Extraction of ECGs for twin pregnancies

Jarvinen, Jarkko (2016) Extraction of ECGs for twin pregnancies. [USQ Project]

Text (Main Project)

Download (5MB) | Preview


The wellbeing of a fetus or fetuses can be monitored by the fetal heart rate (fHR). There are several proposed methods for fHR monitoring; these include fetal phonocardiography (fPCG), fetal cardiography (fCTG) and fetal magnetocardiogram (fMCG). Although, according to the research reviewed, none of these methods are ideal for monitoring or estimating fHR. The fPCG method is highly sensitive to noise and can only be used late in the pregnancy. With fCTG, the ultrasound transducer used for measuring the fHR needs to be properly aligned, otherwise the maternal heart rate (mHR) can be recorded instead of the fHR. In addition, the ultrasound high frequency exposure is not completely proven to be safe for the fetus. fMCG can detect fHR very accurately in comparison to the other methods but the method is unwieldy and expensive; thus not widely used in a clinical environment.

Therefore, there is a need for technology which would be able to provide more information about the cardiac health of a fetus, delivered in a cost-effective, streamlined manner. Based on the research reviewed and captured within this dissertation, non-invasive fetal electrocardiography (fECG) has been identified as a promising fetal cardiac monitoring method and if researched further, has the potential to become the next mainstream approach for monitoring fetal health. Within this dissertation, the fECG extraction methods have been explored and the findings captured. The research revealed that the fECG method can be used from early stages of pregnancy (20 weeks gestational age onwards). It is relatively low cost and does not necessarily require a highly skilled user. Continuous monitoring is also possible. The main challenge identified when using the non-invasive fECG extraction method is poor Signal-to-Noise Ratio (SNR) of the fECG signal on the abdominal signal which consists of fECG, maternal ECG (mECG) and noise.

Eleven different fECG extraction methods were tested as part of this dissertation. The extraction methods were based on Adaptive Methods (AM), Template Subtraction (TS)or Blind Source Separation (BSS). Synthetic test signals were used for the testing the methods. The test signals included five different noise levels across seven different single pregnancy physiological cases and one twin pregnancy case. Each recording included 34 channels (32 abdominal and two maternal reference channels).

For single pregnancy cases all of the extraction methods were able to extract the fECG from the test signals with varying degrees of success. Overall, the BSS-JADE method was the top performing method for single pregnancy cases getting a median F1 score of 99.85%. Furthermore, the twin pregnancy case was tested using BSS methods. The BSS FastICA algorithm using symmetric approach was the top performing method for the twin pregnancy case receiving a median F1 score of 99.93%.

Statistics for USQ ePrint 31426
Statistics for this ePrint Item
Item Type: USQ Project
Item Status: Live Archive
Additional Information: Bachelor of Engineering (Honours) Major Electrical & Electronic Engineering project
Faculty/School / Institute/Centre: Historic - Faculty of Health, Engineering and Sciences - School of Mechanical and Electrical Engineering (1 Jul 2013 - 31 Dec 2021)
Supervisors: Leis, John
Date Deposited: 20 Jul 2017 04:49
Last Modified: 20 Jul 2017 04:49
Uncontrolled Keywords: ECGs; twin pregnancies; fetal magnetocardiogram; electrocardiography; ultrasound transducer; signal-to-noise ratio; template subtraction
Fields of Research (2008): 09 Engineering > 0906 Electrical and Electronic Engineering > 090602 Control Systems, Robotics and Automation
Fields of Research (2020): 40 ENGINEERING > 4007 Control engineering, mechatronics and robotics > 400799 Control engineering, mechatronics and robotics not elsewhere classified
URI: https://sear.unisq.edu.au/id/eprint/31426

Actions (login required)

View Item Archive Repository Staff Only