| A new method of non-invasive blood flow measurement using Doppler ultrasound is presented and investigated. By applying pulsed Doppler ultrasound to make local measurements of blood velocity at constant depth in tissue, a velocity profile is estimated; spatial integration yields volume flow rate. This technique, known as C-mode velocity profiling, is plagued by flow estimation errors on the order of 20-60% due to partial volume effects. The hypothesis of this thesis is that application of attenuation compensation improves the accuracy and precision of C-mode velocity profiling by correcting for the partial volume effect. Systematic and random error for the new technique is investigated using a computational model and a steady state flow experiment. Both model and experiment confirmed the hypothesis, and determined that the remaining primary source of error is due to a correctable effect of signal processing. The mean error of the new flow measurement was {dollar}-{dollar}10 mL/min with a distribution that had a standard deviation of 18 mL/min. For physiologically relevant flow rates, this corresponds to estimation of blood flow to within 5%, with 95% confidence. |
