Investigation of hardware and software techniques to enhance the characteristics of focused ultrasound (FUS) spectra

Abstract

Objective. Microbubble cavitation generated by focused ultrasound (FUS) can induce safe blood-brain-barrier (BBB) opening allowing therapeutic drug passage. Spectral changes in the hydrophone sensor signal are currently used to distinguish stable cavitation from inertial cavitation that can damage the BBB. Gibbs’ ringing, peak intensity loss and peak width increase are well-known distortions evident when using the discrete Fourier transform (DFT) to transform data containing a few hundred points. We investigate overcoming the fact that FUS time signals (10 ms providing 312 500 points sampled at 32 ns intervals) can generate such sharp spectral peaks that variations in their DFT-related distortions can significantly impact the values of the key metrics used for cavitation characterization. Approach. We introduce low-pass filter hardware to improve how the analogue to digital convertor handles high-frequency noise components and the orders of magnitude differences between FUS harmonic intensities. We investigate the enhanced FUS spectral stability and resolution obtained from a new technique, physical sparsification (PH-SP), customized to the a-priori information that all key FUS components are harmonically related. Results are compared with standard DFT optimizations involving time data windowing and Fourier interpolation. Main results. A new simulation model showed peak intensity, widths and metrics modified by small changes in the transformed signal’s length when removing the noisy starting transient of the FUS hydrophone signal or following minor excitation frequency or sampling rate adjustments. 25%–60% area-under-the-curve changes occurred in phantom studies at different pressure levels. Spectral peak sharpness was best optimized and stabilized with PH-SP. Significance. Special FUS characteristics mean starting transients and minor variations in experimental procedures lead to significant changes in the spectral metrics used to monitor cavitation levels. Customizing PH-SP to these characteristics led to sharper, more stable spectra with the potential to track the impact of microbubble environment changes.