On the validity of the linear approximation in the parametric measurement of attenuation in tissues.

A known method for the determination of the attenuation coefficient of tissues involves the measurement of the center frequency downshift of Gaussian spectra. The tissue is generally assumed to have linear frequency dependent attenuation. This assumption results in a simple relationship between the spectral downshift and the attenuation coefficient. However, recent studies have shown that many tissues exhibit nonlinear frequency dependent attenuation. In this paper we investigate the consequences of applying the linear assumption to non-linear cases. These studies show that even a small deviation from linearity results in significant errors in the determination of the attenuation parameters.