Ventricular fibrillation is a cardiac arrhythmia that can result in sudden
death. Understanding and treatment of this disorder would be improved if pa
tterns of electrical activation could be accurately identified and studied
during fibrillation. A feedforward artificial neural network using backprop
agation was trained with the Rule-Based Method and the Current Source Densi
ty Method to identify cardiac tissue activation during fibrillation. Anothe
r feedforward artificial neural network that used backpropagation was train
ed with data preprocessed by those methods and the Transmembrane Current Me
thod. Staged training. a new method that uses different sets of training ex
amples in different stages, was used to improve the ability of the artifici
al neural networks to detect activation. Both artificial neural networks we
re able to correctly classify more than 92% of new test examples. The perfo
rmance of both artificial neural networks improved when staged training was
used. Thus, artificial neural networks may be useful for identifying activ
ation during ventricular fibrillation. (C) 2000 Academic Press.