Positioning Electrodes On The Maternal Abdomen Biology Essay

Published:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

In this chapter, the signal sources, and the proposed protocol for data acquisition is introduced according to the fetal position and orientation. Two different bio-amplifiers are used for acquiring the data, and finally, evaluation and testing the algorithms for off-line simulation and real-time are described. Firstly, performance of the signal processing algorithm depends on its ability to detect the maternal QRS complex peak and in successfully extracting the fetal signal from the maternal abdominal signal, which will enhance the detection of the fetal QRS complex. Secondly, it depends on its ability to detect the MHR and as well as FHR.

The implementation requirements of an effective algorithm capable of real time fetal heart rate detection are:

Choosing suitable extraction algorithm, which distinguishes

oneself from other algorithms, in term of real time extraction.

Developing new effective enhancement technique for

attenuating the interference in the fetal signal.

Implementing peak detection algorithm capable of real-time

maternal as well as fetal heart rate detection.

In this chapter two new algorithms for FECG signal extraction and the detection of FHR as will as MHR are proposed. These algorithms have been developed for real time FECG extraction and detection from ECG signals recorded from the abdominal pregnant women.

PROBLEM FORMULATION

According to the discussion in Chapter II, Noninvasive FECG recording is performed by positioning electrodes on the maternal abdomen (Vullings et al. 2006). The invasive FECG recording involves inserting a direct contact electrode to the fetus head during childbirth that is using fetal scalp electrode to obtain the FECG signal. Invasive solution such as scalp electrode may injuries to the fetus, or uterus may be perforated leading to its infection. Hence a noninvasive solution that is using surface electrodes to aquired signals from pregnant women skin had been chosen to extract the FECG signal (FECGp) from an abdominal composite signal. Figure 3.1 show the sources of signals to be dealt with. The maternal ECG signal (MECGp) is the signal formed by the maternal heart which travels as electrical signal through tissue matter to the chest surface. At the same time this signal propagates inside the maternal body towards the fetal body, which is additively combined with the FECGp to form the signal.

Figure 3. 1 Simplified problem formulation

The primary signal which is of interest, FECGp, is formed by the fetal heart, with unwanted interference from MECGP. Note that additive noise, which may be due to the power line coupling, muscular and respiratory interference, thermal noise (from electronic equipments), and the electrode-skin interface, are not included in this formulation.

To formulate this problem with a noninvasive approach, the target is set on extracting FECGp as the optimal solution. A simple formulation to this problem is shown in Figure 3.1. With reference to this formulation, the mixed sources MECGp and FECGp can be written as follows:

(3.1)

(3.2)

An estimate of the FECGp can then be extracted from the abdominal composite signal using Equation 3.2. This estimation process will be discussed in the next chapter.

THE PROPOSED LEAD SYSTEM

The movement of the fetus depends on the weeks of gestation. Whereas the fetus can move relatively easy within the uterus at 24 weeks, near the end of gestation the fetus gets stuck within the uterus and cannot move easily. In obstetrics it is common use to describe the fetal position and orientation by means of fetal lie, presentation and position (Symonds 1992). The fetal lie is the orientation of the fetus, defined by the axis from head to bottom, in relation to the longitudinal axis of the uterus and describes whether the fetus is in a longitudinal, oblique or transverse lie.

The presentation of the fetus reflects the part of the fetus that is orientated towards the birth channel. Common presentations and the percentage in which they occur are indicated in the Figure 3.2 (Symonds 1992).

Figure 3. 2 Fetal position and orientation in the abdomen

Source: Symonds 1992

A subdivision is made for the fetus facing left or right side of the abdomen with a sub classification for posterior (towards the back), anterior (towards the front of the abdomen) or transverse (towards the side of the mother). Although there are comprehensive studies for lead system in the literature (Oosterom 1989), but for advanced digital signal processing algorithm, no particular lead system is specified. The lack of proper lead systems which can work properly for different age of pregnancy independent of fetal positions, is confirmed in the literature (Sameni et al. 2006).

The fetal position and orientation here are the bases which should be considered for designing new lead system. Furthermore the nearer the fetal heart is to the maternal skin, the more strength the FECG has in the measured AECG. In Figure 3.2 it can be seen that the fetal lies are nearly similar in vertex, face and brow implying that, the proposed lead system will consider these percentages in which they occur, while minimizing the number of abdominal leads attached to the mother.

In previouse work (Najafabadi 2008), ICA was used to extract FECG signal, from signals acquired by proposed lead system was proven to be suitable for acquiring signals from maternal abdominal skin. Based on that, the lead system in this research was developed in a way to semplify the distance between the electrodes, which does not change the effect of the electrode locations.

Beside that ANC is used to extract FECG signal with window signal, which reduces the number of the leads (3 leads: refrence, primary and comon) this is better than other lead system with ANC (5 leads: 3 refrences, primary and comon), which used to extract FECG signal (Zarzoso et al. 2001).

IMPLEMENTED PROTOCOL

The proposed protocol represents the configuration of the leads used for acquiring the signals from the pregnant women abdominal skin. This protocol is developed with respect to the above discussion and based on previous research protocol by Najafabadi (2008) as shown in Figure 3.3.

Figure 3. 3 Proposed lead system

Three abdominal ECG leads (P1, P3, P4), ground (G) and a common (P2) are used in the proposed lead system. The 'Ground' electrode can be placed anywhere as long as it is far from the other electrodes, hence it is placed on the right side from the centre back as shown in Figure 3.3. All the electrode positions of the proposed system are explained in Table 3.1. Three channels are acquired using three electrodes (P1, P3, P4), a common (P2) shared for all electrodes and ground (G).The p1 location is chosen for acquiring maternal signal, while the p3,p4 locations are chosen for acquiring AECG signals (maternal + fetal). The surface distance of the abdominal skin between the sternum and the navel is divided to three equal parts. In Table 3.1 the electrode positions of are explained according to that illustrated in Figure 3.3. The system setup for AECG recording is given in (Apendix A).

Table 3. 1 Electrode positions in proposed lead system

Electrode

Horizontal distance

Vertical distance

Remark

G

Attached to the right side

Shared for all leads

P1

Same vertical line with navel and sternum

sternum

Mainly

pickup MECG

P2

Same vertical line with navel and sternum

1/3 the distance between the sternum and the navel

Common for AECG1,3,4

P3

Same vertical line with navel and sternum

2/3 the distance between the sternum and the navel

AECG

P4

Same vertical line with navel and sternum

10 cm lower than navel

AECG

All the AECG signals acquired from volunteer pregnant women were recorded using the proposed lead system, where all chosen fetuses were singleton, between 36 and 39 weeks of gestation. Subject with twin pregnancies, anterior placed placenta, obese body mass index (BMI>30), gestational diabetes miletus (GDM) and hypertension were excluded from this research.

All the proposed leads from the bio-amplifier are prepared (ECG conductive adhesive electrodes with gel) and placed on the maternal abdominal skin for two to five minutes data acquisition of each case.

HARDWARE

In this study the g-tec bio-amplifier (g.tec Guger Technologies OEG) is used for recording signals from the abdominal skin of pregnant women in order to test the algorithm implemented under Matlab and simulink (Mathworks Inc.) in real-time.

Due to the absence of any electrical contact (Isolation Opto-coupler, patient isolation CF type Optical Signal Isolation of the g-tec bio-amplifier), the instrument used should be safe from electrical shock. As shown in Figure 3.4, all recorded signals for this research are acquired using 24 bit high resolution multi-channel bio-amplifier with multiple choices of sampling frequency. However, in this work the sampling frequency of 256 Hz was chosen. The measured AECG signals were processed in real time and saved in *.mat file format in personal computer. All locations of the electrodes on the maternal abdominal skin and the connections on the bio-amplifier are shown in Figure 3.4. The recorded signals (X1, X2 and X3) are acquired. In Figure 3.5 the front view of the g-tec. Bio-amplifier is shown in (a), while the rear view of the g-tec. is shown in (b).

Figure 3. 4 Locations and references for recording signals

.

Figure 3. 5 (a) front view and (b) rear view of g-tec Data Acquisition System

ABDOMINAL ECG DATABASE

In order to evaluate the performance of the proposed algorithms and lead system and also to investigate its ability in extraction and detection, two types of real recorded data are used: short-term recorded data using the BIOPAC system for off-line processing and off-line simulation processing, and short-term recorded data using the g-tec system for real-time processing.

Data Acquisition Using BIOPAC System

In this work Matlab 7.4 (Matlab script) is used to evaluate the performance of developed algorithm. AECG signals from pregnant women were recorded each of one minute duration, using a number of lead system protocols, and. sampled at 1000 Hz (Najafabadi 2008). For off-line processing the extraction algorithm and the detection algorithm are implemented using Matlab 7.4. Some of these signals were re-sampled at 256 Hz, in order to evaluate off-line the algorithms implemented under simulink.

The signals were recorded using the BIOPAC system. The measured AECG leads were saved in *.mat file format with resolution of 16 bits (resolution of BIOPAC). The recording time is one minute for each signal. FHR is also measured using Doppler ultrasound for comparing with those detected with the proposed algorithm.

Data Acquisition Using g-tec System

Once the implementation of the algorithms is completed, g-tec bio-amplifier is used to acquire other signals from volunteer pregnant women abdomen, each of two to five minutes duration. All recorded signals are acquired using 24 bit high resolution multi-channel bio-amplifier, and the chosen sampling frequency is 256 Hz.

EVALUATION METHOD

Data Set

The proposed algorithms were implemented using Matlab 7.4, and tested using the recorded data of the BIOPAC system. In order to test these algorithms in real time, they were also implemented under simulink (Mathworks Inc.). For this purpose, real recorded data were acquired from volunteer pregnant women in real-time using the g-tec bio-amplifier. The implemented algorithms are tested using signals as illustrated below, and evaluated using sensitivity and the positive predictivity according to section 3.7.2.

Using Matlab 7.4, thirty recorded signals each of five seconds duration were used for comparing between ANC and ICA according to Figure 3.6.

Using Matlab 7.4, twenty recorded signals each of one minute duration

were used for

detecting the MHR.

Under simulink, twenty recorded signals each of five seconds duration

were used for comparing between the RLS and NLMS implementation

of the ANC.

Under simulink, thirty recorded signals each of one minute duration were

used to test FHR detection under Matlab and simulink according to Figure

Pre-processed signals

RLS

ICA

Fetal Extracted signal

Post-processed fetal signal

Post-processed fetal signal

Fetal Extracted signal

Figure 3. 6 Block diagram of the extraction algorithms

Preprocessed

signals

RLS for FECG extraction

Postprocessed fetal signal

Fetal peak detection

Peak position correction

FHR

Figure 3. 7 Block diagram for FHR detection

Comparing to previouse work (Najafabadi 2008), 10 estimated FECGs by ICA, were also used to calculate the FHR of a portion of each signal in order to compare the result with estimated FHR by ultrasound M mode. The FHR average of these signals detected by the proposed algorithm is compared also with the average of estimated FHR by ultrasound M mode.

FHR was detected in real time using g-tec bio-amplifier in order to ascertain the implementation of the algorithm under simulink for real time FHR detection. A small number of recorded data obtained by the g-tec bio-amplifier was used to evaluate the performance of the proposed algorithm implemented.

The formentioned R peaks detections and extraction comparisons explained previously in 3.7.1 are further illustrated in Figures 3.8 and 3.9. The sensitivity and positive predictivity described in 3.7.2 are used to evaluate the performance of the extraction and detection algorithms illustrated in Figures 3.8 and 3.9.

Data recorded by BIOPAC system

Matlab 7.4 (Matlab script)

simulink (Mathworks Inc.)

MHR

Detection

Comparison between ANC and ICA

FHR

Detection

Comparison between RLS and NLMS

Comparison between RLS, ICA and ultrasound

Evaluation

Evaluation

Evaluation

Evaluation

Figure 3. 8 Biopac system data testing and evaluation

Data recorded by g-tec

Real time

FHR Detection

MHR Detection

Evaluation

Evaluation

Figure 3. 9 g-tec data testing and evaluation

Evaluation

All the stages of the proposed algorithms have been implemented in Matlab codes, these are evaluated using AECG signals acquired by the Biopac and g-tec system. The sensitivity and positive predictivity are used to evaluate the performance of the proposed algorithm (ANSI/AAMI EC57 1998). The Sensitivity () is the fraction of real events that are correctly detected and it is defined by:

(3.3)

The positive predictivity () is the fraction of detections that are real events and it is defined by:

(3.4)

Where (False Negatives) denotes the number of missed detections, (False Positives) represents the number of extra detections and (True Positives) is the number of correctly detected QRS complexes.

SUMMARY

The flow chart illustrated in Figure 3.10 represents the implementation of the proposed methodology.

The required data for the proposed methodology are acquired using tow bio-amplifiers, Biopac system is used to acquire data for off-line processing and for off-line simulation. The g-teq system is used to acquire data for real-time processing.

The proposed methodology is illustrated in Figure 3.10, where the proposed algorithms are implemented using three ways of Matlab programming. Firstly, using Matlab 7.4 for off-line processing, secondly, under Matlab®/Simulink® for off-line simulation and finally, under Matlab®/Simulink® for real-time processing. The extraction algorithm is applied in order to separate the fetal signal, and the detection algorithm is used to detect the maternal as well as the fetal peaks.

The proposed algorithms (extraction and detection algorithm) are tested using these recorded data. These tests are evaluated in term of peak detection, where the sensitivity and positive predictivity are used to evaluate the performance of these algorithms.

a- Clinical signals acquired by BIOPAC system

b- Clinical signals acquired by g-tec system

Implementatio of. algorithm under Matlab®/Simulink® for real-time processing

Implementatio of. algorithm under Matlab®/Simulink® for off-line simulation

Data Base of AECG

Implementatio of. algorithm using Matlab 7.4 for off-line processing

Data Base of AECG

Data Base of AECG

Evaluation

Evaluation

Evaluation

Proposed algorithms :

1-Extraction algorithm

2-Detection algorithm

Proposed algorithms :

1-Extraction algorithm

2-Detection algorithm

Proposed algorithms :

1-Extraction algorithm

2-Detection algorithm

Figure 3.10 Flow chart of the proposed methodology.

Writing Services

Essay Writing
Service

Find out how the very best essay writing service can help you accomplish more and achieve higher marks today.

Assignment Writing Service

From complicated assignments to tricky tasks, our experts can tackle virtually any question thrown at them.

Dissertation Writing Service

A dissertation (also known as a thesis or research project) is probably the most important piece of work for any student! From full dissertations to individual chapters, we’re on hand to support you.

Coursework Writing Service

Our expert qualified writers can help you get your coursework right first time, every time.

Dissertation Proposal Service

The first step to completing a dissertation is to create a proposal that talks about what you wish to do. Our experts can design suitable methodologies - perfect to help you get started with a dissertation.

Report Writing
Service

Reports for any audience. Perfectly structured, professionally written, and tailored to suit your exact requirements.

Essay Skeleton Answer Service

If you’re just looking for some help to get started on an essay, our outline service provides you with a perfect essay plan.

Marking & Proofreading Service

Not sure if your work is hitting the mark? Struggling to get feedback from your lecturer? Our premium marking service was created just for you - get the feedback you deserve now.

Exam Revision
Service

Exams can be one of the most stressful experiences you’ll ever have! Revision is key, and we’re here to help. With custom created revision notes and exam answers, you’ll never feel underprepared again.