Signal integrity evaluation for fetal magnetocardiography

The fixed-point independent component analysis (FastICA) algorithm was widely used for extraction of the fetal magnetocardiography (fMCG) signals. Although effective, it may cause signal distortions if not properly designed. In this paper, we developed a simple yet effective method to evaluate the FastICA algorithm from the experimental and synthetic data. The method was based on the detection of simulating maternal MCG and fMCG signals, which derive from two individual electrocardiography (ECG) signals with adjustable frequency and amplitude. Three key performance indexes were proposed to evaluate the performances of the FastICA algorithm quantitatively. The evaluation results turned out to be very effective for parameters optimization of the algorithm and integrity evaluation of the fMCG signal.