Time series and Mel-frequency cepstrum coefficient analyses of venous pulsatile tinnitus vascular sound and flow velocity sensed by transcranial/retroauricular doppler ultrasound approaches

Hsieh, Yue Lin, Hsieh, Yue Da and Wang, Wuqing (2022) Time series and Mel-frequency cepstrum coefficient analyses of venous pulsatile tinnitus vascular sound and flow velocity sensed by transcranial/retroauricular doppler ultrasound approaches. Sensors and Materials, 34 (7). pp. 2791-2807. ISSN 0914-4935

Text (SM3006) - Published Version Available under License Creative Commons Attribution. Download (5MB)

Abstract

Venous pulsatile tinnitus (PT) arises from the motion of blood flow. However, the correlation between flow velocity and amplitude remains undiscovered. In this study, retroauricular colorcoded Doppler (RCCD) and transcranial color-coded Doppler (TCCD) ultrasound examination techniques were deployed to assess the hemoacoustics at the ipsilateral internal jugular vein (IJV), intradiverticular, mainstream sinus, and transverse sinus regions. Auto- and crosscorrelation analyses were used to analyze the correlations between flow velocity and amplitude. Furthermore, the Mel-frequency cepstrum coefficients were calculated, and the Melspectrogram was used to exhibit the human perception of PT. The mainstream sinus flow had the highest coefficient (cross-correlation coefficient = 0.781) among the sensed locations. The cross-correlation coefficient of the IJV was the second largest and close to that of the mainstream sinus flow. The transverse sinus flow had the lowest cross-correlation coefficient. Additionally, the transverse sinus septum was visualized for the first time using the RCCD technique in this study. In conclusion, cross-correlation analysis indicates that the amplitude of vascular sound is highly correlated to the vascular flow velocity. The Mel-spectrogram demonstrates the outcome of the human perception of PT, and its use can be extended to future psychoacoustic studies.

Item Type:	Article
Additional Information:	© 2022 The Author(s).
Divisions:	LSE
Date Deposited:	24 Aug 2022 16:33
Last Modified:	12 Dec 2024 03:14
URI:	http://eprints.lse.ac.uk/id/eprint/116383

Actions (login required)

View Item