About

Gerald Popelka is a Consulting Professor of Otolaryngology and a Senior Scientist at the Rad/Radiological Sciences Laboratory at Stanford University. He is affiliated with the Center for Artificial Intelligence in Medicine & Imaging (AIMI) at Stanford. His work focuses on the application of artificial intelligence in medicine and imaging, contributing to research and development in these fields. As a senior scientist and professor, he is involved in advancing AI-driven healthcare solutions and supporting the mission of AIMI to improve medical imaging and diagnostics through innovative technological approaches.

Research topics

• Computer Science
• Medicine
• Artificial Intelligence
• Radiology
• Psychology
• Neuroscience
• Engineering
• Acoustics
• Telecommunications
• Speech recognition
• Physics
• Paleontology
• Geology
• Audiology

Selected publications

• A deep‐learning model for one‐shot transcranial ultrasound simulation and phase aberration correction

  Medical Physics · 2025-12-31 · 3 citations

  articleOpen access

  BACKGROUND: Transcranial ultrasound is a promising non-invasive neuromodulation technique with applications, including neuronal activity modulation, blood-brain barrier opening, targeted drug delivery, and thermal ablation. Its ability to deliver focused ultrasound waves to precise brain regions has led to over 50 clinical trials targeting conditions such as opioid addiction, Alzheimer's disease, dementia, epilepsy, and glioblastoma. However, skull heterogeneity complicates accurate focal spot prediction and energy delivery, requiring rapid yet precise phase aberration correction in clinical workflows. PURPOSE: To address the trade-off between computational efficiency and accuracy in current focus prediction methods, we introduce TUSNet, a deep learning framework for rapid and accurate transcranial ultrasound pressure field and phase aberration correction computation. METHODS: TUSNet, an end-to-end neural network, was trained to predict both 2D transcranial ultrasound pressure fields and phase corrections. TUSNet was trained on 180432 synthetic skull Computed Tomography (CT) segments, and tested on 1232 real skull CT segments. Its performance was benchmarked against k-Wave, a MATLAB-based acoustic simulation package, evaluating computation speed, focal spot accuracy, phase correction accuracy, and pressure magnitude estimation. RESULTS: faster than k-Wave, while achieving 98.3% accuracy in peak pressure magnitude estimation and a mean focal positioning error of only 0.18 mm relative to k-Wave ground truth. End-to-end training took approximately 8 h on 4x NVIDIA A100 80 GB GPUs. CONCLUSIONS: TUSNet demonstrates that deep learning can provide accurate and rapid estimates of phase aberrations and transcranial pressure fields, offering a promising direction for accelerating ultrasound treatment planning. While the present validation is based on simulated, noise-free ultrasound fields, the results establish a foundation that future experimental studies can build on to assess performance under real-world clinical conditions.

  PublisherOA PDFDOI

• Use of Ultra‐Short Echo Time MRI to Improve Temporal Bone Imaging

  The Laryngoscope · 2024-08-21

  articleOpen access

  OBJECTIVE: The short T2 nature of cortical bone causes it to appear similar to air on MR, forcing clinicians to rely on computed tomography imaging, with its attendant ionizing radiation exposure, to define temporal bone structures. Through the use of novel MR sequences with ultra-short echo times (UTE), short T2 structures are now able to be visualized, allowing for improved understanding of anatomical relationships. METHODS: Eight patients (50% female) undergoing MR imaging of the skull base for diagnostic purposes (62.5% for vestibular schwannoma surveillance) at a tertiary care center were enrolled to evaluate the safety and efficacy of UTE imaging. CT scans were completed in 37.5% of the patients as part of their workup and used for comparison purposes. The repetition time, short echo time, and long echo time for the UTE sequence were 11, 0.032, and 2.2 msec, respectively. RESULTS: The protocol added 6 min to the total scanning time, and all patients tolerated the sequence without issue. The ossicles, mastoid air cells, antrum, and epitympanum were able to be seen and had a high Dice similarity coefficient when compared to CT (>0.5). UTE allowed for clear delineation of all segments of the facial nerve with a signal-to-noise ratio of 35 (although the BRAVO sequences had a superior ratio of 140). Vestibular schwannomas were able to be distinguished from normal brain parenchyma. CONCLUSIONS: UTE is safe and effective for visualizing anatomic structures not normally seen on traditional MRI, potentially allowing for improved surgical planning in patients. LEVEL OF EVIDENCE: 3 Laryngoscope, 134:4691-4696, 2024.

  PublisherOA PDFDOI

• Focal Volume, Acoustic Radiation Force, and Strain in Two-Transducer Regimes

  IEEE Transactions on Ultrasonics Ferroelectrics and Frequency Control · 2024-09-06 · 7 citations

  preprintOpen access

  Transcranial focused ultrasound stimulation (TUS) holds promise for non-invasive neural modulation in treating neurological disorders. Most clinically relevant targets are deep within the brain (near or at its geometric center), surrounded by other sensitive regions that need to be spared clinical intervention. However, in TUS, increasing frequency with the goal of improving spatial resolution reduces the effective penetration depth. We show that by using a pair of 1 MHz, orthogonally arranged transducers we improve the spatial resolution afforded by each of the transducers individually, by nearly 40 fold, achieving a sub-cubic millimeter target volume of $0.24\ mm^3$. We show that orthogonally placed transducers generate highly localized standing waves with Acoustic Radiation Force (ARF) arranged into periodic regions of compression and tension near the target. We further present an extended capability of the orthogonal setup, which is to impart selective pressures--either positive or negative, but not both--on the target. Lastly, we share our preliminary findings that strain can arise from both particle motion and ARF with the former reaching its maximum value at the focus, and the latter remaining null at the focus and reaching its maximum around the focus. As the field is investigating the mechanism of interaction in TUS by way of elucidating the mapping between ultrasound parameters and neural response, orthogonal transducers expand our toolbox by making it possible to conduct these investigations at much finer spatial resolutions, with localized and directed (compression vs. tension) ARF and the capability of applying selective pressures at the target.

  PublisherOA PDFDOI

• Focal Volume, Acoustic Radiation Force, and Strain in Two-Transducer Regimes

  arXiv (Cornell University) · 2024-04-21

  preprintOpen access

  Transcranial focused ultrasound stimulation (TUS) holds promise for non-invasive neural modulation in treating neurological disorders. Most clinically relevant targets are deep within the brain (near or at its geometric center), surrounded by other sensitive regions that need to be spared clinical intervention. However, in TUS, increasing frequency with the goal of improving spatial resolution reduces the effective penetration depth. We show that by using a pair of 1 MHz, orthogonally arranged transducers we improve the spatial resolution afforded by each of the transducers individually, by nearly 40 fold, achieving a sub-cubic millimeter target volume of $0.24\ mm^3$. We show that orthogonally placed transducers generate highly localized standing waves with Acoustic Radiation Force (ARF) arranged into periodic regions of compression and tension near the target. We further present an extended capability of the orthogonal setup, which is to impart selective pressures--either positive or negative, but not both--on the target. Lastly, we share our preliminary findings that strain can arise from both particle motion and ARF with the former reaching its maximum value at the focus, and the latter remaining null at the focus and reaching its maximum around the focus. As the field is investigating the mechanism of interaction in TUS by way of elucidating the mapping between ultrasound parameters and neural response, orthogonal transducers expand our toolbox by making it possible to conduct these investigations at much finer spatial resolutions, with localized and directed (compression vs. tension) ARF and the capability of applying selective pressures at the target.

  PublisherOA PDFDOI

• A deep-learning model for one-shot transcranial ultrasound simulation and phase aberration correction

  arXiv (Cornell University) · 2024-10-25

  preprintOpen access

  Transcranial ultrasound (TUS) has emerged as a promising tool in clinical and research settings due to its potential to modulate neuronal activity, open the blood-brain barrier, facilitate targeted drug delivery via nanoparticles, and perform thermal ablation, all non-invasively. By delivering focused ultrasound waves to precise regions anywhere in the brain, TUS enables targeted energy deposition and is being explored in over fifty clinical trials as a treatment for conditions such as opioid addiction, Alzheimer's disease, dementia, epilepsy, and glioblastoma. However, effective TUS treatment requires careful ultrasound parameter design and precise computation of the focal spot's location and pressure, as skull heterogeneity increases the risk of off-target sonication or insufficient energy delivery to neural tissue. In clinical settings, this phase aberration correction must be computed within seconds. To achieve this, commercial devices often rely on faster methods, such as ray tracing, to predict the focus location and pressure. While computationally efficient, these methods may not always provide the high level of accuracy needed for optimal TUS delivery. We present TUSNet, the first end-to-end deep learning approach to solve for both the pressure field and phase aberration corrections without being bound to the inherent trade-off between accuracy and efficiency. TUSNet computes the 2D transcranial ultrasound pressure field and phase corrections within 21 milliseconds (over $1200\times$ faster than k-Wave, a MATLAB-based acoustic simulation package), achieving $98.3\%$ accuracy in estimating peak pressure magnitude at the focal spot with a mean positioning error of only $0.18$ mm compared to ground truth from k-Wave.

  PublisherOA PDFDOI

• Synthetic Skull CT Generation with Generative Adversarial Networks to Train Deep Learning Models for Clinical Transcranial Ultrasound

  arXiv (Cornell University) · 2023 · 2 citations

  • Computer Science
  • Artificial Intelligence
  • Computer Science

  Deep learning offers potential for various healthcare applications, yet requires extensive datasets of curated medical images where data privacy, cost, and distribution mismatch across various acquisition centers could become major problems. To overcome these challenges, we propose a generative adversarial network (SkullGAN) to create large datasets of synthetic skull CT slices, geared towards training models for transcranial ultrasound. With wide ranging applications in treatment of essential tremor, Parkinson's, and Alzheimer's disease, transcranial ultrasound clinical pipelines can be significantly optimized via integration of deep learning. The main roadblock is the lack of sufficient skull CT slices for the purposes of training, which SkullGAN aims to address. Actual CT slices of 38 healthy subjects were used for training. The generated synthetic skull images were then evaluated based on skull density ratio, mean thickness, and mean intensity. Their fidelity was further analyzed using t-distributed stochastic neighbor embedding (t-SNE), Fréchet inception distance (FID) score, and visual Turing test (VTT) taken by four staff clinical radiologists. SkullGAN-generated images demonstrated similar quantitative radiological features to real skulls. t-SNE failed to separate real and synthetic samples from one another, and the FID score was 49. Expert radiologists achieved a 60\% mean accuracy on the VTT. SkullGAN makes it possible for researchers to generate large numbers of synthetic skull CT segments, necessary for training neural networks for medical applications involving the human skull, such as transcranial focused ultrasound, mitigating challenges with access, privacy, capital, time, and the need for domain expertise.

  PublisherOA PDFDOI

• Development and validation of a computational method to predict unintended auditory brainstem response during transcranial ultrasound neuromodulation in mice

  Brain stimulation · 2023 · 9 citations

  • Computer Science
  • Acoustics
  • Audiology

  BACKGROUND: Transcranial ultrasound stimulation (TUS) is a promising noninvasive neuromodulation modality. The inadvertent and unpredictable activation of the auditory system in response to TUS obfuscates the interpretation of non-auditory neuromodulatory responses. OBJECTIVE: The objective was to develop and validate a computational metric to quantify the susceptibility to unintended auditory brainstem response (ABR) in mice premised on time frequency analyses of TUS signals and auditory sensitivity. METHODS: Ultrasound pulses with varying amplitudes, pulse repetition frequencies (PRFs), envelope smoothing profiles, and sinusoidal modulation frequencies were selected. Each pulse's time-varying frequency spectrum was differentiated across time, weighted by the mouse hearing sensitivity, then summed across frequencies. The resulting time-varying function, computationally predicting the ABR, was validated against experimental ABR in mice during TUS with the corresponding pulse. RESULTS: = 0.97). CONCLUSIONS: To reduce ABR in mice during in vivo TUS studies, 1) reduce the amplitude of a rectangular continuous wave envelope, 2) increase the rise/fall times of a smoothed continuous wave envelope, and/or 3) change the PRF and/or duty cycle of a rectangular or sinusoidal pulsed wave to reduce the gap between pulses and increase the rise/fall time of the overall envelope. This metric can aid researchers performing in vivo mouse studies in selecting TUS signal parameters that minimize unintended ABR. The methods for developing this metric can be adapted to other animal models.

  PublisherOA PDFDOI

• Transcranial ultrasound neuromodulation of the thalamic visual pathway in a large animal model and confirmation of target engagement with MR-ARFI

  Brain stimulation · 2023-01-01

  articleOpen access
  PublisherDOI

• Development and Validation of a Computational Method to Predict Peripheral Auditory Activation During Transcranial Ultrasound Neuromodulation in Mice

  SSRN Electronic Journal · 2023-01-01

  preprintOpen access
  PublisherDOI

• Transcranial ultrasound stimulation: considerations for pulse shaping

  Brain stimulation · 2023-01-01 · 2 citations

  articleOpen accessSenior author
  PublisherDOI

Recent grants

• Administrative Core

  NIH · $5.0M · 2018

• NIH Grant R42DC003614

  NIH · $498k · 2004

Frequent coauthors

• Eunice L.-M. Cheung

  Harvard University

  25 shared

• Mark J. Schnitzer

  Stanford University

  20 shared

• Juergen C. Jung

  Stanford University

  20 shared

• José E. Barrera

  The University of Texas Health Science Center at Houston

  19 shared

• Robert H. Margolis

  16 shared

• Andrew B. Holbrook

  Stanford University

  16 shared

• Nikolas H. Blevins

  Stanford University

  14 shared

• Kim Butts Pauly

  14 shared