Date of Award

3-17-2020

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Systems Engineering

First Advisor

Kamran Iqbal

Abstract

Early detection of fetal neurological disorders is of significant importance however, accessing the fetal brain signal has been a challenging issue. With highly sensitive Superconducting Quantum Interference Device (SQUID) sensors, fetal magnetoencephalography (fMEG) can be recorded and there have been many studies on understanding fetal brain activity through fMEG. The major problem about fMEG data is its low signal to noise ratio. fMEG is extracted from the mixture of maternal heart, fetal heart, fetal brain, and other biological signals from mother and the fetus. Effective removal of interferences is crucial in detecting the low amplitude fetal neurological signals. The objective of this research is to improve the detection of fetal evoked response (ER) by developing an algorithm for better attenuation of cardiac interference and verifying the source of the detected signal by magnetic dipole fitting. We have developed an improved frequency dependent subtraction algorithm to attenuate the cardiac interference regardless of the orientation of fetal heart and head. Normally, the algorithm is designed for removal of cardiac interference from fetal magnetoencephalography (fMEG) but in the case of overlapping fetal heart and brain signal sub-spaces, it has a low attenuation percentage. We have employed minimum norm projection operators (MNPO) to extract maternal and fetal heart signals (mMCG and fMCG) and we have subtracted them in the frequency domain from fMEG. We have shown that the cardiac attenuation has substantially improved with this method. After removal of interference, the detected fetal ER needs to be further verified by checking the source location. The common source verification is done by visually checking the magnetic field map, which is a highly subjective method and requires user intervention. We have established an objective verification method by using magnetic dipole fitting to localize the fetal head. This will enable a reliable detect

Download

Share

COinS