Abstract :
Aim of the present study was to test our heart-rate adaptive match filter procedure for TWA analysis, which includes a) a heart-rate adaptive match filter, that filters out every ECG component but the TWA and quantifies TWA in terms of the duration (TWAD), amplitude (TWAA), and magnitude (TWAM, defined as TWAD times TWAA); and b) a statistical threshold criterion (STC) to identify, among all TWA detected cases, the ones (TWA+) most likely associated to an increased risk of sudden cardiac death. Our test, performed on the database assembled for the PhysioNet/Computers in Cardiology Challenge 2008, yielded the following results. On average, the data set was characterized by TWAD= 63plusmn31 beats, TWAA= 48plusmn77 muV, and TWAM= 3249plusmn5107 beats muV. Based on our STC, TWA+ cases were 21, yielding a ranking score of 0.684. Use of STC, in the perspective of discriminating risky cases, among all detected TWA cases, might have limited our score.
Keywords :
adaptive filters; electrocardiography; filtering theory; medical signal detection; medical signal processing; statistical analysis; ECG component; T-wave alternans; TWA amplitude; TWA analysis; TWA duration; TWA magnitude; electrocardiography; heart-rate adaptive match filter procedure; statistical threshold criterion; sudden cardiac death risk; Adaptive filters; Assembly; Band pass filters; Cardiology; Databases; Electrocardiography; Frequency; Low pass filters; Matched filters; Testing;
