Machine Learning-Assisted Optimization Framework for Single-Element Transcranial Focused Ultrasound Transducer Design...

Transcranial focused ultrasound (t-FUS) is an emerging noninvasive neuromodulation technique with high spatial precision and deep brain penetration. Due to skull-induced attenuation and acoustic aberrations, precise stimulation of deep brain regions in mice remains challenging compared to shallow brain regions.

To develop a machine-learning-assisted framework for single-element transducer design and demonstrate surrogate-assisted optimization of transducer frequency and geometry for deep-brain targeting in mice.

This framework includes a surrogate model consisting of a Random Forest regressor and classifier, trained on acoustic simulation results to predict performance from design parameters. A total of 72 transducer designs were simulated across coronal and sagittal planes, systematically varying frequency (1-6 MHz), radius of curvature (5-7 mm) and f-number (0.58-1.0). Each design was evaluated using five performance metrics: focal length, focal shape, maximum pressure at the focal region, pressure maximum location and sidelobe suppression. The surrogate models were then combined with the Non-Dominated Sorting Genetic Algorithm II to perform multi-objective optimization and identify high-performing transducer designs.

The optimized design produced a compact, symmetric focal region and accurate energy delivery to deep targets, with minimal off-target exposure, even in complex skull anatomy. Results show that lower f-numbers and higher frequencies facilitate precise deep-brain targeting in mice.

This study introduces a machine-learning-assisted framework for low-cost, surrogate-based optimization of single-element t-FUS transducer designs, reducing reliance on heuristic parameter selection. The surrogate was trained on simulations from a single CT-derived mouse head; expanding to additional CT-derived anatomies should improve generalizability across strains, ages and sexes.