shape theory, known external influences can be discriminated from true changes of the system. The external effects are modeled as a transformation group acting on the phase space, and variation of the trajectories not explained by the transformations is accounted for using principal component analysis. The approach suggested is highly adaptable to a wide range of situations and individual differences. The methodology presented is applied to detect abnormalities in electrocardiograms. Results based on measured data indicate that the model developed is resistant to the effects of respiration and body position changes, which are abundant in ambulatory conditions and cause significant morphological artifacts in the signal. The results also show that the detection of an artificially induced acute myocardial infarction is achieved with high performance. Due to its low computational complexity, the method developed can be implemented in real-time. The method developed also adapts to morphological changes caused by various heart conditions.

ECG-derived respiration methods: adapted ICA and PCA.

Respiration is an important signal in early diagnostics, prediction, and treatment of several diseases. Moreover, a growing trend toward ambulatory measurements outside laboratory environments encourages developing indirect measurement methods such as ECG derived respiration (EDR). Recently, decomposition techniques like principal component analysis (PCA), and its nonlinear version, kernel PCA (KPCA), have been used to derive a surrogate respiration signal from single-channel ECG. In this paper, we propose an adapted independent component analysis (AICA) algorithm to obtain EDR signal, and extend the normal linear PCA technique based on the best principal component (PC) selection (APCA, adapted PCA) to improve its performance further. We also demonstrate that the usage of smoothing spline resampling and bandpass-filtering improve the performance of all EDR methods. Compared with other recent EDR methods using correlation coefficient and magnitude squared coherence, the proposed AICA and APCA yield a statistically significant improvement with correlations 0.84, 0.82, 0.76 and coherences 0.90, 0.91, 0.85 between reference respiration and AICA, APCA and KPCA, respectively.

New Vectorcardiographic non-planarity measure of T-wave loop improves separation between healthy subjects and myocardial infarction patients

Principal component analysis of vectorcardio-graphic T-wave loop has been shown to be a potential tool to describe the abnormality of the cardiac repolarization and to predict cardiac events in patients with cardiac disease. In this paper a new method for estimating the non-planarity of the T-wave loop is introduced and tested with healthy subjects and subjects with anterior or inferior myocardial infarction. The method is based on the resamping of T-wave data points with respect to the arc-length, the total least squares plane fitting, the identifying and reordering of the fitted axes, and decomposing the optimal rotation matrix. A recently published related measure, PCA3, was used for comparison purposes. The results showed that the non-planarity of T-wave loop increased significantly in patients with myocardial infarction compared to the healthy group. The new method separated healthy and patient groups with p-value 0.002 while PCA3 only with p-value 0.075. The new method was superior to PCA3 in separating the healthy patients from both infarction types.

A Method for Robustly Determining the Relative Orientation of Vectorcardiographic Loop Structures

A robust method to determine the relative orientations of vectorcardiographic loop structures is presented.

Preliminaries to Finnish Word Prediction

Abstract Commercial word prediction software is thus far mainly available for uninflected languages such as English. In the present study, we investigate word prediction in highly inflected languages, using Finnish as an example. Despite its high degree of case inflection, about one third of word tokens in a Finnish text appear in their uninflected base form. As a result, simple prediction techniques such as word completion, originally developed for English, can be used for investigating characteristics of word prediction in inflected languages. Our preliminary results show that about 45% of characters can roughly be saved in Finnish word prediction in general for uninflected and inflected tokens. The most interesting result of our prediction experiments is, however, showing the distribution of character savings to the most common cases and their cumulative effect on the total percentage of character savings that may be achievable in Finnish word prediction. The major conclusions of the study are that word prediction in a highly inflected language such as Finnish is feasible provided that the case form used with a word appearing in a given context of use can be predicted correctly, at least in some cases, and the cognitive load of the resulting prediction system for the user is not too high when the prediction of the case form fails.

Spectral Data Fusion for Robust ECG-derived Respiration with Experiments in Different Physical Activity Levels.

In this paper, we study instant respiratory frequency extraction using single-channel electrocardiography (ECG) during mobile conditions such as high intensity exercise or household activities. Although there are a variety of ECG-derived respiration (EDR) methods available in the literature, their performance during such activities is not very well-studied. We propose a technique to boost the robustness and reliability of widely used and computationally efficient EDR methods, aiming to qualify them for ambulatory and daily monitoring. We fuse two independent sources of respiratory information available in ECG signal, including respiratory sinus arrhythmia (RSA) and morphological change of ECG time series, to enhance the accuracy and reliability of instant breathing rate estimation during ambulatory measurements. Our experimental results show that the fusion method outperforms individual methods in four different protocols, including household and sport activities.

Automatic Parametrization of Somatosensory Evoked Potentials With Chirp Modeling

In this paper, an approach using polynomial phase chirp signals to model somatosensory evoked potentials (SEPs) is proposed. SEP waveforms are assumed as impulses undergoing group velocity dispersion while propagating along a multipath neural connection. Mathematical analysis of pulse dispersion resulting in chirp signals is performed. An automatic parameterization of SEPs is proposed using chirp models. A Particle Swarm Optimization algorithm is used to optimize the model parameters. Features describing the latencies and amplitudes of SEPs are automatically derived. A rat model is then used to evaluate the automatic parameterization of SEPs in two experimental cases, i.e., anesthesia level and spinal cord injury (SCI). Experimental results show that chirp-based model parameters and the derived SEP features are significant in describing both anesthesia level and SCI changes. The proposed automatic optimization based approach for extracting chirp parameters offers potential for detailed SEP analysis in future studies. The method implementation in Matlab technical computing language is provided online.

Printed, skin-mounted hybrid system for ECG measurements

In this paper we report a design and fabrication process for a screen printed, skin-mounted hybrid system for electrocardiogram (ECG) measurements. The system consists of printed electrodes on a stretchable bandage substrate designed to be attached to the chest, an electronics module, and a data receiving device. The electronics unit is reversibly attached to the single-use electrode bandage to measure the ECG data. The ECG data is then transmitted to a mobile device via Bluetooth Low Energy and the mobile device then displays the data graphically and sends it further a cloud for storing and further analysis. The attained quality of the measured ECG signals is fully satisfactory to compute important cardiac parameters and after preprocessing the signal could be used for more profound analysis of ECG wave shapes.

Analysing performance in a word prediction system with multiple prediction methods

In this article, we investigate the performance of a hybrid prediction system with a phrase prediction utility in English word prediction from two viewpoints. From the application users point of view, measures of effort savings are important in word prediction. Global performance measures such as the average percentage of keystroke or character savings, however, hide rather than display the details of the functioning of the prediction system as a whole. In the present study, we analysed in detail the performance of a prediction system with a phrase prediction utility along with single word prediction. Our preliminary results with a corpus of 383 lexical bundles show that, from a technological viewpoint, the following three parameters affect the practical utility of the phrase prediction method in a hybrid prediction system: (1) cost of selecting an appropriate prediction mode for single word prediction and phrase prediction; (2) token frequency of phrases in the text predicted, and (3) coverage of the phrasal lexicon. We found that all three affect the phrase prediction performance in different proportions. When the percent of ambiguous search keys finding both phrases and single words is 20%, phrase frequency 35%, and coverage of the phrasal lexicon 98%, the character savings percentage for the whole text will be improved by 6% points under optimal conditions. The system is practically useful as long as an appropriate prediction mode can be selected automatically or the cost of disambiguation of a prediction mode is not too high.

Optimal Short Distance Electrode Locations for Impedance Pneumography Measurement from the Frontal Thoracic Area

Electrical impedance pneumography signal is a valuable tool in qualifying better the person’s health condition. It can be used in monitoring of respiration rate, rhythm and tidal volume. Impedance pneumography has also the potential in ambulatory physiological monitoring systems that are increasingly often implemented using plaster-like on-body devices. In such cases, the area of electrode substrate may be limited and therefore, the electrode configuration, which is able to provide both a clinically valuable electrocardiogram signal and accurate pulmonary information, is an issue. EAS is a useful small area electrode configuration that can be used for electrocardiogram measurements. In this work, different two-electrode bipolar pairs of EAS system are tested for impedance pneumography measurements. Two additional electrodes are also considered in these tests. Our results show that the electrode pair S-A provides the most accurate respiration cycle length and is least affected by movement artifact. Additionally, the results show that this electrode pair produces the signals with highest amplitude.