CAD system for impedance cardiography signal analysis Maria Rizzi1,*, Matteo D’Aloia2, and Cataldo Guaragnella1 1

Dipartimento di Ingegneria Elettrica e dell’Informazione, Politecnico di Bari, Italy {maria.rizzi, cataldo.guaragnella}@poliba.it 2 MASVIS s.r.l., Conversano (BA), Italy [email protected] Abstract. Impedance cardiography is a non-invasive method adopted to measure cardiac functions and to monitor hemodynamic parameters. It evaluates stroke volume and some other indices related to cardiac activity, by sensing variations in electrical impedance of thorax produced by changes in blood volume during cardiac cycle. The low signal to noise ratio of impedance cardiogram makes the accurate detection of signal relevant points hard and complicated. In this paper, a computer aided detection system able to provide an easy implementation in design tools is developed. Adopting a parallel computing architecture based on singularity detector, the conceived method analyzed the time-derivative of thoracic impedance signal for characteristic point detection and localization. Experimental results show the approach effectiveness, the method high accuracy and the system noise tolerant capability Keywords: ICG Wavelet transform Peak detector Impedance cardiography Bioimpedance Parallel computing

1

Introduction

The widespread increase in cardiovascular diseases and the technological improvements in diagnostics have contributed to develop of cheap and non-invasive methods for hemodynamic parameter measurements. Some non-invasive methods are not suitable for a continuous monitoring of patient state. Impedance cardiography (ICG) is a technique adopted for diagnosing several heart troubles such as atrial and ventricular dysfunctions, valve disorders, aortic stenosis and vascular diseases. The characteristic to be noninvasive and suitable for long-term and continuous monitoring have caused its introduction in medical environment [1]. In presence of patients which require a continuous monitoring, the introduction in clinical routine of information technologies such as computer aided systems for cardiac signal analysis can [2, 3]:  help physicians in detecting several pathologies;

Page 1 of 5

 improve the patient quality of life and life expectancy;  reduce costs of healthcare services. Many efforts have been done to implement automatic detection of reference points in biological signal. However, existing peak detection algorithms are difficult to automate for generic use because either they rely on a number of parameters that need to be customized for a particular application of the algorithm or they use reference information that is highly specialized for a particular application. Most of the proposed methods make use of filtering technique [4] or adaptive thresholding technique [5, 6]. Both techniques exhibit limitations when real signal are adopted [7]. In fact, the first drawback of filtering-based approach is that frequency variations in the signal under test may adversely affect the method performance. The second problem in filter based algorithms is frequency band overlapping of noise and some biological signals. In this paper a Computer Aided Detection (CAD) system is described. It detects and localizes C points adopting a multi-resolution analysis. The implemented method optimizes the computational time as it processes the ICG signal with a parallel procedure.

2

ICG technique

Impedance cardiography is the study of cardiac function by means of thorax electrical impedance measurements. High frequency (20-100KHz), low intensity current (1-5mA rms) is injected through the thorax by some electrodes and the impedance change is sensed by measuring a voltage across other electrodes. No risk of physiological effects have been found because various tissues of human body are not excitable at this frequency and at this low current level (Patterson, 1989). The impedance variation can be used for diagnostic information and for the stroke volume (SV) estimation by using blood flow appropriate model. Fig. 1 shows a typical impedance waveform obtained from electrodes in which the characteristic points are indicated. Pulsating blood flow through the thoracic aorta causes shifts in the thoracic impedance as a function of changes in blood volume. This oscillating component of the total thoracic impedance can be expressed as its derivative (dZ/dt). Measurements of the changes in the thoracic impedance (dZ/dt waveform) during the cardiac cycle are used to calculate SV. Equations used for SV evaluation, take into account position and value of C-point related to B-point and X-point [8, 9].

Fig. 1. Typical impedance waveform from the thorax of a human subject

3

Implemented method

The method analyses the first derivate of the impedance signal in real time without any pre-filtering procedure, showing an high noise immunity degree. ICG signal is sampled at a frequency of 250 Hz. The input hardware stores sequentially all the samples in a high speed frame which is then processed in real time by the system. For C point detection and localization, the method decomposes the ICG signal into six dyadic scales so to reduce noise sensitivity significantly (Fig.2). The bior3.3 wavelet has been chosen as it makes the perfect signal reconstruction possible

Fig. 2 Decomposition of ICG signal over six scales

According to the power spectra of ICG signal, it is evident that the larger contribute of the true signal is located in scales 4, 5 and 6, while scales 1, 2 and 3 are mostly affected by noise. Adopting a soft treesholding technique for level 1, 2, 3, noise and artefacts have been reduced. Then the signal has been reconstructed in the time domain. Usually, wavelet decomposition algorithms make use of filters in a tree structure. This is unsuitable for implementation by design tools. To overcome these limits, equivalent parallel filter banks are used. However, output signal realignment is necessary to equalize the delay introduced by each filter.

Page 3 of 5

At this step the method localizes R points: a point of maximum value is present in component signals (scales 4-5-6) in the same locations of each singularity in ICG signal. The parallel behaviour of the procedure makes the contemporary search of singularity points in each scale, possible. Making use of a parallel procedure, the proposed method searches local maximum points in the positive region of scale 4, scale 5 and scale 6. Singularities are selected adopting a threshold technique. Various tests have indicated the local maximum in the lower scale as the best point for the real signal peak localization.

4

Performance evaluation

The evaluation of the proposed detection methodology is carried out using real recorded data. Moreover, tests have been repeated adding Gaussian noise with zero average and variable variance. In this situation the method noise immunity has been evaluated. Two parameters are used for the performance evaluation: Sensitivity: Se = TP/(TP + FN) (1) Positive Prediction P = TP/(TP + FP) (2) where:  TP = number of true positive detections;  FN = number of C points present in ICG that the algorithm is not able to detect;  FP = number of C points detected by the algorithm but really not present in ICG Tested frame presents C-peak value fluctuations in the range [1÷1.5Ω/s]. Other local maximum points are all in the negative region. The obtained Se is very satisfactory and appears quite independent from noise (Fig.3). P is fairly good but decreases as noise increases (Fig.4). Anyway, it is to be noted that very heavy noise conditions have been chosen to test the algorithm noise immunity. An additional Gaussian noise signal with v=0.1( /s)2 corrupts heavily the ICG signal; in particular the noise, besides changing the ICG signal shape, introduces many false peaks while cancels a minor number of true peaks.

Fig. 3 Sensitivity

Fig. 4 Positive Prediction

5

Conclusions

In this paper, a real-time procedure for ICG analysis is presented and validated. The adopted method combines threshold technique and wavelet transform and optimizes the computational time as it processes the ICG signal with a parallel procedure. Experimental results show the method high sensitivity parameter. In fact, sensitivity reliable results with minimum interferences from noise and artifact have been obtained.

References [1] Jensen, L., Yakimets, J., Teo, K.K.: A review of impedance cardiography. Heart & Lung 24, 183193 (1995). [2] Rizzi, M., D'Aloia, M., Castagnolo, B.: High sensitivity and noise immune method to detect impedance cardiography characteristic points using wavelet transform. J. App. Sc. 9, 1412-1421 (2009). [3] Rizzi, M., D'Aloia, M., Castagnolo, B.: A new method for ICG characteristic point detection. In: 1th Int. Conf. on Bio-inspired Systems and Signal Processing pp.244-249 (2008) [4] Leski, J., Tkacz, E.: A new parallel concept for QRS complex detector. In: IEEE 14th Annual Int. Conf. on Eng. in medicine and Biology Society, pp.555-556 (1992) [5] Pan, J., Tompkins, W.J.: A real-time QRS detection algorithm. IEEE Trans. Biomed. Eng. 22, 289-297 (1984) [6] Sun, Y., Suppappola, S., Wrublewski, T.A. Microcontroller-based real-time QRS detection. Biomed. Instr. Technol. 26, 313-327 (1992). [7] Sun, Y., Chan, K.L, Krishnan, S.M.: Characteristic wave detection in ECG signal using morphological transform. BMC Cardiovascular Disorders, 5 (2005) [8] Sramek, B. B.: Cardiac output by electrical impedance. Medical Electronics 4, 93-97 (1982) [9] Bernstein, D.P.: A new stroke volume equation for thoracic electrical bioimpedance: theory and rationale. Critical Care Medicine 14, 904-909 (1982)

Page 5 of 5