Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.603537
Title: Theory and applications of linear lead transformations in computerised electrocardiology
Author: Güldenring, Daniel
Awarding Body: University of Ulster
Current Institution: Ulster University
Date of Award: 2013
Availability of Full Text:
Access through EThOS:
Abstract:
An electrocardiogram (ECG) can be used for the identification of different heart diseases. Linear lead transformations (LLTs) allow for the estimation of unrecorded ECG leads through linear combinations of a number of recorded ECG leads. This estimation relies upon statistical inter-lead correlations. Reduced lead systems (RLSs) commonly utilise LLTs for the estimation of the full 12-lead ECG from a reduced number of recorded ECG leads. The RLS used within this thesis utilised the recorded basis leads I, II, V2 and V5 for the estimation of the unrecorded target leads VI, V3, V 4 and V6. An assessment of whether the utilisation of statistical short-term autocorrelations, in addition to the commonly used inter-lead correlations, can improve the estimation performance (EP) of the above RLS has been conducted. The utilisation of statistical short-term autocorrelations was found not to improve the EP during episodes of acute myocardial ischemia (AMI). The effect of AMI on the EP of the aforementioned RLS has been investigated. It was found that AMI can reduce the EP. The inability of LLTs to fully describe the electrical volume conductor properties of the human trunk was found to be the likely source for the reduction in the EP. The EP achieved by the above RLS, has been assessed during episodes of AMI. It was found that a similar EP for target leads VI and V6 was achieved regardless of whether patient-specific or generalised transformation weights were employed. An LLT that allows for the estimation of the Frank vectorcardiogram (yCG) from the Mason-Likar (ML) 12-1ead ECG was developed. The developed LL T and the Kors matrix were used for the estimation of the Frank VCG from ML 12-lead ECG data. The developed LLT was found to allow the determination of the VCG parameters 'spatial ventricular gradient' and 'spatial QRS-T angle' with higher accuracy than when using the Kors matrix.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.603537  DOI: Not available
Share: