Title: Transcranial Ultrasound Stimulation: Optimizing Simulation Paradigms and Modulation of the Neurons

Abstract: Transcranial focused ultrasound stimulation (TUS) is a promising, non-invasive method for neural modulation in treating neurological disorders. Traditional TUS faces challenges in targeting deep brain regions due to the inverse relationship between frequency and penetration depth. Our research demonstrates that orthogonally arranged transducers at 1 MHz can significantly enhance spatial resolution, achieving a sub-cubic millimeter target volume of 0.24 mm3. This configuration produces highly localized standing waves with Acoustic Radiation Force (ARF) near the target, allowing for precise application of direction-specific forces and selective pressures. These advancements enable detailed investigations of ultrasound-neural interactions and hold potential for improved therapeutic outcomes. 

Deep learning applications in healthcare, particularly for transcranial ultrasound, require extensive datasets of curated medical images, which are often scarce due to privacy, cost, and distribution issues. To address this, we developed SkullGAN, a generative adversarial network that generates synthetic skull CT slices for training models. By training SkullGAN on real CT slices, we created high-fidelity synthetic images validated through multiple metrics and expert evaluations. Complementing this, we introduce TUSNet, a deep learning approach that precisely computes the focal spot's location and pressure in TUS, overcoming skull heterogeneity challenges. TUSNet provides accurate phase aberration corrections and pressure field computations within seconds, achieving results 1000 times faster than conventional methods. With 98.3% accuracy in estimating peak pressure values and minimal positioning error, TUSNet significantly enhances the precision and efficiency of TUS applications, supporting ongoing clinical trials and research in treating various neurological conditions.

Please contact Madelyn Bernstein for the Zoom link