← Back to publications

How Do You Simulate Ultrasound Through the Skull Without a Supercomputer?

A simplified summary of the research

2022PLOS ONEOctober 2022Loading citations…

S.M.A.R.T. F.U.S: Surrogate Model of Attenuation and Refraction in Transcranial Focused Ultrasound

Cain, J. A., Visagan, S., Monti, M. M.

  • Methods
  • Modeling
  • Focused Ultrasound
doi:10.1371/journal.pone.0264101
In plain English

What's the Point?

Background

The skull distorts and weakens ultrasound as it enters the brain. To target deep regions safely, researchers need to simulate how each person's skull will bend the beam—but that normally takes hours on a supercomputer.

Methods

We ran over 12,000 full physics simulations across a five-dimensional parameter space (skull thickness, curvature, angle, etc.) and used them to train a fast surrogate model. The result is SMART FUS, an open-source MATLAB toolbox.

Findings

SMART FUS predicts skull-induced attenuation and refraction in seconds instead of hours, with accuracy comparable to full simulations for typical clinical parameters.

Impact

It removes a major technical barrier to adopting transcranial focused ultrasound, letting more labs plan safer, better-targeted studies without high-performance computing.

Abstract

From the manuscript

Transcranial focused ultrasound neuromodulation and therapy require accurate estimates of the pressure field delivered through the skull, but full-wave simulations are computationally expensive. We introduce SMART FUS, an open-source MATLAB toolbox that uses over 12,000 k-Wave simulations across a five-dimensional parameter space to build a surrogate model of skull-induced attenuation and refraction. The tool provides fast, accurate estimates for typical clinical parameter regimes and is intended to lower the barrier to entry for research groups adopting transcranial focused ultrasound.

Full citation

Cain, J. A., Visagan, S., & Monti, M. M. (2022). S.M.A.R.T. F.U.S: Surrogate Model of Attenuation and Refraction in Transcranial Focused Ultrasound. PLOS ONE, 17(10), e0264101.

© 2026 Josh Cain