Assessment Of Variation Of Electrocardiogram Multi-lead T-wave Based On Functional Data Analysis

Electrocardiogram (ECG) is an indicator of cardiac activity level, which is of great value for research of cardiac physiology and pathology. ECG is widely used to analyze and identify sinus arrhythmia, assess damage advancement in cardiac muscles and investigate functionality of cardiac atrium and ventricles. During clinical trials ECG is performed as a primary measure to evaluate effect of drugs on cardiac functions. The main parameter in the ECG that reflects cardiac activity is the QT interval, which represents the temporal distance between the start of the Q wave and the end of the T wave. The quantification of QT interval has so far been based on the manual recognition of ECG features by cardiologists. Naturally this quantification method is prone to involve individual biases in measurement, mainly from detection of the end of a T wave. The examinations using QT interval are thus considered to be highly variable, inefficient and costly. In this work we constructed models of T wave using functional data analysis. Based on the assessment on models of three limb leads and three chest leads out of the total of twelve ECG leads, we analyzed and compared the degree of T wave variation in ECG leads from subjects taking Placebo or Moxifloxacin. Investigation on a large amount of ECG data indicated that the application of Moxifloxacin elevated the degree of T wave variation in ECG leads.