arXiv subjects

Gianmarco Pinton

Publications and source records attributed to Gianmarco Pinton.

16 recordsLinked to original sources

arXiv2026

An Angular Spectrum Method for Nonlinear Propagation in Heterogeneous Tissue with Immersed Sources for Ultrasound

A modified angular spectrum method (ASM) is developed for three-dimensional nonlinear acoustic propagation through heterogeneous tissue, targeting transcranial and therapeutic ultrasound. First, a consistent obliquity correction is applied to both the linear and nonlinear operators of the split-step update. The attenuation and dispersion filter carries a per-mode k/kz factor on the full wavenumber-frequency grid, so each component accumulates absorption and phase over its true path length dz/cos(theta), while the Burgers coefficient is scaled by the power-weighted mean beam obliquity. Second, the retarded-time Burgers update is discretized with a second-order Kurganov-Tadmor central-upwind flux using MUSCL reconstruction and SSP-RK2 time integration, composed with diffraction and attenuation through Strang splitting with adaptive CFL sub-cycling, resolving fully developed shocks without the temporal refinement that the CFL coupling imposes on FDTD. Third, a plane-by-plane source-injection scheme decomposes deeply curved bowl transducers into axial slices injected at their correct propagation depths, preserving the aperture-dependent shock-formation distance. Phase-and-amplitude screens derived from skull CT data model transcranial aberration and insertion loss, and three enhanced absorbing-boundary treatments reduce boundary reflections by a factor of 2.4. Analytical validation reproduces the baffled-piston far-field pattern to 0.014 RMS and the focused-piston focal pressure to within 2.3 percent. In a transcranial benchmark through an ex vivo human skull, the ASM matches the Fullwave 2 focal-plane intensity to 1.1 percent RMS and predicts a 5.4 dB through-skull insertion loss. For a bowl transducer (R = 80 mm, f0 = 1 MHz), the ASM matches the Fullwave 2 focal depth to within 2.3 percent with 9x less memory and 2.9x less wall time.

physics.med-ph↗

arXiv2026

An open-source framework for predicting ultrasound neuromodulation: bridging tissue elastomechanics and neuron firing dynamics

Transcranial focused ultrasound is a non-invasive neuromodulation modality with millimetre-scale resolution, but its biophysical mechanism of action remains unresolved. Exposure is conventionally specified by transducer surface or derated focal pressure, quantities only indirectly related to what matters for therapy: which neurons fire and through which pathway. We address this gap with an end-to-end computational framework that maps a transcranial acoustic field to per-voxel neural firing maps registered to anatomy. The pipeline couples heterogeneous nonlinear full-wave acoustic propagation, viscoelastic shear-wave propagation, Pennes bioheat diffusion, a bilayer-mechanics conversion from tissue strain to membrane tension, and a multi-compartment Hodgkin-Huxley neuron carrying mechanosensitive, cavitation-coupled, calcium-coupled, thermosensitive, astrocytic-gliotransmitter, and mechanosensitive-synaptic pathways. Six candidate mechanisms are implemented as interchangeable components on a shared neuron model, so their firing predictions can be compared directly on the same field, and every numerical parameter is classified by source and bracketed by sensitivity analysis. We demonstrate the framework on a theta-burst sonication delivered through a micro-CT human-skull specimen targeting the left dorsal anterior cingulate cortex, predicting a focal firing zone of approximately 8,500 mm^3 at a focal thermal rise well within ITRUSST consensus safety envelopes. The principal output is a per-voxel firing-volume map resolved jointly with the acoustic, elastic, and thermal field histories that drive it, giving spatially resolved, falsifiable predictions that are testable against high-density extracellular recordings and support parameter estimation, cell-type-resolved mechanism identification, and quantitative safety assessment for ultrasound neuromodulation.

physics.med-ph↗

arXiv2026

A fullwave model of the nonlinear wave equation with multiple relaxations and relaxing perfectly matched layers for high-order numerical finite-difference solutions

Large-scale acoustic simulations underpin the development of ultrasound imaging and therapy, but modeling nonlinearity, frequency-dependent attenuation, and absorbing boundaries in heterogeneous tissue remains computationally demanding. We present Fullwave 2, a unified time-domain formulation with arbitrary power-law tissue attenuation and perfectly matched layers (PMLs) within a single framework suited to high-order finite difference solution. Attenuation and dispersion are encoded directly into complex coordinate-stretched spatial derivatives through multiple relaxation mechanisms. Because the same mechanism describes both interior tissue attenuation and the absorbing boundary, the convolutional PML (C-PML) becomes a special case of the domain-wide model and adds no extra computational burden. The formulation preserves the structure of the d'Alembertian operator, which allows high-order staggered-grid finite difference stencils optimized for long-distance propagation, and a two-stage C-PML with a transition region is introduced to ensure numerical stability in the presence of multiple relaxations. The domain-wide multiple relaxation model reproduces power-law attenuation with less than 5% attenuation error and 0.5% phase-velocity error over 1-20 MHz. The two-stage C-PML reaches reflection coefficients below -49 dB with a compact 4 lambda footprint. Nonlinear propagation is validated against a 1D Burgers solution, with agreement up to the 7^(th) harmonic. The framework is demonstrated on 2D abdominal wall imaging and 3D transcranial rat skull simulations, where it accurately captures complex scattering and aberration artifacts. Fullwave 2 unifies nonlinear propagation, arbitrary power-law attenuation, and absorbing boundaries in a single, computationally efficient time-domain formulation, providing an accurate and scalable wave propagation tool for medical ultrasound research.

physics.med-ph↗

arXiv2026

A reduced viscoelastic FDTD formulation for ultrasound-driven shear wave propagation in soft tissue

Ultrasound-driven shear wave propagation in soft tissue underlies shear wave elastography (SWE) and emerging elastomechanical hypotheses of ultrasonic neuromodulation, both of which require accurate, efficient modeling of radiation-force--induced tissue motion. General-purpose finite-element elastodynamic solvers are often computationally expensive and unnecessarily broad for shear-dominant applications. We derive a reduced viscoelastic formulation by applying near-incompressibility, small-strain linearization, Helmholtz decomposition, and solenoidal force projection to the full Navier equations, yielding a Kelvin--Voigt shear wave equation that retains only the transverse dynamics relevant to radiation-force--induced motion. An explicit finite-difference time-domain (FDTD) implementation with second-order spatial and temporal accuracy enforces the solenoidal body-force constraint via a matrix-free conjugate-gradient Poisson solve. For separable radiation-force sources, a pre-computed projection reduces this cost by one to two orders of magnitude. In homogeneous media the solver recovers the theoretical shear wavespeed to within $<$0.1\%, exhibits clear second-order grid convergence (trace-$L_2$ self-convergence, $p\gtrsim2$), and matches analytical Kelvin--Voigt attenuation and phase speed to within ${\sim}3\%$ and $<$1\% over a 16-point viscosity sweep ($η\in[0,1.5]$~Pa$\cdot$s). An end-to-end RSNA QIBA phantom benchmark recovers shear wave speeds to within ${\sim}1\%$ across a tenfold shear-modulus range ($G=1$--$10$~kPa). The framework accommodates spatial heterogeneity in shear modulus, density, and viscosity, and integrates with acoustic simulators. Transcranial demonstrations through micro-CT skull geometries produce shear displacements of 1.7--5~$μ$m consistent with clinical ARFI.

physics.med-ph↗

arXiv2026

IQ-JEPA: A Joint-Embedding Predictive Architecture with a Hermitian Vision Transformer for Sound Speed and Attenuation Estimation from Ultrasound IQ Data

The speed of sound in tissue is a prerequisite for well-focused imaging and has diagnostic value, but recovering it from raw pulse-echo channel data is fundamentally a nonlinear inverse problem. Learned solvers are fast yet label hungry. Simulated sound-speed labels are expensive, while abundant real channel data is unlabeled. We propose IQ-JEPA to exploit both data types. An encoder is pretrained without labels to predict the latent representation of masked in-phase and quadrature (IQ) regions from visible context, then fine-tuned on simulated maps. Sound speed appears in the IQ signal as a phase difference, invariant to the constant phase offset. The encoder is a Hermitian vision transformer that operates on the complex signal directly. Its attention is equivariant to that phase and its conjugate-product feed-forward is invariant to it, so the encoder reads a quantity analogous to the one classical coherence methods use. On 79,293 Fullwave 2.5 simulations at 2.5 MHz, pretraining on the 63,435 unlabeled acquisitions reaches 15.60 m/s at 10,000 labels. This is a roughly threefold gain in label efficiency over supervised training, growing to over fourfold at 1,000 labels. It is about 2.2x below an InversionNet baseline, and 8.71 m/s at full labels. The gain still grows with more unlabeled pretraining data. Our comparisons point to self-supervision as the dominant factor. The same encoder transfers. Its frozen features expose sound speed and attenuation, and cross-distribution pretraining between layered and abdominal phantoms costs little accuracy. We see this as a first step toward a foundation model for quantitative ultrasound.

cs.LG↗

arXiv2026

Spatially heterogeneous power-law attenuation with multiple relaxation mechanisms for ultrasound modeling

Objective. The soft tissue attenuation laws have a magnitude and frequency dependence that varies across tissue types and generally follow power laws. An accurate model of ultrasound propagation in the human body thus may require spatially heterogeneous power-law attenuation alpha(x,f) = alpha_0(x) f^(y(x)). However, a spatially heterogeneous representation of frequency-dependent attenuation is technically challenging, so existing methods introduce simplifying assumptions. For example, prior approaches such as Fullwave 2 achieved <5% error for individual tissue types but required manual parameter tuning for each (alpha_0, y) pair, limiting the construction of realistic tissue libraries. Approach. We introduce a calibration framework that uses derivative-free optimization to systematically fit relaxation parameters across diverse tissue combinations spanning alpha_0 = 0.0022-1.0 dB/(MHz^y cm) and y = 0.4-2.0. The Nelder-Mead algorithm minimizes complex-wavenumber mismatch. The attenuation is extended to a convolutional perfectly matched layer, where the same relaxation formulation is used in the boundaries. Main results. The method achieves mean errors below 3% over 1-20 MHz with dispersion error of 1.1 +/- 0.8 m/s across the clinically relevant core region (y = 0.7-1.4). Boundary reflections remain below -50 dB for clinically relevant tissue exponents (y <= 1.5). We validated the method with two-layer muscle/fat/liver models and confirmed per-layer accuracy (<2.5% normalized error). A 3D abdominal simulation using the Visible Human dataset demonstrates stable propagation with voxel-level heterogeneity in both alpha_0(x) and y(x). Significance. The open-source multi-GPU implementation (Fullwave 2.5) provides a practical tool for patient-specific therapy planning, training data generation, estimation of acoustic radiation force, quantitative imaging, and inverse problem applications.

physics.med-ph↗

arXiv2026

mach: ultrafast ultrasound beamforming

Purpose: Volumetric ultrafast ultrasound produces massive datasets with high frame rates, dense reconstruction grids, and large channel counts. Beamforming computational demands limit research throughput and prevent real-time applications in emerging modalities such as elastography, functional neuroimaging, and microscopy. Approach: We developed mach, an open-source, GPU-accelerated beamformer with a highly optimized delay-and-sum CUDA kernel and an accessible Python interface. mach uses a hybrid delay computation strategy that substantially reduces memory overhead compared to fully precomputed approaches. The CUDA implementation optimizes memory layout for coalesced access and reuses delay computations across frames via shared memory. We benchmarked mach on the PyMUST rotating disk dataset and validated numerical accuracy against existing open-source beamformers. Results: mach processes 1.1 trillion points per second on a consumer-grade GPU, achieving $>$10$\times$ faster performance than existing open-source GPU beamformers. On the PyMUST rotating disk benchmark, mach completes reconstruction in 0.23~ms, 6$\times$ faster than the acoustic round-trip time to the imaging depth. Validation against other beamformers confirms numerical accuracy with errors below $-60$~dB for Power Doppler and $-120$~dB for B-mode. Conclusions: mach achieves 1.1 trillion points per second throughput, enabling real-time 3D ultrafast ultrasound reconstruction for the first time on consumer-grade hardware. By eliminating the beamforming bottleneck, mach enables real-time applications such as 3D functional neuroimaging, intraoperative guidance, and ultrasound localization microscopy. mach is freely available at https://github.com/Forest-Neurotech/mach

physics.med-ph↗

arXiv2025

Ultrasound Lung Aeration Map via Physics-Aware Neural Operators

Lung ultrasound is a growing modality in clinics for diagnosing and monitoring acute and chronic lung diseases due to its low cost and accessibility. Lung ultrasound works by emitting diagnostic pulses, receiving pressure waves and converting them into radio frequency (RF) data, which are then processed into B-mode images with beamformers for radiologists to interpret. However, unlike conventional ultrasound for soft tissue anatomical imaging, lung ultrasound interpretation is complicated by complex reverberations from the pleural interface caused by the inability of ultrasound to penetrate air. The indirect B-mode images make interpretation highly dependent on reader expertise, requiring years of training, which limits its widespread use despite its potential for high accuracy in skilled hands. To address these challenges and democratize ultrasound lung imaging as a reliable diagnostic tool, we propose LUNA, an AI model that directly reconstructs lung aeration maps from RF data, bypassing the need for traditional beamformers and indirect interpretation of B-mode images. LUNA uses a Fourier neural operator, which processes RF data efficiently in Fourier space, enabling accurate reconstruction of lung aeration maps. LUNA offers a quantitative, reader-independent alternative to traditional semi-quantitative lung ultrasound scoring methods. The development of LUNA involves synthetic and real data: We simulate synthetic data with an experimentally validated approach and scan ex vivo swine lungs as real data. Trained on abundant simulated data and fine-tuned with a small amount of real-world data, LUNA achieves robust performance, demonstrated by an aeration estimation error of 9% in ex-vivo lung scans. We demonstrate the potential of reconstructing lung aeration maps from RF data, providing a foundation for improving lung ultrasound reproducibility and diagnostic utility.

eess.IV↗

arXiv2022

Benchmark problems for transcranial ultrasound simulation: Intercomparison of compressional wave models

Computational models of acoustic wave propagation are frequently used in transcranial ultrasound therapy, for example, to calculate the intracranial pressure field or to calculate phase delays to correct for skull distortions. To allow intercomparison between the different modeling tools and techniques used by the community, an international working group was convened to formulate a set of numerical benchmarks. Here, these benchmarks are presented, along with intercomparison results. Nine different benchmarks of increasing geometric complexity are defined. These include a single-layer planar bone immersed in water, a multi-layer bone, and a whole skull. Two transducer configurations are considered (a focused bowl and a plane piston), giving a total of 18 permutations of the benchmarks. Eleven different modeling tools are used to compute the benchmark results. The models span a wide range of numerical techniques, including the finite-difference time-domain method, angular-spectrum method, pseudospectral method, boundary-element method, and spectral-element method. Good agreement is found between the models, particularly for the position, size, and magnitude of the acoustic focus within the skull. When comparing results for each model with every other model in a cross comparison, the median values for each benchmark for the difference in focal pressure and position are less than 10\% and 1 mm, respectively. The benchmark definitions, model results, and intercomparison codes are freely available to facilitate further comparisons.

physics.comp-ph↗

arXiv2021

Deconstruction and reconstruction of image-degrading effects in the human abdomen using Fullwave: phase aberration, multiple reverberation, and trailing reverberation

Ultrasound image degradation in the human body is complex and occurs due to the distortion of the wave as it propagates to and from the target. Here, we establish a simulation based framework that deconstructs the sources of image degradation into a separable parameter space that includes phase aberration from speed variation, multiple reverberations, and trailing reverberation. These separable parameters are then used to reconstruct images with known and independently modulable amounts of degradation using methods that depend on the additive or multiplicative nature of the degradation. Experimental measurements and Fullwave simulations in the human abdomen demonstrate this calibrated process in abdominal imaging by matching relevant imaging metrics such as phase aberration, reverberation strength, speckle brightness and coherence length. Applications of the reconstruction technique are illustrated for beamforming strategies (phase aberration correction, spatial coherence imaging), in a standard abdominal environment, as well as in impedance ranges much higher than those naturally occurring in the body.

physics.med-ph↗

arXiv2021

Super-resolved shear shock focusing in the human head

Shear shocks, which exist in a completely different regime from compressional shocks, were recently observed in the brain. These low phase speed ($\approx$ 2 m/s) high Mach number ($\approx$ 1) waves could be the primary mechanism behind diffuse axonal injury due to a very high local acceleration at the shock front. The extreme nonlinearity of these waves results in unique behaviors that are different from more commonly studied nonlinear compressional waves. Here we show the first observation of super-resolved shear shock wave focusing. Shear shock wave imaging and numerical simulations in a human head phantom over a range of frequencies/amplitudes shows the super-resolution of shock waves in the low strain and high strain-rate regime. These results suggest that even for mild accelerations injuries as small as a grain of rice on the scale of mm$^2$ can be easily created deep inside the brain.

physics.med-ph↗

arXiv2020

Ultrasound imaging with three dimensional full-wave nonlinear acoustic simulations. Part 2: sources of image degradation in intercostal imaging

Fullwave simulations are applied to an intercostal imaging scenario to determine the sources of fundamental and harmonic image degradation with respect to aberration and reverberation. These simulations are based on Part I of this two part paper, which established the Fullwave simulation methods to generate realistic ultrasound images based directly on the first principles of wave propagation in the human body. The ultasound images are generated based on the first principles of propagation and reflection and they describe interplay between distributed aberration and reverberation clutter. Three imaging scenarios that would not be realizable in vivo are investigated in silico. First, the ribs were completely removed and replaced with fat. Then, the ribs were maintained in their anatomically correct configuration to yield a reference image. Finally the ribs were placed closer together in elevation. The propagation based B-mode images show that of these three scenarios the second, anatomically correct configuration, has the best contrast-to-noise ratio. This is due to two competing effects. First the ribs effectively apodize the fundamental and harmonic beams by 3-5 dB. This effect alone would predict an improvement in image quality. However, the B-mode image quality, measured by the contrast-to-noise ratio degrades by 8%. To explain these changes, it is shown that a second effect, multiple reverberation, must be taken into account. A point spread function analysis shows that when the ribs are placed closer together they generate significantly more reverberation clutter (by 2.4 to 2.9 dB), degrading the image quality even though the beamplot has lower sidelobes. In this intercostal imaging scenario the effects of the ribs on beam shape and reverberation are therefore in competition in terms of image quality and there is an optimal acoustic window that balances them out.

physics.med-ph↗

arXiv2020

Ultrasound imaging of the human body with three dimensional full-wave nonlinear acoustics. Part 1: simulations methods

Simulations of three dimensional ultrasound propagation in heterogeneous media are computationally intensive due to the constraints arising from the large size of the domain, which is on the order of hundreds of wavelengths, and the small size of scatterers, which are much smaller than a wavelength. Consequently, three dimensional ultrasound imaging simulations are currently based on models that simplify the propagation physics. Here the full three dimensional wave physics is simulated with finite differences to generate ultrasound images of the human body based directly on the first principles of propagation and backscattering. The Visible Human project, a 3D data set of the human body that was generated with photographs of 0.33 mm cryosections, is converted into 3D acoustical maps. A full-wave nonlinear acoustic simulation tool is used to propagate ultrasound into the liver with a 2D intercostal ultrasound array in a $93 \times 39 \times 22$ mm domain with $6\times10^8$ points. Imaging metrics, based on the beamplots, root-mean-square phase aberration, spatial coherence lengths, and contrast-to-noise ratio are used to characterize the image quality. It is shown that the harmonic image quality is better than the fundamental image quality due, in part, to a narrower beam profile. The root-mean-square estimate of aberration after propagation through the simulated body wall is shown to be low (23.4 ns), consistently with previous reports of aberration measured experimentally in a human body wall. The spatial coherence measured at the transducer surface indicates that a transducer array element size of $<0.81 λ$ would be required to fully sample the acoustic field. These simulated three dimensional ultrasound images based directly on propagation physics provide a platform to investigate the sources of image degradation in three dimensions (included in Part II).

physics.med-ph↗

arXiv2020

Quantitative sub-resolution blood velocity estimation using ultrasound localization microscopy ex-vivo and in-vivo

Super-resolution ultrasound imaging relies on the sub-wavelength localization of microbubble contrast agents. By tracking individual microbubbles, the velocity and flow within microvessels can be estimated. However, a 2D super-resolution image can only localize bubbles with sub-wavelength resolution in the imaging plane whereas the resolution in the elevation plane is limited by conventional beamwidth physics. Since ultrasound imaging integrates echoes over the elevation dimension, velocity estimates at a single location in the imaging plane include information throughout the imaging slice thickness. This slice thickness is typically a few orders or magnitude larger than the super-resolution limit. It is shown here that in order to estimate the velocity, a spatial integration over the elevation direction must be considered. This operation yields a multiplicative correction factor that compensates for the elevation integration. A correlation-based velocity estimation technique is then presented. Calibrated microtube phantom experiments are used to validate the proposed velocity estimation method and the proposed elevation integration correction factor. It is shown that velocity measurements are in agreement with theoretical predictions within the considered range of flow rates (10 to 90 $μ$L/min). Then, the proposed technique is applied to two in-vivo mouse tail experiments imaged with a low frequency human clinical transducer with human clinical concentrations of microbubbles. In the first experiment, a vein was visible with a diameter of 140 $μ$m and a peak flow velocity of 0.8 mm/s. In the second experiment, a vein was observed in the super-resolved image with a diameter of 120~$μ$m and with maximum local velocity of $\approx$~4.4~mm/s. It is shown that the parabolic flow profiles within these micro-vessels are resolvable.

physics.med-ph↗

arXiv2018

Super resolution imaging through the human skull

High resolution transcranial ultrasound imaging in humans has been a persistent challenge for ultrasound due to the imaging degradation effects from aberration and reverberation. These mechanisms depend strongly on skull morphology and they have high variability across individuals. Here we demonstrate the feasibility of human transcranial super-resolution imaging using a geometrical focusing approach to concentrate energy at the region of interest, and a phase correction focusing approach that takes the skull morphology into account. It is shown that using the proposed focused method, we can image a 208$μ$m microtube behind a human skull phantom in both an out-of-plane and an in-plane configuration. Individual phase correction profiles for the temporal region of the human skull were calculated and applied to transmit-receive a custom-focused super-resolution imaging sequence through a human skull phantom, targeting the microtube, at 68.5mm in depth, at 2.5 MHz. Microbubble contrast agents were diluted to a concentration of 1.6$\times$10$^6$ bubbles/mL and perfused through the microtube. It is shown that by correcting for the skull aberration, the RF signal amplitude from the tube improved by a factor of 1.6 in the out-of-plane focused emission case. The lateral registration error of the tube's position, which in the uncorrected case was 990 $μ$m, was reduced to 50$μ$m in the corrected case as measured in the B-mode images. Sensitivity in microbubble detection for the phase corrected case increased by a factor of 1.5 in the out-of-plane imaging case, while in the in-plane case it improved by a factor of 1.3 while achieving an axial registration correction from an initial 1885$μ$m error for the uncorrected emission, to a 284$μ$m error for the corrected counterpart. These findings suggest that super-resolution may be used more generally as a clinical imaging modality in the brain.

physics.med-ph↗

arXiv2017

Shear shock waves are observed in the brain

The internal deformation of the brain is far more complex than the rigid motion of the skull. An ultrasound imaging technique that we have developed has a combination of penetration, frame-rate, and motion detection accuracy required to directly observe, for the first time, the formation and evolution of shear shock waves in the brain. Experiments at low impacts on the traumatic brain injury scale demonstrate that they are spontaneously generated and propagate within the porcine brain. Compared to the initially smooth impact, the acceleration at the shock front is amplified up to a factor of 8.5. This highly localized increase in acceleration suggests that shear shock waves are a fundamental mechanism for traumatic injuries in soft tissue.

cond-mat.soft↗