Chinese researchers have proposed a new neural network for screening heart disease quickly and accurately, according to the Chinese Academy of Sciences Thursday.
Atrial fibrillation (AF) is one of the most common sustained chronic heart diseases in the elderly, associated with high mortality, strokes and heart failure.
The early detection of AF is necessary for averting disability or death. However, AF detection remains uncertain due to its episodic pattern.
Currently, the AF screening method is to analyze ECG recordings in 1-2 minutes. Thus it is hard to use in wearable devices, which can monitor and capture AF recordings quickly and accurately.
The new neural network was proposed to screen out AF recordings from ECG recordings. The experimental results show that the proposed neural network achieved 96.99 percent of classification accuracy on ECG recordings of 5 seconds.
The best classification accuracy, 98.13 percent, was obtained on ECG recordings of 20 seconds.
The excellent AF screening performance of the new neural network can satisfy most of the elderly for daily monitoring on wearable devices. (Xinhua)
86-10-68597521 (day)
86-10-68597289 (night)
86-10-68511095 (day)
86-10-68512458 (night)
cas_en@cas.cn
52 Sanlihe Rd., Xicheng District,
Beijing, China (100864)
Copyright © 2002 - Chinese Academy of Sciences
