Modifying the cosine model for nonstationary blood flow signal simulation.

Objective: The most important features of the Doppler signal are its nonstationary nature and randomness, which make simulating these kinds of signals difficult. Our intent was to create realistic simulations of these signals.

Methods: We generated the Doppler signal for the carotid artery by using the cosine model. We also modified the cosine model to simulate the femoral Doppler signal for the whole cardiac cycle.

Results: In auditory comparisons, the simulation results compared favorably with the original data. No appreciable differences were found between the real clinical signals and the simulations for the normal carotid and femoral arteries. However, the calculated normalized root mean square errors were high.

Conclusions: The cosine model is simple to implement and requires input of only the spectrum shape variation during a cardiac cycle. In addition, it is more informative than other modeling methods and therefore is preferred over other methods.