Noise sensitivity of three surface ECG fibrillation detection algorithms

The widening application of automatic external defibrillators (AEDs) presen ts very strong requirements for external electrocardiogram (ECG) signal ana lysis. Highly accurate detection of shockable rhythms is required, aimed to approach the maximum of 100% sensitivity and specificity. In a previous study the performance of five well known detection algorithms was assessed by test signals from the ECG-signal databases of the American Heart Association (AHA) and the Massachusetts Institute of Technology (MIT ). The results obtained were used as a basis for testing the noise sensitiv ity of three of these algorithms. Realistic noise was obtained by simulatio n and recording of signal disturbance by various motions during resuscitati on and defibrillation episodes (body shudder convulsions and gasps, cable m ovement, car transportation). The sensitivity and specificity of the detect ion algorithms were evaluated using electrocardiogram signals mixed with th ese noises.