Comparison of the autoregressive modeling and fast Fourier transformation in demonstrating Doppler spectral waveform changes in the early phase of atherosclerosis

In this study, we have performed fast Fourier transformation (FFT) and autoregressive (AR) signal processing of the Doppler signals at a nonstenotic arterial site in patients with atherosclerosis and healthy volunteers. We have not only utilized Doppler sonograms, but also facilitated the power spectral density distribution graphs using AR modeling and FFT. Our preliminary analysis show that AR modeling has a higher efficacy in demonstrating Doppler spectral waveform changes in the preclinic or silent phase of atherosclerosis. AR has especially revealed an outstanding difference in the calculation for frequency level of maximum power spectral density.