Ultrasonic detection of cardiovascular flow disturbances

Blood flow that is disturbed or turbulent may have a significant effect on the development of cardiovascular disease. A method is presented here for detecting periods of disturbed flow using autocorrelograms of the audio signal from a pulsed ultrasound Doppler velocity meter (PUDVM). Autocorrelograms describe quantitatively how the form of a signal changes over time. We produced steady laminar and turbulent pipe flow in a hydraulic test tank, and computed autocorrelograms of the audio signal of the centerline velocity as detected by the PUDVM using fast Fourier transform techniques. We have shown that the autocorrelation coefficient averaged over a short length of time (64 ms) is significantly higher for laminar than for turbulent flow. We have also produced pulsatile flow in our hydraulic tank and computed the mean autocorrelation coefficient at different phases of the flow cycle. The regions of disturbed and undisturbed flow were predicted from the steady flow results. The disturbed flow first appears during the period of the highest forward velocities. These results indicate that the mean autocorrelation coefficient can serve as an indicator of the presence of flow disturbances.