Characteristics of innocent and stenotic cervical bruits

Abstract

The purpose of this study was to identify the stenotic and innocent bruit and to recognize possible identifying features of each. Treating the acoustical signal as random stationary noise and examining the histogram for differences in variation, the first moment was found to be unsatisfactory for differentiating between innocent and stenotic bruits. It was found that the two families of bruits could be reliably separated (with up to 85 per cent accuracy) by counting the number of major peaks in a periodogram (a form of the energy density spectrum) of the bruit. From the results of a t test, rejecting the null hypothesis at the 0.01 level, significant differences were also found between the following: the mean zero-crossing frequency for stenotic bruits (273.5 Hz) and for innocent bruits (175.3 Hz); the average energy-spectrum mean frequency for stenotic bruits (130.7 Hz) and for innocent bruits (82.0 Hz); the mean of the energy-spectrum 90 per cent bandwidth for stenotic bruits (188.4 Hz) and for innocent bruits (123.3 Hz). It is then possible to construct a reliable method of diagnosis for the cervical bruit based on the energy density spectrum. Employment of additional analysis could considerably increase this reliability. Such a method of diagnosis could employ recordings make by individual physicians and analyzed in a central location and/or a central facility to which patients would be sent. It is also possible to construct an electronic system which would have as its purpose the analysis of, or diagnosis of, bruits. For example, instruments ★ ★ Saicor Real Time Spectrum Analyzer, Signal Analysis Industries Corp., 12 Di Tomas Court, Copiague, N. Y. 11726. are commercially available which will produce the energy-density spectrum directly from the analogue signal from a contact microphone. Using one of these instruments and the appropriate associated electronics, it would be possible to construct a system which could be used for the diagnosis of bruits either from a recording of the sounds or from a direct microphone input of the sounds from the patient. No method will be 100 per cent reliable, but the complementary methods of signal analysis, and the experience of the physician should lead to increased reliability in the diagnosis of cervical bruits.