Analysis of sounds from occluded arteries by using a linear array of microphones

The development of a non-invasive method for measuring of disturbances and obstructions in flow patterns inside the arterial system is a valuable tool for medical diagnostic purposes. A complete mapping of the pressure field and its correlation to the flow field and to the characteristics of the occlusion (stenosis) are goals that have not yet been attained, and are of extreme importance to the medical, scientific, and bioengineering communities. Disturbed blood flow can generate arterial sounds that can be heard directly over the affected area and transmitted to another area over the skin. In recent years, several investigators suggested that audio-frequency analysis of vascular sounds may provide a non-invasive method that may be used to quantify arterial stenosis. A new high resolution array signal processing technique is proposed as a diagnostic tool to utilize the sounds that originate in occluded arteries, specifically determination of abnormal flow conditions. These sounds can be measured by non-invasive, sensitive and acoustic microphone linear array placed on the skin. The use of "Inverse Problem" technique proposed to localize and characterize abnormal sounds being generated in occluded arteries. The solution is obtained by deriving an accurate set of linear equations related to "input/output" signals obtained in array of microphones, and then stored in a "weight" matrix. A computer simulation for non-coherent sources located at different coordinates show a quite agreeable solution of a problem, by enhancing the spatial locations of the simulated sources. We built an experiment model that demonstrates different flow conditions in constricted tubes. Measuring sounds are picked by a linear array of 16 microphones placed separated from the constricted tube. In an experimental system we examine first the frequency content of constricted tube and then we derive a spatial resolution by using an inverse problem technique. Representative computer simulations and - easured results obtained in experimental model demonstrate the potential of our approach.