This paper describes a novel technique for the cancellation of the ventricu
lar activity for applications such as P-wave or atrial fibrillation detecti
on. The procedure was thoroughly tested and compared with a previously publ
ished method, using quantitative measures of performance. The novel approac
h estimates, by means of a dynamic time delay neural network (TDNN), a time
-varying, nonlinear transfer function between two ECG leads. Best results w
ere obtained using an Elman TDNN with nine input samples and 20 neurons, em
ploying a sigmoidal tangencial activation in the hidden layer and one linea
r neuron in the output stage. The method does not require a previous stage
of QRS detection. The technique was quantitatively evaluated using the MIT-
BIH arrhythmia database and compared with an adaptive cancellation scheme p
roposed in the literature. Results show the advantages of the proposed appr
oach, and its robustness during noisy episodes and QRS morphology variation
s.