An open-source framework for synthetic post-dive Doppler ultrasound audio generation

Abstract

Doppler ultrasound (DU) measurements are used to detect and evaluate venous gas emboli (VGE) formed after decompression. Automated methodologies for assessing VGE presence using signal processing have been developed on varying real-world datasets of limited size and without ground truth values preventing objective evaluation. We develop and report a method to generate synthetic post-dive data using DU signals collected in both precordium and subclavian vein with varying degrees of bubbling matching field-standard grading metrics. This method is adaptable, modifiable, and reproducible, allowing for researchers to tune the produced dataset for their desired purpose. We provide the baseline Doppler recordings and code required to generate synthetic data for researchers to reproduce our work and improve upon it. We also provide a set of pre-made synthetic post-dive DU data spanning six scenarios representing the Spencer and Kisman-Masurel (KM) grading scales as well as precordial and subclavian DU recordings. By providing a method for synthetic post-dive DU data generation, we aim to improve and accelerate the development of signal processing techniques for VGE analysis in Doppler ultrasound.

Fig 1. Schematic of the experimental setup used for laboratory Doppler ultrasound bubble recordings.

A peristaltic pump was used to provide a water flow through a wall-less vessel situated in a gelatin phantom, where air bubbles were injected and measured with Doppler ultrasound above the phantom at different angles.

Fig 2. Flow diagram detailing the synthetic data generation process.

Fig 3. Graphical representation of KM grade (222) synthetic data generation process.

A) Original cardiac data with windows showing all cardiac cycles segmented by orange vertical lines. Based on the KM2 value of 2, 50% of the cardiac windows are selected for bubble-placement (6/12). B) For each selected window, a number of bubbles based on KM1 is chosen and placed into that cardiac window, bubble-audio is shown in orange. C) Summed cardiac and bubble-only audios to generate the final synthetic combined audio recording.

Fig 4

Examples of Spencer grade 4 for 3 conditions: 1) Subclavian data with VGE placed at any point along the signal. 2) Precordial data with VGE placed at any point in the signal, and 3) Precordial data with VGE placed only within the regions with amplitude less than 50% of the maximum amplitude between cardiac cycles.

Fig 5. Directory tree representation of directory system containing the generated synthetic data upon execution.