Separating independent components in heart period signal / 生物医学工程学杂志

ABSTRACT

To extract sub-signal of heart period signal (HPS), a new statistical signal processing approach, namely independent component analysis (ICA) was addressed. Electrocardiosignal (ECS) was acquired from ten volunteers. ECS was sampled 8 minutes when the volunteer was in supine position, and then when the same volunteer was in erect position. HPS was extracted from ECS. According to time-delay, HPS was divided into five groups as mixed signals. Five signals were reconstructed into two groups by ICA. The rebuilt signals were transformed by Fourier transformation. One centralized in low frequency (called IC1); the other did in high frequency (called IC2). The power of IC1 was significantly increased (P<0.01) while that of IC2 showed no significant change (P>0.05), and the ratio of IC1 to total power also significantly increased with the change from supine position to erect position. Comparsion between the two postural results reveals that IC1 may express sympathetic activity, and IC2 represents parasympathetic activity. Sympathetic and parasympathetic nervous functions can be evaluated respectively and quantitatively by use of data and graphs from the two decomposed components.