Computational Neural Models for Body Surface ECG Data Analysis

Cardiovascular heart disease (CHD) is a chief publichealth priority worldwide. The 12-Lead Electrocardiogram (ECG) is astandard procedure in diagnosing CHDs such as MyocardialInfarction (MI). Nevertheless due to sparse spatial sampling it islimited in identifying cardiac abnormalities. Alternatively inBody Surface Cardiac Mapping (BSCM) a higher number of ECGs arerecorded. Hence BSCM provides a more comprehensive picture of electrocardiographic information than is possiblewith the 12-lead ECG. This work has two main objectives. Firstly todevelop a classification framework for an accurate and earlydiagnosis of acute MI. This decision support system encompassescomputational neural models with the input space based on BSCM.Secondly since MI is localised on the torso surface and due to thehigh number of electrocardiographic leads involved in BSCM it isdesirable to find an optimal reduced lead set for acute MI detection.By building an additional layer of knowledge between thecardiologist and clinical practice this work not only enhances final MIclassification performance but allow the discovery of newelectrocardiographic MI markers.