The impact of CT image parameters and skull heterogeneity modeling on the accuracy of transcranial focused ultrasound simulations

Abstract Objective . Low-intensity transcranial ultrasound stimulation (TUS) is a promising non-invasive brain (NIBS) technique. TUS can reach deeper areas and target smaller regions in the than other NIBS techniques, but its application humans hampered by lack of straightforward reliable procedure to predict induced exposure. Here, we examined how skull modeling affects computer simulations TUS. Approach We characterized ultrasonic beam after transmission through sheep with hydrophone performed computed tomography (CT) image-based experimental setup. To study model’s impact, varied: CT acquisition parameters (tube voltage, dose, filter sharpness), image interpolation, segmentation parameters, acoustic property maps (speed-of-sound, density, attenuation), transducer-position mismatches. compared impact parameter changes on model predictions measurement agreement. Spatial-peak intensity location, total power, Gamma metric (a measure for distribution differences) were used as quantitative criteria. Modeling-based sensitivity analysis was also two human head models. Main results Sheep attenuation assignment transducer positioning had most important spatial peak (overestimation up 300%, respectively 30%), followed sharpness tube voltage (up 20%), requiring calibration mapping functions. Positioning skull-heterogeneity-structure strongly affected (gamma tolerances exceeded > 80%, 150%, focus-volume water), necessitating personalized modeling. Simulation models consistently demonstrate high skull-heterogeneity model, tuning, shifts, magnitude which depends underlying structure complexity. Significance Our reveals importance properly accurately reproducing position. The raise red flags when translating approaches among clinical sites without proper standardization and/or recalibration imaging parameters.