System and method for monitoring ultrasound therapy

Abstract

A computer-implemented method and corresponding system for monitoring ultrasound therapy are described. The method dynamically classifies, by a processor, cavitation effects of at least one ultrasound stimulus on microbubbles introduced into in a biological setting for use in the ultrasound therapy. The cavitation effects are observed via an ultrasound receiver and dynamically classified, by the processor, into stable cavitation effects and inertial cavitation effects. The at least one ultrasound stimulus targets biological tissue by causing the stable and inertial cavitation effects during the ultrasound therapy which responsively induce distinct bio-effects on the biological tissue in the biological setting. The method outputs, by the processor, a representation of an action to be taken for the ultrasound therapy. The action is a function of the stable and inertial cavitation effects dynamically classified, thereby enabling controlled ultrasound-based treatments, minimizing unintended effects on surrounding tissue.

Description

5200.2434002System and Method for Monitoring Ultrasound TherapyRELATED APPLICATIONS

[0001] This application claims the benefit of U. S. Provisional Application No.63 / 723,170, filed on November 21, 2024, and U. S. Provisional Application No. 63 / 783,884, filed on April 4, 2025. The entire teachings of the above applications are incorporated herein by reference.BACKGROUND

[0002] Passive cavitation imaging (PCI) has gained significant attention for its role in monitoring cavitation-facilitated therapeutic applications, including targeted drug delivery, focal ablation, and other therapeutic applications for non-limiting examples. These therapeutic applications often utilize a focused ultrasound (FUS) transducer to induce acoustic cavitation — with or without the introduction of pre-seeded bubbles — enabling spatially precise biological effects. For example, bubble cavitation can promote the controlled release and delivery of therapeutics, allowing treatment to access otherwise inaccessible regions, such as across the blood-brain barrier. Both preclinical and clinical studies have demonstrated enhanced therapeutic efficacy when cavitation-based approaches are applied for treating conditions, such as Alzheimer’s disease, glioblastoma, and other pathologies for non-limiting examples.SUMMARY

[0003] An example embodiment disclosed herein may be used for localizing and classifying cavitation activities, such as for localizing and classifying cavitation activities of passive cavitation imaging (PCI) dynamically and for guiding and controlling ultrasound therapies.

[0004] According to an example embodiment, a computer-implemented method for monitoring ultrasound therapy comprises dynamically classifying, by a processor, cavitation effects of at least one ultrasound stimulus on bubbles in a biological setting. The bubbles are introduced into the biological setting for the ultrasound therapy. The cavitation effects are observed via an ultrasound receiver and dynamically classified, by the processor, into stable cavitation effects and inertial cavitation effects. The stable and inertial cavitation effects- 1 - 4239812.v15200.2434002induce distinct bio-effects on biological tissue in the biological setting. The biological tissue is targeted by the at least one ultrasound stimulus. The computer-implemented method further comprises outputting, by the processor, a representation of an action to be taken for the ultrasound therapy. The action is a function of the stable and inertial cavitation effects dynamically classified.

[0005] The computer-implemented method can further comprise determining, by the processor, respective spatial locations, within the biological tissue, of the stable and inertial cavitation effects dynamically classified. The computer-implemented method can further comprise outputting, by the processor, representations of the respective spatial locations determined. The representations of the respective spatial locations determined and output can enable an operator or controller to make an informed adjustment to the at least one ultrasound stimulus.

[0006] The computer-implemented method can further comprise causing an ultrasound device communicatively coupled to the processor to deliver the at least one ultrasound stimulus. The at least one ultrasound stimulus can be focused upon a therapeutic target within the biological tissue.

[0007] The dynamically classifying can include applying a mathematical decomposition to measurements acquired via the ultrasound receiver to extract components of the measurements. The measurements can represent the cavitation effects of the at least one ultrasound stimulus on the bubbles. The dynamically classifying can be based on one or more subsets of the components extracted via the mathematical decomposition applied. At least a portion of the stable cavitation effects or the inertial cavitation effects can be associated with each subset of the one or more subsets of the components extracted. The dynamically classifying can further include comparing at least a portion of the one or more subsets of the components extracted.

[0008] The computer-implemented method can further comprise, by the processor, causing the ultrasound receiver to acquire the measurements.

[0009] The computer-implemented method can further comprise introducing a shift to the measurements. The shift can be based on a time delay experienced by a receiver element of the ultrasound receiver. The computer-implemented method can further comprise computing a covariance of the measurements. The covariance can be associated with a similarity between each measurement of the measurements. The dynamically classifying can include extracting the components of the covariance of the measurements computed.- 2 - 4239812.v15200.2434002

[0010] The measurements can be time-domain measurements. The applying can include applying the mathematical decomposition to the time-domain measurements. Applying the mathematical decomposition can include applying an Eigen decomposition to extract at least a subset of the components. The subset of the components extracted using the Eigen decomposition can be associated with frequency components of the measurements.

[0011] The stable cavitation effects classified can be associated with oscillation of the bubbles in the biological setting and the inertial cavitation effects classified can be associated with expansion and collapse of the bubbles in the biological setting.

[0012] The computer-implemented method can further comprise, by the processor, rendering a representation of the stable and inertial cavitation effects dynamically classified. The representation can be rendered with respect to the biological setting. The computer-implemented method can further comprise outputting the representation rendered toward displaying the representation rendered on an electronic display. The electronic display can be communicatively coupled to the processor.

[0013] The action to be taken can include (i) adjusting the at least one ultrasound stimulus, (ii) modifying a dosage of the bubbles introduced into the biological setting, or a combination of (i) and (ii) for non-limiting examples.

[0014] According to another example embodiment, a computer-based system for monitoring ultrasound therapy comprises at least one processor and at least one memory. The at least one memory has encoded thereon a sequence of instructions which, when loaded and executed by the at least one processor, causes the computer-based system to dynamically classify cavitation effects of at least one ultrasound stimulus on bubbles in a biological setting. The bubbles are introduced into the biological setting for the ultrasound therapy. The cavitation effects are observed via an ultrasound receiver and dynamically classified into stable cavitation effects and inertial cavitation effects. The stable and inertial cavitation effects induce distinct bio-effects on biological tissue in the biological setting. The biological tissue is targeted by the at least one ultrasound stimulus. The sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to output a representation of an action to be taken for the ultrasound therapy. The action is a function of the stable and inertial effects dynamically cavitation classified.

[0015] Alternative computer-based system embodiments parallel those disclosed above in connection with the example computer-implemented method embodiment.- 3 - 4239812.v15200.2434002

[0016] According to another example embodiment, a non-transitory computer-readable medium has encoded thereon a sequence of instructions which, when loaded and executed by at least one processor, causes the at least one processor to dynamically classify cavitation effects of at least one ultrasound stimulus on bubbles in a biological setting. The bubbles are introduced into the biological setting for the ultrasound therapy. The cavitation effects are observed via an ultrasound receiver and dynamically classified into stable cavitation effects and inertial cavitation effects. The stable and inertial cavitation effects induce distinct bioeffects on biological tissue in the biological setting. The biological tissue is targeted by the at least one ultrasound stimulus. The sequence of instructions further causes the processor to output a representation of an action to be taken for the ultrasound therapy. The action is a function of the stable and inertial cavitation effects dynamically classified.

[0017] Alternative non-transitory computer-readable medium embodiments parallel those disclosed above in connection with the example computer-implemented method embodiment.

[0018] It should be understood that example embodiments disclosed herein can be implemented in the form of a method, apparatus, system, or non-transitory computer readable medium with program codes embodied thereon.BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The patent or application file contains at least one drawing executed in color. Copies of this patent or patent application publication with color drawing(s) will be provided by the Office upon request and payment of the necessary fee.

[0020] The foregoing will be apparent from the following more particular description of example embodiments, as illustrated in the accompanying drawings in which like reference characters refer to the same parts throughout the different views. The drawings are not necessarily to scale, emphasis instead being placed upon illustrating embodiments.

[0021] FIG. 1 A is a diagram of an example embodiment of a focused ultrasound (FUS) therapy system that includes an example embodiment of a computer-based system for monitoring ultrasound therapy.

[0022] FIG. 1B is a block diagram of another example embodiment of a FUS system that includes an example embodiment of a computer-based system for monitoring ultrasound therapy.

[0023] FIG. 2 is a flow diagram of an example embodiment of a computer-implemented method for monitoring ultrasound therapy.- 4 - 4239812.v15200.2434002

[0024] FIG. 3 is a flow diagram of another example embodiment of a computer-implemented method for monitoring ultrasound therapy.

[0025] FIG. 4 is a diagram of an example embodiment of a cavitation imaging setup configured to implement an example embodiment of a computer-implemented method for monitoring ultrasound therapy.

[0026] FIGS. 5A-C are plots illustrating point spread functions (PSFs) of a delay-sum-integrate (DSI) beamformer, Robust Capon Beamformer (RCB), and an example embodiment of a beamformer, respectively.

[0027] FIGS. 6A-C are plots illustrating in-silico passive cavitation images generated using a Vokurka model and beamformed using a Delay-Sum-Integrate (DSI) beamformer, a Robust Capon Beamformer (RCB), and an example embodiment of a beamformer, respectively.

[0028] FIG. 6D is a plot of axial beamwidth profiles of the passive cavitation images of FIGS. 6A-C.

[0029] FIG. 6E is a plot of lateral beamwidth profiles of the passive cavitation images of FIGS. 6A-C.

[0030] FIG. 7 is a plot illustrating in silico passive cavitation images of point-source resolvability in images generated using a DSI beamformer, an RCB, and an example embodiment of a beamformer.

[0031] FIG. 8A is a plot illustrating representative cavitation images of a 5-source simulation generated using a DSI beamformer, a RCB, and an example embodiment of a beamformer.

[0032] FIG. 8B is a plot of mean-squared intensities (MSIs) indicative of artifact reduction ability of the beamformers of FIG. 8 A at different bubble cluster sizes.

[0033] FIGS. 9A-C are plots illustrating beamformed images of a pure tone source and a Vokurka source using parameter m of 2, 3, and 4, respectively, according to an example embodiment.

[0034] FIG. 9D is a plot of frequency spectra of the pure tone source and the Vokurka source.

[0035] FIGS. 9E and 9F are plots of eigenvalue distribution of the pure tone source and the Vokurka source, respectively, of FIGS. 9A-C.- 5 - 4239812.v15200.2434002

[0036] FIG. 10A is a plot illustrating images, reconstructed using example embodiments of a beamformer with parameters m = 1, m = 8, and m = 20, respectively, of a domain including inertial and stable cavitation sources.

[0037] FIG. 10B is a plot of reconstructed contrast values of a stable source and an inertial source generated using example embodiments of beamformers sweeping across parameter m, including the parameter m values of 1, 8, and 20 of FIG. 10 A.

[0038] FIG. 10C is a plot of frequency spectra of a simulated stable source and a simulated inertial source.

[0039] FIG. 11 A is a plot illustrating images, reconstructed using example embodiments of beamformers with parameters m = 1, m = 8, and m = 20, respectively, of a domain including inertial and stable cavitation sources, wherein an amplitude of the inertial sources are set to a factor of two times an amplitude of the stable sources.

[0040] FIG. 11B is a plot of reconstructed contrast values of a stable source and an inertial source generated using example embodiments of beamformers sweeping across parameter m, including the parameter m values of 1, 8, and 20 of FIG. 11 A.

[0041] FIG. 11C is a plot of frequency spectra of a simulated stable source and a simulated inertial source.

[0042] FIG. 12A is a plot illustrating images, reconstructed using example embodiments of beamformers with parameters m = 1, m = 8, and m = 20, respectively, of a domain including inertial and stable cavitation sources, wherein an amplitude of the inertial sources are set to a factor of four times an amplitude of the stable sources.

[0043] FIG. 12B is a plot of reconstructed contrast values of a stable source and an inertial source generated using example embodiments of beamformers sweeping across parameter m, including the parameter m values of 1, 8, and 20 of FIG. 12 A.

[0044] FIG. 12C is a plot of frequency spectra of a simulated stable source and a simulated inertial source.

[0045] FIG. 13 illustrates schematically an in vitro experimental setup for evaluating methods of monitoring ultrasound therapy, the in vitro experimental setup including double tube phantoms.

[0046] FIG. 14A is a plot illustrating cavitation images, acquired using the experimental setup of FIG. 13, beamformed using a DSI beamformer, an RCB (a = 10), and example embodiments of beamformers with parameter m of 1, 8, and 40, respectively.- 6 - 4239812.v15200.2434002

[0047] FIG. 14B is a plot of sizes of -3dB and -6dB regions of a main lobe of the cavitation images of FIG. 14 A.

[0048] FIG. 14C is a plot of example spectra of an inertial cavitation source, stable cavitation source, and a combination of the inertial cavitation source and the stable cavitation source acquired using the experimental setup of FIG. 13.

[0049] FIG. 15 is a plot illustrating, in a normalized linear scale, cavitation images corresponding to the cavitation images of FIG. 14 A.

[0050] FIG. 16 illustrates beamformed images of the experimental setup of FIG. 10 generated using an RCB.

[0051] FIG. 17 illustrates beamformed images of the experimental setup of FIG. 10 generated using example embodiments of beamformers, e.g., PAD AM.

[0052] FIG. 18 illustrates an in vitro experimental setup for evaluating methods of monitoring ultrasound therapy in a rat phantom.

[0053] FIG. 19A is a plot illustrating images acquired using the experimental setup of FIG. 18, including a B-mode ultrasound image and beamformed images using a DSI beamformer, an RCB, and example embodiments of beamformers with parameter m = 1 and m= 10, respectively.

[0054] FIG. 19B is a plot of frequency spectra of radiofrequency (RF) data at three look points in the image of FIG. 19A, with an inset indicating the positions of the look points.

[0055] FIG. 19C is a plot quantifying harmonic, ultraharmonic, and inharmonic components at the look points of FIG. 19B.

[0056] FIG. 20 is a plot illustrating, in a normalized linear scale, cavitation images corresponding to the cavitation images of FIG. 19 A.

[0057] FIG. 21 illustrates beamformed images of the experimental setup of FIG. 18 generated using an RCB.

[0058] FIG. 22 illustrates beamformed images of the experimental setup of FIG. 18 generated using example embodiments of beamformers, e.g., PAD AM.

[0059] FIG. 23A is a plot illustrating a spectrogram of a focused ultrasound (FUS) passive cavitation detector (PCD) signal of cavitation effects induced in a two-tube experimental setup using an FUS stimulus, according to an example embodiment.

[0060] FIG. 23B is a plot illustrating a spectrogram of ultrasound probe channel data of cavitation effects induced in a two-tube experimental setup using an FUS stimulus, according to an example embodiment.- 7 - 4239812.v15200.2434002

[0061] FIGS. 23C and 23D are plots illustrating images formed from ultrasound probe data, including the ultrasound probe channel data of FIG. 23B. The images are beamformed using an example embodiment of a beamformer using parameters of m between 1 and 60 for a tube including inertial cavitation effects and a tube including stable cavitation effects.

[0062] FIG. 23E is a plot illustrating a best m parameter for detecting inertial cavitation and stable cavitation for the tubes including the inertial cavitation effects and the stable cavitation effects of FIGS. 23C and 23D.

[0063] FIG. 23F is a plot of power of harmonic, ultraharmonic, and broadband frequencies in the PCD signal of FIG. 23 A.

[0064] FIGS. 23G and 23H are plots illustrating images beamformed using an example embodiment of a beamformer with parameter m = 1 and m = 20. The images are beamformed using data at a time point of t= 1 ms of ultrasound probe data including the ultrasound probe channel data of FIG. 23B.

[0065] FIGS. 23I and 23J are plots illustrating images beamformed using an example embodiment of a beamformer with parameter m = 1 and m = 20, respectively. The images are beamformed using data at a time point of t= 4 ms of ultrasound probe data including the ultrasound probe channel data of FIG. 23B.

[0066] FIG. 24 is a schematic view of a computer network in which embodiments may be implemented.

[0067] FIG. 25 is a block diagram illustrating an example embodiment of a computer node in the computer network of FIG. 24.DETAILED DESCRIPTION

[0068] A description of example embodiments follows.

[0069] Passive cavitation imaging has explored various beamforming methods to optimize spatial resolution, suppress imaging artifacts, and maintain computational efficiency. These factors may be useful for the clinical translation of Focused Ultrasound (FUS) therapies, where precise cavitation localization and dose control may be used to minimize off-target effects, e.g., off-target cavitation effects. Clinical applications of FUS therapy may include, as non-limiting examples, delivery of therapeutic agents through anatomical barriers (e.g., the blood brain barrier) and cancer treatment, including targeted ablation of tumors. However, the stronger mode of cavitation, which may commonly be- 8 - 4239812.v15200.2434002referred to as inertial cavitation, can be destructive at high acoustic pressures, potentially causing either undesired or intentional damage and cell death due to mechanical stress and hyperthermia.

[0070] Commonly used methods for cavitation localization or dose control, such as Delay-Sum-Integrate (DSI) and Robust Capon Beamforming (RCB), may have demonstrated utility, but may be limited by either significant artifacts or the need for a nonphysical input parameter. To address these challenges, an example embodiment of a method and system described herein can enhance resolution and introduce a physically grounded parameter for signal characterization without compromising computational speed and robustness.

[0071] An example embodiment disclosed herein, which may include an example embodiment referred to herein as Passive Acoustic Dynamic Differentiation and Mapping (PAD AM), may adapt the Multiple Signal Classification method to the time domain to improve cavitation localization and classification. An example embodiment, e.g., PAD AM, can incorporate a physically meaningful input parameter that dynamically reflects the frequency richness of a received signal.

[0072] An example embodiment can achieve, for example, using PAD AM, up to a 6-fold improvement in lateral beamwidth compared to RCB, and a 4-fold reduction in mean-square artifact intensity reduction. Input parameters for an example embodiment can provide a novel physical insight, which may be useful for differentiation between stable and inertial cavitation based on spectral content. Such differentiation between stable and inertial cavitation may reduce reliance on empirically tuned or arbitrary thresholds and simplifies integration into therapy workflows.

[0073] With their ability to improve resolution, reduce artifacts, and provide computational efficiency, embodiments, e.g., PAD AM, may represent a promising advancement for precise cavitation localization and therapy monitoring.

[0074] As disclosed herein, embodiments, e.g., PAD AM, can include a time-domain passive cavitation imaging method that can offer superior resolution and artifact reduction compared to DSI and RCB. Its physically intuitive input parameter enables dynamic differentiation between stable and inertial cavitation, enhancing precision in the monitoring and control of FUS therapy.

[0075] Non-invasive control and monitoring of cavitation facilitated FUS therapy may represent a challenge for adoption of FUS-based therapeutic techniques, traditionally addressed using MRI-based techniques or PCI, the latter being more cost-effective and time- 9 - 4239812.v15200.2434002efficient (T. Sun, Y. Zhang, C. Power, P. M. Alexander, J. T. Sutton, M. Aryal, N.Vykhodtseva, E. L. Miller, and N. J. McDannold, “Closed-loop control of targeted ultrasound drug delivery across the blood-brain / tumor barriers in a rat glioma model,” Proceedings of the National Academy of Sciences, vol. 114, no. 48, pp. E10281-E10290, Nov. 2017, N. McDannold, Y. Zhang, J. G. Supko, C. Power, T. Sun, C. Peng, N. Vykhodtseva, A. J.Golby, and D. A. Reardon, “Acoustic feedback enables safe and reliable carboplatin delivery across the blood-brain barrier with a clinical focused ultrasound system and improves survival in a rat glioma model,” Theranostics, vol. 9, no. 21, pp. 6284-6299, Aug. 2019, C. D. Arvanitis and N. McDannold, “Integrated ultrasound and magnetic resonance imaging for simultaneous temperature and cavitation monitoring during focused ultrasound therapies,” Medical Physics, vol. 40, no. 11, p. 112901, 2013). PCI can use a passive listening device to estimate cavitation power and localize the treatment, which can make PCI a topic of significant interest in array signal processing (C. Coviello, R. Kozick, J. Choi, M. Gyongy, C. Jensen, P. P. Smith, and C.-C. Coussios, “Passive acoustic mapping utilizing optimal beamforming in ultrasound therapy monitoring,” The Journal of the Acoustical Society of America, vol. 137, no. 5, pp. 2573-2585, May 2015). High-resolution beamforming methods may be useful for improving the clinical viability of this passive imaging. Conventional approaches, such as Delay-Sum Integrate (DSI) and the Robust Capon Beamformer (RCB), have been used for cavitation power estimation and localization, and each may display inherent trade-offs in resolution, speed, and artifact suppression (M. Gyongy, M. Arora, J. A. Noble, and C. C. Coussios, “Use of passive arrays for characterization and mapping of cavitation activity during HIFU exposure,” in 2008 IEEE Ultrasonics Symposium, Nov. 2008, pp. 871-874, K. J. Haworth, V. A. Salgaonkar, N. M. Corregan, C. K. Holland, and T. D. Mast, “Using Passive Cavitation Images to Classify High-Intensity Focused Ultrasound Lesions,” Ultrasound in medicine & biology, vol. 41, no. 9, p. 2420, Jun. 2015, M. Gyongy and C.-C. Coussios, “Passive spatial mapping of inertial cavitation during HIFU exposure,” IEEE transactions on bio-medical engineering, vol. 57, no. 1, pp. 48-56, Jan. 2010, C. H. Farny, R. G. Holt, and R. A. Roy, “Temporal and spatial detection of HIFU-induced inertial and hot-vapor cavitation with a diagnostic ultrasound system,” Ultrasound in Medicine & Biology, vol. 35, no. 4, pp. 603-615, Apr. 2009, C. Coviello, R. Kozick, J. Choi, M. Gyongy, C. Jensen, P. P. Smith, and C.-C. Coussios, “Passive acoustic mapping utilizing optimal beamforming in ultrasound therapy monitoring,” The Journal of the Acoustical Society of America, vol. 137, no. 5, pp. 2573-2585, May 2015, K. J. Haworth, N. G. Salido, M. Lafond,- 10 - 4239812.v15200.2434002D. S. Escudero, and C. K. Holland, “Passive Cavitation Imaging Artifact Reduction Using Data-Adaptive Spatial Filtering,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 70, no. 6, pp. 498-509, Jun. 2023).

[0076] Classical Passive Beamforming Techniques: The Delay-And-Sum (DAS) beamformer may remain a commonly used method due to its computational efficiency (M. Gyongy, M. Arora, J. A. Noble, and C. C. Coussios, “Use of passive arrays for characterization and mapping of cavitation activity during HIFU exposure,” in 2008 IEEE Ultrasonics Symposium, Nov. 2008, pp. 871-874, K. J. Haworth, V. A. Salgaonkar, N. M. Corregan, C. K. Holland, and T. D. Mast, “Using Passive Cavitation Images to Classify High-Intensity Focused Ultrasound Lesions,” Ultrasound in medicine & biology, vol. 41, no.9, p. 2420, Jun. 2015, M. Gyongy and C.-C. Coussios, “Passive spatial mapping of inertial cavitation during HIFU exposure,” IEEE transactions on bio-medical engineering, vol. 57, no. 1, pp. 48-56, Jan. 2010, C. H. Farny, R. G. Holt, and R. A. Roy, “Temporal and spatial detection of HIFU-induced inertial and hot-vapor cavitation with a diagnostic ultrasound system,” Ultrasound in Medicine & Biology, vol. 35, no. 4, pp. 603-615, Apr. 2009). DAS can estimate pixel intensity by summing delayed signals from multiple transducer channels based on their respective distances to the pixel location (V. Perrot, M. Polichetti, F. Varray, and D. Garcia, “So you think you can DAS? A viewpoint on delay-and-sum beamforming,” Ultrasonics, vol. 111, p. 106309, Mar. 2021). A subsequent variant, Delay-Sum Integrate (DSI), extends DAS to cases wherein the incoming wavefront time is not known, or there are many wavefronts to consider. By integrating across all available time indices, DSI can provide a long-exposure-like reconstruction of cavitation events (S. Norton and I. Won, “Time exposure acoustics,” IEEE Transactions on Geoscience and Remote Sensing, vol. 38, no. 3, pp. 1337-1343, May 2000). DSI may be implemented in either the time or frequency domain, with accuracy and speed improvements achieved by selecting frequencies associated with stable or inertial cavitation (K. J. Haworth, K. B. Bader, K. T. Rich, C. K. Holland, and T. D. Mast, “Quantitative Frequency-Domain Passive Cavitation Imaging,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 64, no. 1, pp. 177-191, Jan. 2017). Although DSI can be well-suited for real-time applications due to its computational efficiency (S. Bae, K. Liu, A. N. Pouliopoulos, R. Ji, and E. E. Konofagou, “Real-Time Passive Acoustic Mapping With Enhanced Spatial Resolution in Neuronavigation-Guided Focused Ultrasound for Blood-Brain Barrier Opening,” IEEE transactions on bio-medical engineering, vol. 70, no. 10, pp. 2874-2885, Oct. 2023, H. A. S.- 11 - 4239812. vl5200.2434002Kamimura, S.-Y. Wu, J. Grondin, R. Ji, C. Aurup, W. Zheng, M. Heidmann, A. N.Pouliopoulos, and E. E. Konofagou, “Real-Time Passive Acoustic Mapping Using Sparse Matrix Multiplication,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 68, no. 1, pp. 164-177, Jan. 2021, E. Lyka, C. M. Coviello, C. Paverd, M. D. Gray, and C. C. Coussios, “Passive Acoustic Mapping Using Data-Adaptive Beamforming Based on Higher Order Statistics,” IEEE Transactions on Medical Imaging, vol. 37, no. 12, pp. 2582-2592, Dec. 2018, R. M. Jones, D. McMahon, and K. Hynynen, “Ultrafast three dimensional microbubble imaging in vivo predicts tissue damage volume distributions during nonthermal brain ablation,” Theranostics, vol. 10, no. 16, pp. 7211-7230, 2020), DSI can suffer from comparatively poor resolution and a characteristic tail artifact, a nonphysical effect resulting from the overlapping wavefront delays across channels.

[0077] To mitigate these limitations, adaptive beamformers have been explored, which may notably include the Capon beamformer and an enhanced version of the Capon beamformer that can prevent self-nulling, the Robust Capon Beamformer (RCB) (C.Coviello, R. Kozick, J. Choi, M. Gybngy, C. Jensen, P. P. Smith, and C.-C. Coussios, “Passive acoustic mapping utilizing optimal beamforming in ultrasound therapy monitoring,” The Journal of the Acoustical Society of America, vol. 137, no. 5, pp. 2573-2585, May 2015, K. J. Haworth, N. G. Salido, M. Lafond, D. S. Escudero, and C. K. Holland, “Passive Cavitation Imaging Artifact Reduction Using Data-Adaptive Spatial Filtering,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 70, no. 6, pp. 498-509, Jun. 2023, S. Bae, K. Liu, A. N. Pouliopoulos, R. Ji, and E. E. Konofagou, “Real-Time Passive Acoustic Mapping With Enhanced Spatial Resolution in Neuronavigation-Guided Focused Ultrasound for Blood-Brain Barrier Opening,” IEEE transactions on bio-medical engineering, vol. 70, no. 10, pp. 2874-2885, Oct. 2023). Several studies have shown that RCB improves localization accuracy compared to DSI and that imaging results may align well with measured bio-effects in vitro (K. J. Haworth, K. B. Bader, K. T. Rich, C. K.Holland, and T. D. Mast, “Quantitative Frequency -Domain Passive Cavitation Imaging,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 64, no. 1, pp.177-191, Jan. 2017). However, despite the advantages of RCB with respect to resolution, RCB may be computationally intensive and may require tuning a steering vector uncertainty parameter, a (C. Coviello, R. Kozick, J. Choi, M. Gybngy, C. Jensen, P. P. Smith, and C.-C. Coussios, “Passive acoustic mapping utilizing optimal beamforming in ultrasound therapy- 12 - 4239812. vl5200.2434002monitoring,” The Journal of the Acoustical Society of America, vol. 137, no. 5, pp. 2573-2585, May 2015). This parameter, defined as:||a(0o) - u(0o+ ^)l|2 < E (1)quantifies the allowable uncertainty in the steering vector a(9) (J. Li, P. Stoica, and Z. Wang, “On robust Capon beamforming and diagonal loading,” IEEE Transactions on Signal Processing, vol. 51, no. 7, pp. 1702-1715, Jul. 2003). Selecting an optimal £ can be performed once for an imaging application; however, the parameter may contain no tangible meaning in the context of cavitation physics. Some works suggest tuning it to around 8.45 (K. J. Haworth, N. G. Salido, M. Lafond, D. S. Escudero, and C. K. Holland, “Passive Cavitation Imaging Artifact Reduction Using Data-Adaptive Spatial Filtering,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 70, no. 6, pp. 498-509, Jun. 2023), but the parameter can be of arbitrary order of magnitude for a given application, field of view, or source location among other factors. Eigenspace-based RCB has been proposed to reduce the variance of the input parameter (S. Lu, H. Hu, X. Yu, J. Long, B. Jing, Y. Zong, and M. Wan, “Passive acoustic mapping of cavitation using eigenspace-based robust Capon beamformer in ultrasound therapy,” Ultrasonics Sonochemistry, vol. 41, pp. 670-679, Mar.2018), but £ remains an empirically determined factor rather than a physically measurable parameter. Multiple Signal Classification (MUSIC) is established in far-field array sensing (such as radar and sonar applications) for high-resolution and high signal to noise ratio (SNR) direction finding. MUSIC is a frequency-domain direction-of-arrival beamformer that uses eigenvalue analysis assuming the presence of m scatterers (R. Schmidt, “Multiple emitter location and signal parameter estimation,” IEEE Transactions on Antennas and Propagation, vol. 34, no. 3, pp. 276-280, Mar. 1986). The incoming signal can be assumed to be modeled as the summation of discrete sinusoidal sources and noise, which is divisible into two orthogonal subspaces: the signal subspace and the noise subspace, and linear algebra can be used to separate the two. MUSIC was previously explored by Polichetti et al (M. Polichetti, F. Varray, B. Gilles, J.-C. B era, and B. Nicolas, “Use of the Cross-Spectral Density Matrix for Enhanced Passive Ultrasound Imaging of Cavitation,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 68, no. 4, pp. 910-925, Apr. 2021) for PCI, wherein the authors noted the beamformer’s ability to distinguish low-intensity sources with a trial-error estimate of the parameter m. The computational study noted that a single bubble source contributed more than one eigenvector to the signal subspace in some cases, opening the question of the physical meaning of the MUSIC parameter in the context of PCI.- 13 - 4239812. vl5200.2434002

[0078] Proposed Approach: Cavitation emissions from multiple bubble sources may be generated with the same drive frequency, and the stochastic time variances involved may be on smaller scales than the sampling rates involved in passive imaging. Thus, the signals may be correlated, and the MUSIC parameter m may not correspond to the exact number of sources in the incoming signal, consistent with the observations in Polichetti et al (M.Polichetti, F. Varray, B. Gilles, J.-C. B era, and B. Nicolas, “Use of the Cross-Spectral Density Matrix for Enhanced Passive Ultrasound Imaging of Cavitation,” IEEE Transactions on Ultrasonics, Ferroelectrics, and Frequency Control, vol. 68, no. 4, pp. 910-925, Apr. 2021). Instead, broadband signals present in inertial cavitation may represent a signal subspace that may be uncorrelated with the harmonic and ultraharmonic signals associated with stable cavitation. Furthermore, for a signal containing both inertial and stable cavitation emission, there may exist a value of m for that can separate the two phenomena.

[0079] To explore the role of the parameter m and address the limitations of existing beamforming approaches, a time-domain adaptation of the MUSIC direction-finding method to the field of PCI may be introduced. An example embodiment of a time-domain adaption of the MUSIC method may be referred to as Passive Acoustic Dynamic Differentiation and Mapping (PAD AM), which may emphasize the embodiment’s unique ability to dynamically classify the spectral richness of acoustic emissions. Embodiments, e.g., PAD AM, can leverage time-domain MUSIC to overcome the trade-offs commonly encountered in existing beamformers — namely, between resolution, computational speed, and artifact suppression. Embodiments can also offer a physically grounded interpretation of the m parameter by exploiting its capacity to distinguish frequency signatures based on the rank characteristics of the incoming signal’s spatial covariance matrix.

[0080] FIG. 1 A is a diagram of an example embodiment of a focused ultrasound (FUS) therapy system 100a that includes an example embodiment of a computer-based system 116a for monitoring ultrasound therapy. The FUS therapy system 100a can include an ultrasound transducer 102 configured to deliver a focused ultrasound stimulus 104a to a biological setting 106a, for example, the brain for non-limiting example, wherein the ultrasound therapy may be useful for, as a non-limiting example, enabling therapeutic agents to penetrate through the blood brain barrier.

[0081] In some embodiments, the ultrasound transducer 102a can be operated or positioned by an operator 112, for example, a clinician or a researcher. The biological setting 106a can include, for example, a biological tissue or an anatomical location of a human or an- 14 - 4239812. vl5200.2434002animal. Further, bubbles (not illustrated but further described herein with reference to FIGS. IB and 4) may be introduced to the biological setting 106a. The focused ultrasound stimulus 104a can cause cavitation effects, e.g., stable cavitation and inertial cavitation effects, on the bubbles, the cavitation effects generating a signal 108a that can be detected by an ultrasound receiver 110a, which may be an ultrasound probe or an imaging probe (not shown). The ultrasound receiver 110a can be communicatively coupled 114 to the computer-based system 116a for monitoring ultrasound therapy. The computer-based system 116a may be configured to dynamically classify and localize inertial and stable cavitation associated with ultrasound therapy. In some embodiments, the computer-based system 116a may be operated by the operator 112.

[0082] The computer-based system 116a may be further configured to output a representation 118a of an action to be taken for the ultrasound therapy. The representation 118a of the action can be based on the inertial and stable cavitation effects classified. The representation 118a may be helpful for guiding the operator 112 in delivering ultrasound therapy or enabling the operator 112 to make an informed adjustment to the ultrasound therapy.

[0083] FIG. IB is a block diagram of another example embodiment of a focused ultrasound (FUS) system 100b that includes a computer-based system 116b for monitoring ultrasound therapy. The computer-based system 116b can comprise at least one processor 184 and at least one memory 190. The at least one memory 190 can have encoded thereon a sequence of instructions which, when loaded and executed by the at least one processor 184, can cause the computer-based system 116b to dynamically classify cavitation effects 127 of at least one ultrasound stimulus 104b on bubbles 120, in a biological setting 106b. The bubbles 120 can be introduced into the biological setting 106b for the ultrasound therapy. The cavitation effects 127 can be observed via an ultrasound receiver 110b and dynamically classified into stable cavitation effects 121 and inertial cavitation effects 123. The stable cavitation effects 121 and inertial cavitation effects 123 can induce distinct bio-effects on biological tissue 125 in the biological setting 106b. The biological tissue 125 can be targeted by the at least one ultrasound stimulus 104b. The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to output a representation 118b of an action to be taken for the ultrasound therapy. The action can be a function of the stable cavitation effects 121 and inertial cavitation effects 123 dynamically classified.- 15 - 4239812. vl5200.2434002

[0084] The representation 118b of the action may serve as feedback that enables an operator or controller to make an informed adjustment to the at least one ultrasound stimulus. It should be understood that the representation 118b may be formatted or scaled in a manner known in the art such that the representation 118b can be interpreted by the operator or controller in a manner that enables the adjustment to be made to the at least one ultrasound stimulus, thereby enabling a controlled ultrasound-based treatment with minimized unintended effects on surrounding tissue that surrounds the biological tissue 125 that is targeted.

[0085] The stable cavitation effects 121 and the inertial cavitation effects 123 may generate a signal(s) 108b detectable by the ultrasound receiver 110b. The ultrasound receiver 110b may be communicatively coupled 114b to the computer-based system 116b. In some embodiments, the ultrasound receiver 110b may be coupled to the computer-based system 116b via a wired or wireless connection. In some embodiments, the computer-based system 116 may be configured to access data acquired by an ultrasound receiver 110b without being directly coupled to the ultrasound receiver 110b.

[0086] Bubbles as disclosed herein, for example, the bubbles 120, can be microbubbles, nanobubbles, or a combination thereof for non-limiting examples.

[0087] The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to determine respective spatial locations, within the biological tissue, of the stable and inertial cavitation effects dynamically classified. The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to output representations of the respective spatial locations determined. The representations of the respective spatial locations determined and output can enable an operator or controller to make an informed adjustment to the at least one ultrasound stimulus.

[0088] The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to cause an ultrasound device communicatively coupled to the processor to deliver the at least one ultrasound stimulus. The at least one ultrasound stimulus can be focused upon a therapeutic target within the biological tissue.

[0089] The dynamically classifying, by the at least one processor 184 of computer-based system 116b, can include applying a mathematical decomposition to measurements acquired via the ultrasound receiver to extract components of the measurements. The measurements- 16 - 4239812. vl5200.2434002can represent the cavitation effects of the at least one ultrasound stimulus on the bubbles. The dynamically classifying can be based on one or more subsets of the components extracted via the mathematical decomposition applied. At least a portion of the stable cavitation effects or the inertial cavitation effects can be associated with each subset of the one or more subsets of the components extracted. The dynamically classifying can further include comparing at least a portion of the one or more subsets of the components extracted.

[0090] The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to cause the ultrasound receiver to acquire the measurements.

[0091] The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to introduce a shift to the measurements. The shift can be based on a time delay experienced by a receiver element of the ultrasound receiver. The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to compute a covariance of the measurements. The covariance can be associated with a similarity between each measurement of the measurements. The dynamically classifying can include extracting the components of the covariance of the measurements computed.

[0092] The measurements can be time-domain measurements. The applying can include applying the mathematical decomposition to the time-domain measurements. Applying the mathematical decomposition can include applying an Eigen decomposition to extract at least a subset of the components. The subset of the components extracted using the Eigen decomposition can be associated with frequency components of the measurements.

[0093] The stable cavitation effects classified can be associated with oscillation of the bubbles in the biological setting and the inertial cavitation effects classified can be associated with expansion and collapse of the bubbles in the biological setting.

[0094] The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to render a representation of the stable and inertial cavitation effects dynamically classified. The representation can be rendered with respect to the biological setting. The sequence of instructions, when loaded and executed by the at least one processor 184, can further cause the computer-based system 116b to output the representation rendered toward displaying the representation rendered on an electronic display. The electronic display can be communicatively coupled to the processor.- 17 - 4239812. vl5200.2434002

[0095] The action to be taken can include (i) adjusting the at least one ultrasound stimulus, (ii) modifying a dosage of the bubbles introduced into the biological setting, or a combination of (i) and (ii) for non-limiting examples.

[0096] The computer-based system 116b may be employed as the computer-based system 116a of FIG. 1 A and may be configured to implement a computer-implemented method for monitoring ultrasound therapy, such as disclosed below with reference to FIG. 2.

[0097] FIG. 2 is a flow diagram of an example embodiment of a computer-implemented method for monitoring ultrasound therapy (201). The computer-implemented method starts 203 and comprises dynamically classifying, by a processor, cavitation effects of at least one ultrasound stimulus on bubbles in a biological setting (205). The bubbles can be introduced into the biological setting for the ultrasound therapy. The cavitation effects can be observed via the ultrasound receiver and dynamically classified, by the processor, into stable cavitation effects and inertial cavitation effects. The stable and inertial cavitation effects can induce distinct bio-effects on biological tissue in the biological setting. The biological tissue can be targeted by the at least one ultrasound stimulus. The computer-implemented method further comprises outputting, by the processor, a representation of an action to be taken for the ultrasound therapy, the action being a function of the stable and the inertial cavitation effects dynamically classified (207). The computer-implemented method thereafter ends (209) in the example embodiment.

[0098] FIG. 3 is a flow diagram of another example embodiment of a computer-implemented method 301 for monitoring ultrasound therapy. The computer-implemented method 301 can start 303 by acquiring 305 measurements of a biological tissue or a biological setting. In some embodiments, the measurements can include raw voltage measurements from sensor elements overtime. The computer-implemented method 301 can further include generating 307 one or more images of the biological tissue / setting, wherein a given image can be based on a number of assumed sources m and a number of locations (e.g., pixels) of the biological tissue / setting. The generating 307 can include applying 309 a delay to ultrasound measurements. The delay can be based on a geometric distance or time of flight between a given sensor unit of a receiver, for example, the receiver 110 described herein with reference to FIG. 1, and a location (e.g., a pixel) of the biological tissue or setting. The generating 307 can further include computing 311 a covariance between time-domain measurements of each sensor unit of the receiver and decomposing 313 the covariance, e.g., into eigenvalues and eigenvectors. The generating 307 can further include computing 315 a- 18 - 4239812. vl5200.2434002value for the location in the biological setting based on a noise subspace (subset of eigenvectors) or the covariant matrix. The noise subspace may be determined by the number of assumed sources m. In some embodiments, the computing 315 the value may be based on an inverse of the noise subspace.

[0099] The steps of the generating 307, / .<?., the applying 309, the computing 311, the decomposing 313, and the computing 315, may be repeated for the number of locations and for each number of assumed sources m desired.

[0100] The computer-implemented method 301 further includes classifying inertial versus stable cavitation based on the values computed for the locations and the one or more images. In some embodiments, each image can be generated using a given number of assumed sources m. As such, multiple images can be generated by repeating the generating 307 with different values of m. The computer-implemented method 301 can subsequently end 319.

[0101] FIG. 4 is a diagram of an example embodiment of a cavitation imaging setup 400 implementing a method for monitoring ultrasound therapy. The cavitation imaging setup 400 can include a focused ultrasound transducer 402 configured to generate a stimulus 404 that can cause cavitation effects 421 in bubbles 420, e.g., lipid nanoparticles, in a biological setting 406, e.g., targeted vasculature. In some embodiments, the bubbles 420 can include microbubbles, nanobubbles, or a combination thereof for non-limiting examples. The cavitation effects 421 in the bubbles 420 can cause bioeffects on a biological tissue 425 of the biological setting. The cavitation effects can further cause a signal 408 detectable by a receiver 410, e.g., an ultrasound probe. The receiver 410 can include a sensor array 422 configured to detect the signal 408, which may be a radiofrequency (RF) signal. A detected signal 424 can be used to generate a beamformed image 426 via an example embodiment of a method for monitoring ultrasound therapy as disclosed herein.

[0102] The action can include, as an example embodiment, changing a delivery of the focused ultrasound (e.g., target location, direction, angle, dosage) or of the microbubbles (e.g., dosage or concentration).

[0103] Measurements of the signal 408 by the receiver 410 can represent the effects of the at least one ultrasound stimulus on the bubbles (e.g., the stable cavitation and the inertial cavitation of the bubbles 420). The method 201 of FIG. 2 and the method 301 of FIG. 3 may be useful for classifying the effects represented by the measurements as stable cavitation,- 19 - 4239812. vl5200.2434002inertial cavitation, or combination thereof and for localizing the effects within a biological setting.

[0104] While embodiments described herein are directed towards monitoring ultrasound therapy in the brain, it should be understood that the computer-implemented methods and systems described herein can be applied to other therapeutic applications. As non-limiting examples, focused ultrasound may be applied towards cancer treatments, e.g., tumor ablation, or kidney diseases, including kidney stones.

[0105] While embodiments described herein are directed towards cavitation effects of at least one ultrasound stimulus on microbubbles, it should be understood that the computer-implemented methods and systems described herein can be applied to types of bubbles. As non-limiting examples, the cavitation effects can include cavitation effects of at least one ultrasound stimulus on nanobubbles.

[0106] The remainder of this includes: Methods that include 1) a derivation of the PAD AM signal model, 2) an explanation of MUSIC in the frequency domain, 3) a derivation of PAD AM in the time domain, 4) an introduction of the computational models used, 5) an introduction of the experimental methods used, and 6) an explanation of the metrics used to compare beamformer performances. The Results section presents findings from both in silico and in vitro experiments, analyzed using the aforementioned metrics. The Discussion section compares PAD AM’s capabilities and limitations against DSI and RCB. The Conclusion summarizes key findings and suggests directions for future research.

[0107] METHODS

[0108] Signal Model

[0109] The original MUSIC is a frequency domain direction-of-arrival beamformer that uses eigenvalue analysis and assumes the presence of m uncorrelated scatterers (R. Schmidt, “Multiple emitter location and signal parameter estimation,” IEEE Transactions on Antennas and Propagation, vol. 34, no. 3, pp. 276-280, Mar. 1986). It can divide the received radio frequency (RF) data into a noise and signal subspace, and can assume the incoming signal can be modeled as the summation of discrete sinusoidal sources and noise:-’'('Mx / tj=•'4(MXD)5(DX / C) +n(Mxkj (2)

[0110] Here, A is the matrix of steering vectors corresponding to AT channels with D actual sources, 5 is the signal vector for a snapshot with k samples, and n is the noise vector.- 20 - 4239812. vl5200.2434002For convenience, k = 1 may be used for the remainder of the signal model. Next, the spatial covariance matrix Rxx may be defined as the expected value of the sample covariance matrix:Rxx= E{xxH}=ARsAH+ <2I (3)

[0111] Where the signal covariance matrix Rs= E{ss"j is of size D *D, and the spatial covariance matrix is of size AT *M. To satisfy the assumption that Rxx = E{xx"k multiple snapshots may be used:(4)

[0112] Because Rxx = ARsAH+ u2 / , the rank of Rsis equal to the number of sources £>, and the rank of the noise matrix o2I is zero, the rank of RxX is also D. This means that there are D nonzero eigenvalues of Rxx corresponding to the signal, and M~D near-zero eigenvalues corresponding to noise. The eigenvalue decomposition of Rxx is:Rxx= VMT1(5)

[0113] Where A = “1kiln and t is the i’th eigenvalue, and Fis a matrix of the eigenvectors, sorted in the order of the largest to smallest eigenvalues.

[0114] Classical MUSIC Reconstruction

[0115] With the signal model established, reconstruction using MUSIC may proceed as follows: The incoming signal may either be accumulated or divided into “snapshots”,X(ntxnc), which may be averaged in the frequency-domain, nt is the number of time samples, and ncis the number of transducer channels, so x is a matrix with nt rows and nccolumns. The spatial covariance matrix estimate can also generated during this step by taking the inner product:1X,„. = — yjQX, (6)' '~L~z4239812. vl5200.2434002

[0116] Where nsis the number of snapshots used. Note that Xi is the frequency-domain representation of the i’th snapshot, requiring a Fourier transform. Next, the eigenvalue decomposition of the averaged spatial covariance matrix estimate can becomputed:X = FAF-1(7)

[0117] From this decomposition, the number m of scatterers may be selected, assuming the signal rank D of X. The m largest eigenvalues are selected from the decomposition, where m < M.21,22,... Am G A (8)

[0118] The associated eigenvectors from each of these m eigenvalues may form the signal subspace, Vs G (M x / ??), and the remaining nc- m + 1 eigenvectors may form the noise subspace, Vn G (M *M~ ni). For each direction (or pixel location in the near-field), the pseudo-spectral intensity can be calculated as the inverted inner product of the steered noise subspace:

[0119] PMUSIC(^) = (9)with A((|)) as the steering vector. Notably, from the derivation provided hereinabove, MUSIC may not yield a power estimate. MUSIC may instead be inversely proportional to the noise power; when the noise subspace is orthogonal and uncorrelated (z.e., the signal subspace is highly correlated), then the MUSIC pseudo-spectrum is large.

[0120] PAD AM - MUSIC Time-Domain

[0121] The MUSIC method may be adapted for the time-domain to form embodiments of a method for cavitation imaging and therapy monitoring, e.g., the PAD AM method. The input signal x(nt,nc) can be delayed in time according to the geometric distance between each transducer’s position p(x, y) and a pixel location a(x, y), using the one-way time of flight at the speed of sound, following the DSI approach:- 22 - 4239812. vl5200.2434002„. r |p('O + A. Hwhere i is the index of a channel or transducer. The delays may be computed using a circular shift including all samples, or myPsamples corresponding to the hyperbola shape of the delay across channels may be omitted, as done in this work, and indicated by ^thyPin (10).

[0122] Next, the delayed signal can be treated as nt - myPsnapshots in time, and averaged in the time-domain:X = V X.' -X (11)

[0123] As in the frequency-domain approach, the eigen-analysis can be performed, but in this case, the entire decomposition is real-valued:x = V AV-1(12)

[0124] The pixel value can be calculated from the noise subspace as before, but because a delay has already been applied, the subspace inner product may no longer need to be steered during this step. Specifically, the steering vector A(<j>) can be replaced with a vector of ones, which is equivalent to the two-dimensional summation of the matrix elements:„. 1E E HA / The method is laid out step-by-step in Method 1 described hereinbelow.4239812. vl5200.2434002Method 1: PAD AM Time-DomainInput: Delay Matrix D(nxXnyXns) containing the delays in units of indices at the sample rate for the one-way time of flight from each pixel in the lateral (nx) and axial (ny) directions to each transducer element ns). Signal matrix S(ntXns), with the RF data for nt time points and the nstransducer elements, Number of PAD AM assumed sources, mOutput: Beamformed Image B(nxXny)1 S(ntxns) <" RF data2 for z <— 1 to nx,j <— 1 to nydo3 for k <— 1 to nsdo4 S <— circshift(S, Dij,k)5 R(nsxns) zeros6 for n <— 1 to nt do7 R = R + SnS^8 PAP-1<- eig(R)9 A, V <- sort (A)10 Vnoise <— V(-m + 1):rls11Bij — — - - ^noise^noise

[0125] In comparison to classical MUSIC, embodiments described herein, e.g., the PAD AM time-domain approach, may offer the advantage of more extensive snapshot averaging, at the sacrifice of a potentially expensive average calculation recalculating the eigenvalue decomposition for each pixel. One significant advantage of this time-domain method may include its simplicity. The frequency domain method can require careful management of snapshot lengths to retain relevant frequency information when splitting a frame of RF data into snapshots. Moreover, the time-domain method may eliminate the need for windowing, overlapping, and dividing the signal into deliberate snapshots, offering a more straightforward approach without sacrificing essential data integrity.

[0126] Computational Bubble Model

[0127] Vokurka’s bubble signal model may provide a computational framework for simulating the time-series behavior of cavitating microbubbles using stochastic distributions.- 24 - 4239812. vl5200.2434002Vokurka’s model may offer a unique platform for testing beamformer performance, as the location of a bubble source can be arbitrarily controlled in both space and time. The model may be based on Vokurka’s observations on cavitation behavior, where a cavitation event can be modeled over time by the equation (K. Vokurka, “A simple model of a vapor bubble,” The Journal of the Acoustical Society of America, vol. 81, no. 1, pp. 58-61, Jan. 1987, S. Buogo and K. Vokurka, “Intensity of oscillation of spark-generated bubbles,” Journal of Sound and Vibration, vol. 329, no. 20, pp. 4266-4278, Sep. 2010, (K. Vokurka, “Cavitation noise modeling and analyzing,” in CD-ROM Proceedings of Forum Acousticum, vol. 2002, 2002):(14)where to represents the time of the cavitation event, is a stochastic variable representing a small random time shift, and 0, referred to by Vokurka as the “time-condition”, controls the width of the cavitation event’s build-up and ramp-down. Q can also be manipulated stochastically. The process may be repeated at regular time periods Z, corresponding to a FUS pressure cycle’s period, and for many point-scatterers:where NPis the number of periods, and Nb is the number of scatterers. To model the RF data received at a transducer’s location, x(t) can be delayed by the propagation time between the scatterer’s location and the transducer, and scaled by - to account for proportional spherical spreading. Additionally, a bandpass filter can be applied to mimic the frequency response of an imaging probe. A virtual probe may be simulated. The virtual probe used herein consists of a 64-element 82mm linear array with a center frequency of 3.21 MHz, a pitch of 1.28mm, and a pass-band of 1.2 to 5.2 MHz. Image reconstruction was performed for a 128x128 pixel grid using a 1000 time-point integration window sampled at 12.8 MHz.

[0128] Experimental Methods and Materials- 25 - 4239812. vl5200.2434002

[0129] To validate the computational results and evaluate beamformer performance of an example embodiment of cavitation mapping, two in vitro models may be used. In both models, in-house lipid microbubbles were diluted and infused through tubing using a syringe with an 18-gauge tip. The microbubbles can be prepared using DSPC and DSPE-PEG2000 at a 2.53:1 weight ratio, with perfluoropropane as the gas core. The resulting bubbles may be poly disperse ranging from 200nm to IO / / ?? in diameter, with the majority measuring 1 - 2 m.The concentration can be approximately 5* 109bubbles per mL. For imaging, the bubbles may be diluted 1:1000 in phosphate buffered saline (PBS) solution.

[0130] According to an example embodiment, in the double tube phantom setup, further described hereinbelow with reference to FIG. 13, two tubes, for example, 1 mm inner-diameter PTFE tubes, can be positioned in the focal region of a 500-kHz FUS transducer. The tubes were spaced apart axially to the FUS focal direction by 5 mm so that the bottom tube was in the center of the focus at 24.5 mm, and the top tube at 29.45 mm. Additionally, the top tube was spaced apart laterally by 2.3 mm so that the expected pressure at the far-tube was approximately 0.3 times the pressure amplitude in the center of the focus. Cavitation can be induced using a peak rarefactional pressure of 1.5 MPa in the focus for the near-tube, corresponding to 0.45 MPa in the far tube, with a pulse length of 10 ms and a repetition frequency of 10 Hz. An imaging probe can be aligned along the radial axis of the FUS focus, at 25mm from the far tube. The imaging probe can be selected to analyze the high harmonics and ultraharmonics of the incoming signal while also providing an adequate frequency range for higher resolution B-Mode images. The imaging probe may be used to transmit a basic B-mode flash sequence for localization, followed by passive RF data acquisition for cavitation imaging. Both B-mode and passive cavitation images may be captured using a computer-based system, e.g., a computer-based ultrasound system, sampled at 33.25 MHz.Beamforming can be performed on a computer communicatively coupled to the computer-based ultrasound system. Inertial cavitation activity can be confirmed in the near-tube region, while stable cavitation behavior can be observed in the far-tube, as assessed by passive cavitation detection.

[0131] To test the robustness of each beamformer with a skull phantom, a rat skull may be positioned in water, e.g., at 25 mm below the FUS transducer, as illustrated herein with respect to FIG. 18. The imaging probe can be placed at a 90-degree angle to the FUS focal axis, aligned parallel to the anterior-posterior axis of the skull. The skull was rotated 45 degrees toward the probe to ensure that both the FUS beam and resulting cavitation emissions- 26 - 4239812. vl5200.2434002passed through the parietal bone. A 1.2 mm hole was drilled on the frontal bone, and a 1 mm inner-diameter PTFE tube was threaded through the cranial cavity. Cavitation can be induced using 500-kHz FUS at a peak negative pressure of 1.5 MPa for 10 seconds, with a pulse length of 10 ms and a pulse repetition frequency of 10 Hz. The computer-based ultrasound system can be used to acquire RF-Data and generate B-mode localization images, sampled at 33.25 MHz. According to an embodiment, cavitation image reconstruction can be performed on a 128x128 pixel grid using a 2048 time-point integration.

[0132] Comparison Metrics

[0133] To effectively compare the performance of the beamformers, several metrics may be used. The point spread function (PSF) of a beamformer can be an ideal metric, although the PSF may be spatially dependent: every pixel can exhibit a slightly different point-spread.

[0134] FIGS. 5A-C are plots illustrating point spread functions (PSFs) 530a-c of a delaysum-integrate (DSI) beamformer 530a, a Robust Capon Beamformer (RCB) 530b, and an example embodiment of a beamformer 530c, respectively. Lateral full-width at halfmaximum (FWHM) maps are shown in millimeters with respect to an axial position and a lateral position across the field of view of a receiver, e.g., the receiver described herein with reference to FIG. 4.

[0135] The full-width at half-maximum (FWHM) can be used herein as the measure of beam profile width in both the axial and lateral directions.

[0136] The mean-square intensity (MSI) of the pixels can serve as a measure of the amount of relative noise or cumulative artifact presence in an image. MSI can encompass both signal and artifacts; however, since the size of cavitating bubbles can be on the micrometer scale, an order of magnitude smaller than the millimeter scale used in ultrasound imaging applications, the cavitation source can be point-like and much smaller than a single pixel. For the figures included herein, a single, stationary cavitation source (such as those simulated by Vokurka’s model) could produce minimal signal intensity, while the overwhelming majority of observed pixel intensity can come from artifacts and the pointspread function. The beamformers used in this work may not be energy-preserving, so MSI can be used to quantify artifacts and the effects of point-spread. It is calculated by:- 27 - 4239812. vl5200.2434002(16)Where Nx, Nyare the dimensions of the image in pixels, and Pij is the pseudo-intensity at pixel i,j. As used herein, MSI may be useful for approximating the amount of noise in a beamformed image, comparing beamformers for differently sized bubble clusters.

[0137] RESULTS

[0138] Localization, Speed, and Artifact Reduction

[0139] To compare embodiments, e.g., PAD AM, to RCB and DSI, the Vokurka model can be used with a single cavitation source positioned axially at 40 mm and centered laterally in a virtual probe’s field of view. The images can be beamformed, and the output images can be normalized on a linear scale for comparison, as PAD AM isTABLE I: Beam Widths (mm)DSI PADAMAxial Lateral Axial Lateral-3 -IB 2,58 Q216 L02 0.072-6 dB 9.36 0.744 1.98 0.12-10 d. B 15.06 132 3.42 0.312not a power-based beamformer and the intensity of the images may not be compared directly.

[0140] FIGS. 6A-C are plots illustrating in-silico passive cavitation images 630a-c generated using a Vokurka model and beamformed using a Delay-Sum-Integrate (DSI) beamformer 630a, a Robust Capon Beamformer (RCB) 630b, and an example embodiment of a beamformer 630c, respectively. The image 630c may be generated using the example embodiment with parameter m = 1. Each image 630a-c of FIGS. 6A-C is normalized between 0 and 1, with pixel intensities representing pseudo-power. The images 630a-c show detected intensities along an axial position and a lateral position of a view of a receiver, e.g., the receiver 410 described herein with reference to FIG. 4.

[0141] As indicated by the image 630c, embodiments, e.g., PAD AM, can exhibit a point-like source stretched by the axial resolution of the system, without a tail artifact, whereas the image 630b generated using RCB can display a much larger area without a tail, and the image 630a generated using DSI can show the classical X-shaped tail artifact.- 28 - 4239812. vl5200.2434002

[0142] FIG. 6D is a plot 630d of axial beamwidth profiles of the passive cavitation images of FIGS. 6A-C. The plot 630d indicates intensity along an axial position of images formed using DSI 632a, RCB 634a, and the example embodiment 636a.

[0143] FIG. 6E is a plot 630e of lateral beamwidth profiles of the passive cavitation images of FIGS. 6A-C. The plot 630e indicates intensity along an axial position of images formed using DSI 632b, RCB 634b, and the example embodiment 636b.

[0144] The plots 630d-e of FIGS. 6D and 6E, respectively, may be helpful in visualizing beam widths along axial and lateral dimensions for each of the three beamforming techniques used in FIGS. 6A-C. Further, the plots 630d-e of FIGS. 6D and 6E can be generated using a linear array at a point-source depth of 130 mm. Results regarding the beamwidths of FIGS.6D and 6E are summarized in Table I.

[0145] To evaluate computational performance of an example embodiment, e.g., PAD AM, beamforming was performed in MATLAB on a 100 * 100 pixel image with 2,000 timepoints using an Intel i9-13900K CPU. The average computation times were 2.15 seconds for DSI, 6.33 seconds for RCB, and 5.69 seconds for PAD AM. However, PAD AM can offer a significant advantage over RCB when testing multiple parameters: the eigenvalue decomposition can be re-used for each parameter desired and simply indexed, whereas RCB may employ both the Lagrange multiplier and Newton’s methods to re-solve the optimization problem for a new a. The times for beamforming a set of images for several numbers of parameters are listed in Table II.

[0146] The Vokurka model can further be used to quantify resolvability. Two bubble sources, initially co-located, can be incrementally separated laterally, and a one dimensional lateral profile can be extracted at the depth of the sources for each separation distance. These line-images were concatenated to visualize how resolution changes with increasing source separation. (Note: in FIG. 3, each line-image was individually normalized from 0 to 1 to highlight relative contrast.) To mitigate stochastic variability inherent in the Vokurka model, 30 line-images were averaged at each spacing.

[0147] FIG. 7 is a plot illustrating in silico passive cavitation images 730a-c of pointsource resolvability in images generated using a DSI 730a beamformer, an RCB 730b, and an example embodiment of a beamformer 730c. The image 730c of the example embodiment may be generated using parameter / ?? = 1. Each column of the panels (the images 730a-c) is a lateral line-image taken at a focus-depth of two point sources which step away from each other from left to right, indicated as the bubble spacing. While the RCB beamformer appears- 29 - 4239812. vl5200.2434002to produce sharper profiles at larger separations, the RCB image 730b exhibits a persistent central artifact even when sources are well resolved. A similar artifact is observed in the image 730a DSI. Using the definition of resolution as the minimum distance at which two point sources remain distinguishable, the observed resolutions are approximately 1 mm for the image 730a of DSI, 0.65 mm for the image 730b of RCB, and 0.5 mm for the image 730c of the embodiment based on a visual estimate. These results may highlight the improved spatial discrimination of embodiments described herein, particularly in closely spaced cavitation scenarios.

[0148] To simulate a bubble cloud within the therapeutic focal region, Vokurka-modeled bubbles can be randomly distributed within an ellipsoidal volume approximating the FUS focal zone. These clustered bubble distributions can serve as ideal test cases for evaluating artifact removal, as tail artifacts between adjacent bubbles can overlap and create misleading signals in regions devoid of actual bubbles.

[0149] FIG. 8A is a plot illustrating representative cavitation images 830a-c of a 5-source simulation generated using a DSI beamformer 830a, an RCB 830b, and an example embodiment of a beamformer 830c, respectively. The images 830a-c show relative intensities across axial positions and lateral positions of a simulated field of view. True source locations are overlaid in each of the beamformed images as X’s,.e.g, the X 838. The image 830c generated using the example embodiment uses parameter m = 3.

[0150] FIG. 8B is a plot 830d of mean-squared intensities (MSIs) indicative of artifact reduction ability of the beamformers of FIG. 8 A at different bubble cluster sizes. Bubble clusters of size 10, 20, 50, and 80 are simulated and correspondingly plotted. The cluster size may also be referred to as cloud size.

[0151] As indicated in FIGS. 8A and 8B, embodiments may demonstrate significantly lower MSI values compared to DSI and RCB, which may highlight superior performance of embodiments in cavitation localization and tail artifact suppression. Despite a thorough parameter sweep aimed at optimizing RCB performance, RCB consistently underperformed relative to embodiments. The elliptical output patterns commonly observed in RCB reconstructions appear to stem from the fixed 8 coefficient, which may artificially constrain the steering vector and limit spatial accuracy. FIGS. 8A and 8B may suggest that embodiments, e.g., PAD AM, are highly effective at reducing noise and enhancing spatial localization in scenarios involving clustered cavitation.TABLE II: Parameter Beamforming Time (seconds)- 30 - 4239812. vl5200.2434002RCB ( ) PAQAM 6 )1 parameter 6.33 5.6920 parameters 22.04 8.5250 parameters 43.41 13.5

[0152] The PAD AM Parameter mEmbodiments as described herein may operate by assuming a number of sources m from which the signal and noise subspaces can be determined using accepted and rejected eigenvectors. Selecting a number of sources, and therefore a number of eigenvalues, can mean that at each pixel, only the eigenvectors corresponding to the m largest eigenvalues are included. The largest m eigenvectors may correspond to the most correlated sources in the spatial covariance matrix at that pixel. Therefore, selecting the m largest eigenvectors may effectively filter out uncorrelated signals. Additional eigenvalues in the spatial covariance matrix may represent “weaker” sources at other locations with lower correlation, while nearzero eigenvalues primarily correspond to noise. According to some embodiments, the exact number of nonzero eigenvalues present in a signal’s spatial covariance matrix may be two times the number of frequencies in that signal (W. D. Penny, “Signal Processing Course,” www.fil.ion.ucl.ac.uk / ~wpenny / , 2009).

[0153] Based upon this relationship between m and the number of frequencies in a signal, increasing m to larger integers may introduce a leakage signal from other sources, as well as potential noise. This relationship may be further described hereinbelow with respect to FIGS.5A-5D, wherein an image may be beamformed using an embodiment, e.g., P D AM, with a sinewave source with amplitude of 1 MPa and frequency f = 3.21 MHz on the left, and a Vokurka source with a mean peak amplitude of 1 MPa on the right.

[0154] FIGS. 9A-C are plots illustrating beamformed images 930a-c of a pure tone source and a Vokurka source using parameter m of 2 (930a), 3 (930b), and 4 (930c), respectively, according to an example embodiment. The images 930a-c show recovered intensities across axial positions and lateral positions of a simulated field of view.

[0155] FIG. 9D is a plot 930d of frequency spectra (power spectral density versus frequency) of the pure tone source and the Vokurka source.

[0156] FIGS. 9E and 9F are plots 930e-f, respectively, of eigenvalue distribution of the pure tone source and the Vokurka source, respectively, of FIGS. 9A-C. As illustrated in plot 930f, the Vokurka source’s eigenvalue pattern may include 16 eigenvalues that represent the- 31 - 4239812. vl5200.24340028 in-band harmonics of the drive frequency, which was 500 kHz, which is illustrated in the plot 930d).

[0157] Sinewave signals may generate a spatial covariance matrix that has exactly two non-zero eigenvalues, as illustrated in the plot 93 Oe, which can be strong compared to the many non-zero eigenvalues of the Vokurka source, which is illustrated in the plot 930f).

[0158] At m = 2 (image 930a), only the sine-wave may be visible in the entire image, as the sine wave is highly correlated and may dominate the signal, even at the Vokurka source’s pixel location. However, at m = 3 (image 930b), more eigenvalues are included than the sinewave’s alone, which may allow one eigenvalue from the Vokurka source. The Vokurka source appears at approximately half the intensity of the sinewave source, which may be attributed to two eigenvectors belonging to the sinewave and only one from the Vokurka source. At m = 4 (image 930c), both regions appear similar in intensity. Beyond m = 4, more eigenvectors correspond to the Vokurka source, which shift the visual weight towards the right. Notably, the eigenvectors in embodiments, e.g., PAD AM, may not be scaled by their eigenvalues, which may lead to a stacking effect. Admitting too many sources can result in overlaying eigenvectors from closely located sources, causing 2 - 3 pixels to dominate the image for a single source. As such, admitting too many sources can misidentify the strongest source’s location due to factors like the tail artifact if m is set too high.

[0159] A synthetic frequency component separation test can be performed using Vokurka-modeled sources to illustrate the ability of embodiments, e.g., PAD AM, to distinguish cavitation types. According to an example embodiment, five sources can be arranged in a rectangular configuration, with one located at the center. The time-conditioning parameter (“9”) can be increased for the source positioned at the center, which may result in longer pulse widths and a different harmonic-to-inharmonic energy ratio, serving as a proxy for stable cavitation, while the outer sources with low 9 can model inertial cavitation.

[0160] FIG. 10A is a plot illustrating images 1030a-c, reconstructed using example embodiments of beamformers with parameters m = 1 (1030a), m = 8 (1030b), and m = 20 (1030c), respectively, of a domain including inertial and stable cavitation sources. Each panel (the images 1030a-c) is normalized. The images 1030a-c show reconstructed intensities across axial positions and lateral positions of a simulated field of view. As described hereinabove, the central source of each of the images 1030a-c models a stable cavitation source and the outer sources of each of the images 1030a-c model inertial cavitation sources.- 32 - 4239812. vl5200.2434002

[0161] FIG. 10B is a plot 1030d of reconstructed contrast values of a stable source and an inertial source generated using example embodiments sweeping across parameter / ??, including the parameter m values of 1, 8, and 20 of FIG. 10 A.

[0162] FIG. 10C is a plot 1030e of frequency spectra (power spectral density versus frequency) of a simulated stable source and a simulated inertial source. With respect to the plot 1030d FIG. 10B and the plot 1030e of FIG. 10C, the stable source corresponds to the “Center” source and the inertial source corresponds to the “Top Left” source as described hereinabove with reference the images 1030a-c of FIG. 10 A.

[0163] It may be noted that the contrast of the outer sources slightly increases with increasing / ??, as seen in the plot 1030d of FIG. 10B, but the change in contrast of the outer sources is overshadowed by that of the center source.

[0164] As shown in the plot 1030e of FIG. 10C, three dominant peaks may be associated with the inertial sources, while the remaining peaks may correspond to the stable source. Consequently, the leading eigenvalues in the image may primarily originate from the inertial sources until m = 8 and 9, where contributions from the stable source may begin to emerge. The contrast of the sources as shown in FIG. 10B may highlight this effect.

[0165] With respect to the plot 1030d of FIG. 10B and the plot 1030e of FIG. 10C, at low m, the stable cavitation signal can be suppressed due to the dominant low-frequency components from the inertial sources, as reflected in the contrast values of the plot 1030d and frequency spectra of the plot 1030e. The strongest spectral peaks in the plot 1030e originate from the peripheral inertial sources. At higher values of m (8 and 20), the stable cavitation source’s higher-frequency components are incorporated via eigen-decomposition, resulting in increased contrast in the reconstructed image. Contrast can be calculated asIImmaaxx+IImmiinn wherein I represents intensity in a given image and Imax and / min are the highest and low intensity values in the given image..

[0166] According to an embodiment, the aforementioned finding, which may include an ability to dynamically tune reconstructed contrast of sources based on the parameter m (number of eigenvalues / eigenvectors), suggests that embodiments, e.g., PAD AM, may enable differentiation between stable and inertial cavitation within an image, a capability further explored in the in vitro experiments described hereinbelow.

[0167] FIG. 11 A is a plot illustrating images 1130a-c, reconstructed using example embodiments of beamformers with parameters m = 1 (1130a), m = 8 (1130b), and m = 20- 33 - 4239812. vl5200.2434002(1130c), respectively, of a domain including inertial and stable cavitation sources, wherein an amplitude of the inertial sources are set to a factor of two times an amplitude of the stable sources. The images 1130a-c show reconstructed intensities across axial positions and lateral positions of a simulated field of view

[0168] FIG. 1 IB is a plot 1130d of reconstructed contrast values of a stable source and an inertial source generated using example embodiments sweeping across parameter m, including the parameter m values of 1, 8, and 20 of FIG. 11 A. Similar to FIG. 10B, the contrast can be calculated asImax Imin.Imax+Imin

[0169] FIG. 11C is a plot 1130e of frequency spectra (power spectral density versus frequency) of a simulated stable source and a simulated inertial source.

[0170] FIG. 12A is a plot illustrating images 1230a-c, reconstructed using example embodiments of beamformers with parameters m = 1 (1230a), m = 8 (1230b), and m = 20 (1230c), respectively, of a domain including inertial and stable cavitation sources, wherein an amplitude of the inertial sources are set to a factor of four times an amplitude of the stable sources. The images 1230a-c show reconstructed intensities across axial positions and lateral positions of a simulated field of view.

[0171] FIG. 12B is a plot 1230d of reconstructed contrast values of a stable source and an inertial source generated using example embodiments sweeping across parameter m, including the parameter m values of 1, 8, and 20 of FIG. 12 A. Similar to FIG. 10B, the contrast can be calculated asImax Imin.Imax+Imin

[0172] FIG. 12C is a plot 1230e of frequency spectra (power spectral density versus frequency) of a simulated stable source and a simulated inertial source.

[0173] In Vitro Models

[0174] In vitro experiments may be used to further evaluate the ability of embodiments presented herein, e.g., PAD AM, to provide superior artifact reduction and harmonic differentiation, as seen in the computational models.

[0175] FIG. 13 illustrate schematically an in vitro experimental setup 1300 for evaluating methods of monitoring ultrasound therapy, the in vitro experimental setup including double tube phantoms 1342, 1344. The experimental setup 1300 includes a focused ultrasound (FUS) transducer 1302 configured to deliver a focused ultrasound stimulus 1304 to a first tube 1342 and a second tube 1344. The first tube 1342 and the second tube 1344 may be fluidically coupled to syringes 1346, 1348. The syringes 1346, 1348 may include microbubbles 1320,- 34 - 4239812. vl5200.2434002e.g., lipid microbubbles, and may be configured to cause the microbubbles 1320 to flow through the first tube 1342 and the second tube 1344. An imaging probe 1310 may be positioned to detect a signal 1308 that may be associated with stable cavitation effects 1350 and inertial cavitation effects 1352 of the first tube 1342 and the second tube 1344, respectively, resulting from excitation by the focused ultrasound stimulus 1304 from the FUS transducer 1302.

[0176] An imaging plane (not illustrated) of the imaging probe 1310 may intersect a cross-section of the two tubes 1342, 1344 containing microbubbles 1320. The FUS transducer 1302, which may be a 500-kHz FUS transducer, can deliver pulses (FUS stimulus 1304) to induce cavitation, while the imaging probe 1310 passively receives the resulting acoustic emissions ( / .<?., the stable cavitation effect 1350 and the inertial cavitation effect 1352).According to an example embodiment, the two tubes 1342, 1344 can be fixed in place by a plastic device, e.g., a plastic phantom holder 1356, with holes drilled specifically to place the tubes 1342, 1344 in the center of a focus 1354 of the focused ultrasound and off-center from the focus 1354 so that inertial cavitation may happen primarily in the near (bottom) tube 1342, and stable cavitation happens primarily in the far (top) tube 1344.

[0177] FIG. 14A is a plot illustrating cavitation images 1430a-e, acquired using the experimental setup of FIG. 13, beamformed using a DSI beamformer (1430a), an RCB using parameter a = 10 (1430b), and example embodiments of beamformers with parameter m = 1 (1430c), m = 8 (1430d), and m =40 (1430e). The images 1430a-e of FIG. 14A may be generated using the aforementioned beamformers and overlaid on a B-mode image for localization The images 1430a-e may further show reconstructed intensities at axial positions and lateral positions within a field of view of a receiver, for example, the receiver 1310 of FIG. 13. Cross-sections of tubes, e.g., the tubes 1442, 1444, may be visible in the B-mode image (indicated by dotted white circles). The left tube 1442 includes inertial cavitation effects (which may be generated using a FUS target pressure of 1.5 MPa) and the right tube includes stable cavitation effects.

[0178] Compared to RCB (image 1430b) and DSI (image 1430a), embodiments (images 1430c-e) correctly isolate the inertial cavitation to the near-tube at low m (image 1430c) and also reveals the stable cavitation in the far-tube at higher m values starting at m = 8 (image 1430d). This inclusion of stable cavitation can further be seen when more frequency components are included at m = 40 (image 1430e). The images 1430c-e may provide further evidence that embodiments can isolate inertial cavitation regions from stable ones by- 35 - 4239812. vl5200.2434002selecting low m, and that increasing m can find stable cavitation even if it is at a much lower pressure amplitude.

[0179] FIG. 14B is a plot 1430f of sizes of -3dB and -6dB regions of a main lobe of the cavitation images of FIG. 14 A. As indicated by the plot, embodiments (e.g., PAD AM) using parameter m = 1 may achieve the strongest localization, as indicated by the smallest area size of the -3dB and -6dB lobes.

[0180] FIG. 14C is a plot 1430g of example spectra (power spectral density versus frequency) of an inertial cavitation source 1432, stable cavitation source 1434, and a combination 1436 of the inertial cavitation source and the stable cavitation source acquired using the experimental setup of FIG. 13. The plot 1430g may provide confirmation of inertial and stable cavitation behaviors by passive cavitation detection.

[0181] FIG. 15 is a plot illustrating, in a normalized linear scale, cavitation images 1530a-e corresponding to the cavitation images 1430a-e of FIG. 14 A.

[0182] FIG. 16 illustrates beamformed images 1630a-h of the experimental setup of FIG.10 generated using an RCB. The images 1630a-h are reconstructed using RCB parameter a = 0.001 (1630a), 0.01 (1630b), 0.1 (1630c), 1 (1630d), 2 (1630e), 8 (1630f), 10 (1630g), 20 (163 Oh). All images are scaled from 0 to -20 dB. The images 1630a-e may show reconstructed intensities at axial positions and lateral positions within a field of view of a receiver, for example, the receiver 1310 of FIG. 13.

[0183] FIG. 17 illustrates beamformed images 1730a-l of the experimental setup of FIG.10 generated using example embodiments of a beamformer, e.g., PAD AM. The images 1730a-l are reconstructed using parameter m = 1 (1730a), 3 (1730b), 5 (1730c), 10 (1730d), 12 (1730e), 15 (1730f), 20 (1730g), 30 (1730h), 40 (1730i), 60 (1730j), 80 (1730k), 100 (1730l). All images are scaled from 0 to -20 dB. The images 1730a-e may show reconstructed intensities at axial positions and lateral positions within a field of view of a receiver, for example, the receiver 1310 of FIG. 13.

[0184] Embodiments of methods for monitoring ultrasound therapy may be further evaluated in the presence of skull-induced aberration using an in vitro model comprising a skull.

[0185] FIG. 18 illustrates an in vitro experimental setup 1800 for evaluating methods of monitoring ultrasound therapy in a rat phantom 1860. The rat phantom 1860 includes a thin polytetrafluoroethylene (PTFE) tube (not visible) routed through a cranial cavity of a rat skull and the rat skull is submerged in water. The experimental setup includes an ultrasound- 36 - 4239812. vl5200.2434002transducer 1802 configured to deliver an FUS stimulus 1804, e.g., FUS pulses, from the top down. The experimental setup 1800 further includes an imaging probe 1810 configured to record signals, e.g, cavitation signals, from the rat phantom along an imaging plane 1862.

[0186] FIG. 19A is a plot illustrating images acquired of the experimental setup of FIG.18, including a B-mode ultrasound image (1930a) and beamformed images using a DSI beamformer (1930b), an RCB (1930c), and example embodiments of beamformers with parameter m = 1 (1930d) and m= 10 (1930e). The images 1930a-e may show reconstructed intensities at axial positions and lateral positions within a field of view of a receiver, for example, the receiver 1810 of FIG. 18. In the image 1930a, the rat skull 1964 may be visible in white at a depth (axial position) of 29 mm. The embedded tube of the experimental setup 1800 of FIG. 18 may be visible at a depth (axial position) of 35 mm and is indicated by the arrow in the image 1930a. As illustrated in FIG. 19 A, embodiments, as shown in images 1930d-e can demonstrate superior localization and improvement over the image 1930c generated using RCB, particularly in isolating the off-target cavitation lobe observed near (8, 30) mm at the skull base in the image 1930d.

[0187] FIG. 19B is a plot 1930f of frequency spectra (power spectral density versus frequency) of radiofrequency (RF) data at three look points in the image 1930e of FIG. 19 A, with an inset 1964 indicating the positions of the look points. The look points include an inertial cavitation (IC) look point 1966, a stable cavitation (SC) look point 1968, and a no bubble area 1970. The plot 1930f includes a corresponding frequency spectrum 1932 of the IC look point, frequency spectrum 1934 of the SC look point, and frequency spectrum 1936 of the no bubble look point

[0188] FIG. 19C is a plot 1930g quantifying harmonic, ultraharmonic, and inharmonic components at the look points of FIG. 19B. The plot may confirm that stable cavitation (SC) versus inertial cavitation (IC) can be dynamically distinguished in images generated by embodiments, e.g., PAD AM, which may be indicated by the IC / SC ratio

[0189] As illustrated by FIG. 19 A, all three beamformers localized the cavitation source near the true position of the tube observed in B-mode imaging, as indicated by the arrow. The images 1930d, 1930e generated using embodiments demonstrated markedly improved localization compared to the image 1930b generated using DSI, particularly at low values of m. The image 1930c generated using RCB was able to effectively suppress most of the tail artifacts; however, the embodiments may be more effective, which may allow for easier identification of an off-target cavitation source at the skull base (30.5 mm axially, e.g., point- 37 - 4239812. vl5200.24340021972 of the image 1930d) in addition to the microbubble cavitation in the tube (at 36.6 mm axially, e.g., the point 1976 of the image 1930d). Increasing m to 10 widens the image lobes, which may suggest stable cavitation to the right of the focus, as seen in the image 1930e.

[0190] Presence of stable cavitation may be further investigated by selecting, with respect to FIGS. 19A and 19B, three look-points from 1) the main lobe 1966 of the image 1930e, 2) a side-lobe area 1968 of the image 1930e, and 3) the periphery with no bubbles 1970 of the image 1930e as a control. The beamformed frequency spectrum can be calculated by applying delay-and-sum across channels of an imaging probe, e.g., the imaging probe 1810 of FIG. 18, for the time RF-data delayed for each location and taking the Fourier Transform.

[0191] Harmonic, inharmonic, and broadband components of the spectrums of FIG. 19B can be isolated and used to calculate a ratio of inertial-cavitation to stable cavitation power as:in the plot 1930g of FIG. 19C, which shows a low IC / SC ratio for the stable look-point and high ratio for the IC look-point, confirming that mostly stable cavitation occurs at the rightlobe look point. As such, embodiments, e.g., PAD AM, may be capable of localizing stable cavitation when comparing to a low m image in the presence of the skull barrier and with aberration effects by increasing m, which may suggest that the embodiments can perform classification well in-vivo.

[0192] FIG. 20 is a plot illustrating, in a normalized linear scale, cavitation images 2030a-d corresponding to the cavitation images 1930b-e of FIG. 19 A.

[0193] FIG. 21 illustrates beamformed images 2130a-l of the experimental setup of FIG.18 generated using an RCB. The images 2130a-l are reconstructed using RCB parameter a = 0.0001 (2130a), 0.001 (2130b), 0.01 (2130c), 0.1 (2130d), 0.5 (2130e), 1 (2130f), 2 (2130g), 5 (2130h), 10 (2130i), 20 (2130j), 30 (2130k), 50 (21301). All images are scaled from 0 to -10 dB. The images 1930a-e may show reconstructed intensities at axial positions and lateral positions within a field of view of a receiver, for example, the receiver 1810 of FIG. 18.

[0194] FIG. 22 illustrates beamformed images 2230a-l of the experimental setup of FIG.18 generated using example embodiments of a beamformer, e.g., PAD AM. The images- 38 - 4239812. vl5200.24340022230a-l are reconstructed using parameter m = 1 (2230a), 2 (2230b), 3 (2230c), 5 (2230d), 8 (2230e), 10 (2230f), 12 (2230g), 15 (2230h), 20 (2230i), 30 (2230j), 34 (2230k), and 38 (2230l). All images are scaled from 0 to -10 dB. The images 1930a-e may show reconstructed intensities at axial positions and lateral positions within a field of view of a receiver, for example, the receiver 1810 of FIG. 18.

[0195] DISCUSSION

[0196] As disclosed herein, embodiments of methods for monitoring ultrasound therapy, for example, PAD AM, in the time domain can be a highly effective approach for passive cavitation localization. The embodiments consistently outperformed DSI and RCB in both lateral and axial resolution, as well as point-spread function size for the observed areas in the imaging plane. Furthermore, PAD AM exhibited superior tail and reflection artifact reduction compared to RCB for both singular sources and clusters, both in-silico and in-vitro, with and without a skull.

[0197] A key advantage of embodiments such as PAD AM may include the parameter / ??, which can reflect the frequency richness of incoming signals. Although embodiments such as PAD AM operate in the time domain, this parameter can enable spatial filtering based on frequency content. As a result, embodiments can offer a physically meaningful and intuitive avenue to distinguish between different frequency spectra and, consequently, between cavitation mechanisms.

[0198] As discussed in the hereinabove, stable and inertial cavitation can induce distinct bio-effects: stable cavitation can facilitate targeted vascular permeation, whereas inertial cavitation can enable tissue ablation. By adjusting / ??, users can localize sources or clusters ( / .<?., setting / ?? to 1) to find inertial-like cavitation sources, or increase m to visualize stable cavitation regions, for example, if ultraharmonics are present. The embodiments’ ability to dynamically distinguish cavitation mechanisms may have significant implications in image-guided FUS therapy, particularly for applications requiring selective monitoring of stable versus inertial cavitation. This monitoring of stable versus inertial cavitation may be useful when avoiding unintended bio-effects from inertially cavitating bubbles in sensitive areas or in off-target zones frequency seen in-vivo.

[0199] Compared to RCB, embodiments disclosed herein can be easier to use and more computationally efficient. Unlike RCB, which requires parameter tuning ( / .<?., a) to optimize performance, which may often result in inconsistent image quality, the embodiments, e.g., PAD AM, may not require parameter searches, making it more user-friendly. Additionally,- 39 - 4239812. vl5200.2434002the embodiments may be computationally faster than RCB, as they may rely on eigenvalue decompositions rather than matrix inversions for each pixel. Furthermore, the embodiments’ flexible handling of signal and noise subspaces can allow for rapid computation of multiple images with varying m values, rather than rerunning the method separately for each case. An example embodiment of a scenario of selecting the parameter / ?? can involve selecting 10-15 test values ranging from low (1) to high (40), and noting at which / ?? the cavitation footprint expands significantly, which may reveal quieter sources with wider frequency-spectra. The embodiments may also be a candidate for parallelization and speedup by using methods that identify the largest several eigenvalues (namely, the power method) instead of every eigenvalue.

[0200] However, a key limitation of the embodiments disclosed herein may include that the embodiments are not a power-based beamformer, and should not be used by itself to monitor cumulative energy absorption. The embodiments may assign pixel values inversely proportional to noise correlation, which may essentially measure what is not noise.

[0201] Future work could explore using images generated using the embodiments, e.g., PAD AM, as masks over a power-based beamformer to generate comprehensive dosage maps. Additionally, some embodiments may currently require manual selection of the parameter / ??; future work will focus on developing data-driven or biologically informed strategies for adaptive selection of this parameter. Third, while other methods such as Delay-Multiply-and-Sum (DMAS) and sparse reconstruction may be relevant in broader ultrasound contexts, the other methods were not directly compared here. DMAS can offer limited resolution improvements in PCI (compared to DSI) and has not seen widespread adoption, while sparse reconstruction can be more applicable to compressed sensing or accelerated imaging — beyond the scope of the current fully sampled study. Nevertheless, these methods may hold promise for extending the capabilities of embodiments disclosed herein and may be explored future works aimed at real-time or resource-constrained applications. Lastly, extending embodiments to the frequency domain would enable direct frequency selection as it does with DSI, allowing for isolation of the ultraharmonic and harmonic bands associated with stable cavitation.

[0202] Several experimental limitations may also be present with respect to the in-silico and in-vitro methods. The Vokurka model, while useful, may assume a smooth ramp-down of the impulse pressure, which oversimplifies actual cavitation behavior. The experiments described hereinabove also varied the Vokurka 6 time-condition parameter to mimic bubble- 40 - 4239812. vl5200.2434002dynamics. While altering 0 modified the frequency spectrum and the ratio of harmonic to inharmonic energy, altering 0 did not generate ultraharmonics that may commonly be observed in cavitation behavior. In addition, the inharmonic content in the Vokurka model may mainly arise from aliasing rather than a well-defined inertial cavitation mechanism, which may limit the Vokurka model’s ability to fully replicate stable and inertial cavitation. However, the fundamental principle underlying the embodiments disclosed herein, differentiating sources dynamically based on frequency content, remains effectively demonstrated and was further validated by the in-vitro results.

[0203] In the phantom studies, the double-tube experimental setup was designed to capture both inertial cavitation in the near-tube and stable cavitation far-tube simultaneously. Cavitation effects may also be investigated at lower concentrations of microbubbles in the tubes to track individual bubble dynamics over time as they migrate into a FUS focal region. Another limitation of the current in vitro study rat skull experiment may include the skull model thickness. The embodiments performed well in the presence of a thin rat skull at a low drive frequency, showing little focal aberration consistently with the other beamformers. However, higher-frequency components such as harmonics and ultraharmonics above 2 MHz may include wavelengths comparable to skull thickness, and thus may experience greater distortion. Future studies may examine the performance of embodiments, e.g., P D AM, through thicker and more acoustically complex barriers, such as primate skulls. In vivo studies will also be critical for assessing the embodiments’ ability to differentiate cavitation regimes in a more realistic biological environment.

[0204] While embodiments disclosed herein may not compensate for skull-induced focal shifts, experimental results may suggest that the embodiments remain robust in the presence of absorption, noise, and structural inhomogeneities — conditions that often degrade performance in conventional beamformers.

[0205] CONCLUSION

[0206] Embodiments of a method for monitoring ultrasound therapy, e.g., PAD AM, are disclosed herein as a time-domain PCI method. Performance of the embodiments are compared with respect to established beamformers such as DSI and RCB. that the embodiments can offer enhanced resolution, robust artifact suppression, and a physically interpretable and intuitive input parameter that can isolate stable and inertial cavitation. The embodiments can also achieve improved computational efficiency over RCB by eliminating the need for per-pixel matrix inversions. Most notably, the embodiments’ ability to- 41 - 4239812. vl5200.2434002differentiate cavitation mechanisms through its parameter m may represents a meaningful advancement in PCI. This dual capability, not only localizing cavitation but also characterizing its nature dynamically, can hold significant promise for guiding and controlling ultrasound therapies.

[0207] According to an example embodiment, an experimental setup (not shown) for evaluating methods for monitoring ultrasound therapy can include two flow tubes containing bubbles, such as microbubbles or nanobubbles for non-limiting examples. The experimental setup can further include a focused ultrasound emitter configured to deliver a focused ultrasound stimulus (e.g., sonication) with a FUS focal zone overlapping the two flow tubes. A first tube of the two flow tubes can be a near tube positioned directly in the focus and the FUS stimulus may cause inertial cavitation effects in the first tube. A second tube of the two flow tubes can be a far tube disposed, for non-limiting example, 7 mm away from the first tube and the FUS stimulus may cause stable cavitation effects in the second tube.

[0208] In addition, cavitation effects in the two flow tubes, which may be induced by the FUS stimulus, may be detected using a passive cavitation detector (PCD) or an ultrasound probe. In an embodiment, a passive cavitation detector can be single-element hydrophone with a narrow-band frequency response that can be targeted at a frequency range of expected cavitation. In some embodiments, the PCD can be located at a the center of a bowl of an FUS transducer or emitter, and the resonant frequency can be around 750 kHz (which can be the first ultraharmonic of a 500KHz FUS), for nonlimiting example. The PCD data may be included because it can provide independent evidence of stable versus inertial cavitation behavior. The PCD signal may be useful for validating the cavitation regime, which in turn supports classifications generated by example embodiments of beamformers described herein, e.g., PAD AM. In some embodiments, the ultrasound probe can include a conventional diagnostic ultrasound probe. The ultrasound probe can be sensitive to a wider frequency response in the higher range but may be unable see the low-end frequency response between the fundamental and first few harmonics. Both types of probes may be relevant in characterizing cavitation behavior - in tissue, higher frequencies can have a harder time propagating over long distances (or through barriers like the skull), whereas lower frequencies can go much further. On the other hand, higher frequencies can allow higher spatial resolution beamforming.

[0209] FIG. 23A is a plot illustrating a spectrogram 2330a of a focused ultrasound (FUS) passive cavitation detector (PCD) signal of cavitation effects induced in a two-tube- 42 - 4239812. vl5200.2434002experimental setup using an FUS stimulus, according to an example embodiment. The spectrogram 2330a illustrates spectral components of the passive cavitation detector signal over time.

[0210] FIG. 23B is a plot illustrating a spectrogram 2330b of ultrasound probe channel data of cavitation effects induced in a two-tube experimental setup using an FUS stimulus, according to an example embodiment. The spectrogram 2330b illustrates spectral components of data from a central channel of an ultrasound probe over time.

[0211] As indicated in FIGS. 23A and 23B, when the FUS stimulus is activated, inertial cavitation can be observed coming from the near tube, and stable cavitation can be observed from the far tube. Across the 5 ms in the burst (e.g., the FUS stimulus), microbubbles in the near tube may collapse before they can be replenished and, thus, a broadband intensity of the signal may decay around 3 ms into stable cavitation (as may be evident in the spectrogram 2330b).

[0212] Images generated by an example embodiment of a beamformer, e.g., PAD AM, may be useful for dynamically classifying and localizing the cavitation effects such as disclosed with reference to FIGS. 23C and 23D below.

[0213] FIGS. 23C and 23D are plots illustrating images 2330c, 2330d formed from ultrasound probe data, including the ultrasound probe channel data of FIG. 23B. The images 2330c, 2330d are beamformed using an example embodiment of a beamformer with parameters of m between 1 and 60 for a tube including inertial cavitation effects 2330c and a tube including stable cavitation effects 2330d. The images 2330c, 2330d may be used as a mask over images formed using a DSI beamformer. As described hereinabove, the tube including the inertial cavitation effects can be the near tube and the tube including the stable cavitation effects can be the far (distal) tube of the two tube experimental setup. An amount of power in a small region surrounding each tube can be measured for parameters of m between 1 and 60. The image 2330c may indicate that low values for m result in inertial cavitation in the near tube. The image 2330d may indicate that a value of 20 for m may maximize isolation of stable cavitation for the far tube during an early portion of a burst or stimulus.

[0214] FIG. 23E is a plot 2330e illustrating a best m parameter for detecting inertial cavitation and stable cavitation for the tubes including the inertial cavitation effects and the stable cavitation effects of FIGS. 23C and 23D.- 43 - 4239812. vl5200.2434002

[0215] FIG. 23F is a plot 2330f of power of harmonic, ultraharmonic, and broadband frequencies in the PCD signal of FIG. 23 A. As microbubbles become exhausted, less inertial cavitation may occur in the near tube, and there may be less separation of inertial and stable cavitation effects. The parameter m may follow the spectral patterns of the PCD signal, as shown in the plot 2330f.

[0216] FIGS. 23G and 23H are plots illustrating images 2330g, 2330h beamformed using an example embodiment of a beamformer with parameter m = 1 (the image 2330g) and m = 20 (the image 2330h), respectively. The images 2330g, 2330h are beamformed using data at a time point of t= 1 ms of ultrasound probe data including the ultrasound probe channel data of FIG. 23B. At t = 1 ms, a clear separation of stable and inertial effects may be evident between the tubes (the tubes including inertial and stable cavitation effects) for the image 2330g with a low value of m and the image 2330h with a high value of m, which may be due to both tubes cavitating stably.

[0217] FIGS. 231 and 23J are plots illustrating images 233 Oi, 2330j beamformed using an example embodiment of a beamformer with parameter m = 1 (the image 233 Oi) and m = 20 (the image 2330j), respectively. The images 2330i, 2330j are beamformed using data at a time point of t= 4 ms of ultrasound probe data including the ultrasound probe channel data of FIG. 23B. At t = 4 ms, no or minimal distinction may be seen between the tubes (the tubes including inertial and stable cavitation effects) for the image 2330i with a low value of m and the images 2330j with a high value of m, which may be due to both tubes cavitating stably.

[0218] Colorbars for the images 2330g-j of FIGS. 23 G- J, respectively, may be provided on a linear scale.

[0219] Computer Support

[0220] FIG. 24 is a schematic view of a computer network in which embodiments may be implemented. Client computer(s) / devices 50 and server computer(s) 60 provide processing, storage, and input / output (I / O) devices executing application programs and the like. Client computer(s) / device(s) 50 can also be linked through communications network 70 to other computing devices, including other client device(s) / processor(s) 50 and server computer(s) 60. The communications network 70 can be part of a remote access network, a global network (e.g., the Internet), cloud computing servers or service, a worldwide collection of computers, local area or wide area networks, and gateways that currently use respective protocols (e.g., TCP / IP, Bluetooth®, etc.) to communicate with one another. Other electronic device / computer network architectures are also suitable.- 44 - 4239812. vl5200.2434002

[0221] FIG. 25 is a block diagram illustrating an example embodiment of a computer node (e.g., client processor(s) / device(s) 50 or server computer(s) 60) in the computer network 70 of FIG. 24. Each computer node 50, 60 contains system bus 79, where a bus is a set of hardware lines used for data transfer among components of a computer or processing system. The system bus 79 is essentially a shared conduit that connects different elements of a computer system (e.g., processor, disk storage, memory, I / O ports, network ports, etc.) that enables transfer of information between the elements. Attached to the system bus 79 is an I / O devices interface 82 for connecting various input and output devices (e.g., keyboard, mouse, display(s), printer(s), speaker(s), etc.) to the computer node 50, 60. A network interface 86 allows the computer node to connect to various other devices attached to a network (e.g., the network 70 of FIG. 24). A memory 90 provides volatile storage for computer software instructions 92a and data 94a used to implement embodiments of the present disclosure (e.g., the method 201 of FIG. 2, the method 301 of FIG. 3, etc.). A disk storage 95 provides nonvolatile storage for the computer software instructions 92b and data 94b used to implement an embodiment of the present disclosure. A central processor unit (CPU) 84 is also attached to the system bus 79 and provides for execution of computer instructions.

[0222] In an embodiment, the processor routines 92a-92b and data 94a-94b are a computer program product (generally referenced as 92), including a non-transitory, computer readable medium (e.g., a removable storage medium such as DVD-ROM(s), CD-ROM(s), diskette(s), tape(s), etc.) that provides at least a portion of the software instructions for the disclosure methods. The computer program product 92 can be installed by any suitable software installation procedure, as is well known in the art. In another embodiment, at least a portion of the software instructions may also be downloaded over a cable, communication, and / or wireless connection. In other embodiments, the disclosure programs are a computer program propagated signal product embodied on a propagated signal on a propagation medium (e.g., a radio wave, an infrared wave, a laser wave, a sound wave, or an electrical wave propagated over a global network such as the Internet, or other network(s)). Such carrier medium or signals provide at least a portion of the software instructions for the present disclosure routines / program 92.

[0223] In alternative embodiments, the propagated signal is an analog carrier wave or digital signal carried on the propagated medium. For example, the propagated signal may be a digitized signal propagated over a global network (e.g., the Internet), a telecommunications network, or other networks (such as the network 70 of FIG. 24). In one embodiment, the- 45 - 4239812. vl5200.2434002propagated signal is a signal that is transmitted over the propagation medium over a period of time, such as the instructions for a software application sent in packets over a network over a period of milliseconds, seconds, minutes, or longer. In another embodiment, the computer readable medium of the computer program product 92 is a propagation medium that the computer system 50 may receive and read, such as by receiving the propagation medium and identifying a propagated signal embodied in the propagation medium, as described above for computer program propagated signal product.

[0224] Generally speaking, the term “carrier medium” or transient carrier encompasses the foregoing transient signals, propagated signals, propagated medium, storage medium, and the like.

[0225] In other embodiments, the program product 92 may be implemented as a so-called Software as a Service (SaaS), or other installation or communication supporting end-users.

[0226] Embodiments or aspects thereof may be implemented in the form of hardware including but not limited to hardware circuitry, firmware, or software. If implemented in software, the software may be stored on any non-transient computer readable medium that is configured to enable a processor to load the software or subsets of instructions thereof. The processor then executes the instructions and is configured to operate or cause an apparatus to operate in a manner as described herein.

[0227] Further, hardware, firmware, software, routines, or instructions may be described herein as performing certain actions and / or functions of the data processors. However, it should be appreciated that such descriptions contained herein are merely for convenience and that such actions in fact result from computing devices, processors, controllers, or other devices executing the firmware, software, routines, instructions, etc.

[0228] It should be understood that the flow diagrams, block diagrams, and network diagrams may include more or fewer elements, be arranged differently, or be represented differently. But it further should be understood that certain implementations may dictate the block and network diagrams and the number of block and network diagrams illustrating the execution of the embodiments be implemented in a particular way.

[0229] Accordingly, further embodiments may also be implemented in a variety of computer architectures, physical, virtual, cloud computers, and / or some combination thereof, and, thus, the data processors described herein are intended for purposes of illustration only and not as a limitation of the embodiments.- 46 - 4239812. vl5200.2434002

[0230] The teachings of all patents, published applications and references cited herein are incorporated by reference in their entirety.

[0231] While example embodiments have been particularly shown and described, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the scope of the embodiments encompassed by the appended claims.- 47 - 4239812. vl

Claims

5200.2434002CLAIMSWhat is claimed is:

1. A computer-implemented method for monitoring ultrasound therapy, the computer- implemented method comprising:dynamically classifying, by a processor, cavitation effects of at least one ultrasound stimulus on bubbles in a biological setting, the bubbles introduced into the biological setting for the ultrasound therapy, the cavitation effects observed via an ultrasound receiver and dynamically classified, by the processor, into stable cavitation effects and inertial cavitation effects, the stable and inertial cavitation effects inducing distinct bio-effects on biological tissue in the biological setting, the biological tissue targeted by the at least one ultrasound stimulus; andoutputting, by the processor, a representation of an action to be taken for the ultrasound therapy, the action being a function of the stable and inertial cavitation effects dynamically classified.

2. The computer-implemented method of Claim 1, further comprising:determining, by the processor, respective spatial locations, within the biological tissue, of the stable and inertial cavitation effects dynamically classified; andoutputting, by the processor, representations of the respective spatial locations determined, the representations of the respective spatial locations determined and output enabling an operator or controller to make an informed adjustment to the at least one ultrasound stimulus.

3. The computer-implemented method of Claim 1, further comprising causing an ultrasound device communicatively coupled to the processor to deliver the at least one ultrasound stimulus, the at least one ultrasound stimulus being focused upon a therapeutic target within the biological tissue.

4. The computer-implemented method of Claim 1, wherein the dynamically classifying includes applying a mathematical decomposition to measurements acquired via the ultrasound receiver to extract components of the measurements, wherein the measurements represent the cavitation effects of the at least one ultrasound stimulus- 48 - 4239812. vl5200.2434002on the bubbles, and wherein the dynamically classifying is based on one or more subsets of the components extracted via the mathematical decomposition applied.

5. The computer-implemented method of Claim 4, further comprising, by the processor, causing the ultrasound receiver to acquire the measurements, wherein at least a portion of the stable cavitation effects or the inertial cavitation effects is associated with each subset of the one or more subsets of the components extracted, and wherein the dynamically classifying further includes comparing at least a portion of the one or more subsets of the components extracted.

6. The computer-implemented method of Claim 4, further comprising:introducing a shift to the measurements, the shift based on a time delay experienced by a receiver element of the ultrasound receiver; andcomputing a covariance of the measurements, the covariance associated with a similarity between each measurement of the measurements, wherein the dynamically classifying includes extracting the components of the covariance of the measurements computed, wherein the measurements are time-domain measurements, and wherein the applying includes applying the mathematical decomposition to the time-domain measurements.

7. The computer-implemented method of Claim 4, wherein the measurements are time-domain measurements and wherein the applying includes applying the mathematical decomposition to the time-domain measurements.

8. The computer-implemented method of Claim 4, wherein applying the mathematical decomposition includes applying an Eigen decomposition to extract at least a subset of the components, and wherein the subset of the components extracted using the Eigen decomposition are associated with frequency components of the measurements.

9. The computer-implemented method of Claim 1, wherein the stable cavitation effects classified are associated with oscillation of the bubbles in the biological setting and wherein the inertial cavitation effects classified are associated with expansion and collapse of the bubbles in the biological setting.

10. The computer-implemented method of Claim 1, further comprising, by the processor:- 49 - 4239812. vl5200.2434002rendering a representation of the stable and inertial cavitation effects dynamically classified, the representation rendered with respect to the biological setting; andoutputting the representation rendered toward displaying the representation rendered on an electronic display, the electronic display communicatively coupled to the processor.

11. The computer-implemented method of Claim 1, wherein the action to be taken includes (i) adjusting the at least one ultrasound stimulus, (ii) modifying a dosage of the bubbles introduced into the biological setting, or a combination of (i) and (ii).

12. A computer-based system for monitoring ultrasound therapy, the computer-based system comprising:at least one processor; andat least one memory, the at least one memory having encoded thereon a sequence of instructions which, when loaded and executed by the at least one processor, causes the computer-based system to:dynamically classify cavitation effects of at least one ultrasound stimulus on bubbles in a biological setting, the bubbles introduced into the biological setting for the ultrasound therapy, the cavitation effects observed via an ultrasound receiver and dynamically classified into stable cavitation effects and inertial cavitation effects, the stable and inertial cavitation effects inducing distinct bio-effects on biological tissue in the biological setting, the biological tissue targeted by the at least one ultrasound stimulus; and output a representation of an action to be taken for the ultrasound therapy, the action being a function of the stable and inertial cavitation effects dynamically classified.

13. The computer-based system of Claim 12, wherein the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to:determine respective spatial locations, within the biological tissue, of the stable and inertial cavitation effects dynamically classified; and- 50 - 4239812. vl5200.2434002output representations of the respective spatial locations determined, the representations of the respective spatial locations determined and output enabling an operator or controller to make an informed adjustment to the at least one ultrasound stimulus.

14. The computer-based system of Claim 12, wherein the stable cavitation effects classified are associated with oscillation of the bubbles in the biological setting and wherein the inertial cavitation effects classified are associated with expansion and collapse of the bubbles in the biological setting and wherein the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to:cause an ultrasound device communicatively coupled to the processor to deliver the at least one ultrasound stimulus, the at least one ultrasound stimulus being focused upon a therapeutic target within the biological tissue, wherein the action to be taken includes (i) adjusting the at least one ultrasound stimulus, (ii) modifying a dosage of the bubbles introduced into the biological setting, or a combination of (i) and (ii).

15. The computer-based system of Claim 12, wherein, to dynamically classify the cavitation effects, the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to apply a mathematical decomposition to measurements acquired via the ultrasound receiver to extract components of the measurements, wherein the measurements represent the cavitation effects of the at least one ultrasound stimulus on the bubbles, and wherein dynamically classifying the cavitation effects is based on one or more subsets of the components extracted via the mathematical decomposition applied.

16. The computer-based system of Claim 15, wherein at least a portion of the stable cavitation effects or the inertial cavitation effects is associated with each subset of the one or more subsets of the components extracted and wherein, to dynamically classify the cavitation effects, the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to compare at least a portion of the one or more subsets of the components extracted.- 51 - 4239812. vl5200.243400217. The computer-based system of Claim 15, wherein the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to:introduce a shift to the measurements, the shift based on a time delay experienced by a receiver element of the ultrasound receiver, and compute a covariance of the measurements, the covariance associated with a similarity between each measurement of the measurements, and wherein, to dynamically classify the cavitation effects, the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to extract the components of the covariance of the measurements computed, wherein the measurements are time-domain measurements, and wherein the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to apply the mathematical decomposition to the time-domain measurements.

18. The computer-based system of Claim 15, wherein, to apply the mathematical decomposition to the measurements, the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to apply an Eigen decomposition to extract at least a subset of the components, and wherein the subset of the components extracted using the Eigen decomposition are associated with frequency components of the measurements.

19. The computer-based system of Claim 15, wherein the sequence of instructions, when loaded and executed by the at least one processor, further causes the computer-based system to:render a representation of the stable and inertial cavitation effects dynamically classified, the representation rendered with respect to the biological setting; and output the representation rendered toward displaying the representation rendered on an electronic display, the electronic display communicatively coupled to the processor.

20. A non-transitory computer-readable medium having encoded thereon a sequence of instructions which, when loaded and executed by at least one processor, causes the at least one processor to:- 52 - 4239812. vl5200.2434002dynamically classify cavitation effects of at least one ultrasound stimulus on bubbles in a biological setting, the bubbles introduced into the biological setting for the ultrasound therapy, the cavitation effects observed via an ultrasound receiver and dynamically classified into stable cavitation effects and inertial cavitation effects, the stable and inertial cavitation effects inducing distinct bio-effects on biological tissue in the biological setting, the biological tissue targeted by the at least one ultrasound stimulus; andoutput a representation of an action to be taken for the ultrasound therapy, the action being a function of the stable and inertial cavitation effects dynamically classified.- 53 - 4239812. vl

Images