Time-domain separation of interfering waves in cancellous bone using bandlimited deconvolution: simulation and...

Area of Science:

• Biomedical Engineering
• Materials Science
• Ultrasound Physics

Background:

• Cancellous bone analysis uses through-transmission ultrasound, generating fast and slow waves.
• Assessing bone properties from these waves is challenging due to signal overlap.
• Wave properties are crucial for understanding bone's material and micro-architectural characteristics.

Purpose of the Study:

• To evaluate two signal decomposition methods for separating overlapping fast and slow waves in cancellous bone ultrasound.
• To assess the accuracy and precision of bandlimited deconvolution and MLSP+CF for wave property estimation.
• To determine if these methods are suitable for cancellous bone analysis.

Main Methods:

• Applied bandlimited deconvolution to decompose ultrasound signals.
• Applied modified least-squares Prony's method with curve-fitting (MLSP+CF) for signal decomposition.
• Validated methods using plastic and Zerdine® samples with bone-like wave velocities.

Main Results:

• Phase velocity estimates were highly accurate (within 0.4% for slow waves, 1.2% for fast waves).
• Midband signal loss estimates showed good accuracy (within 1.7% for fast waves, 3.7% for slow waves).
• Methods demonstrated accuracy comparable to simulations of published cancellous bone parameters.

Conclusions:

• Bandlimited deconvolution and MLSP+CF accurately decompose overlapping fast and slow waves in cancellous bone ultrasound.
• These methods offer sufficient precision for cancellous bone analysis, with estimation error likely less than experimental error.
• Improved wave separation enhances the potential for detailed assessment of bone characteristics.