A COMPUTATIONAL APPROACH TO MODELING FOCUSED ULTRASOUND EFFECTS ON PERIPHERAL NERVES

Focused ultrasound shows significant potential for modulating peripheral nerve activity, yet the mechanisms of action and the optimal stimulation parameters remain unclear.
We present an automated computational pipeline (with supporting experimental data) that predicts the effect of ultrasound stimulation on axonal structures in realistic biophysical environments and enables systematic exploration of putative underlying interaction mechanisms (membrane bending, intramembrane cavitation, flexoelectricity, mechanosensitive ion channel, and thermomodulation) across a wide range of stimulation parameters.
Our approach leverages detailed, histology-based nerve models (fascicles, perineurium, epineurium, connective tissue) with axonal trajectories (AxonDeepSeg), and integrates them either in an anatomical rodent model, or in a lab-on-a-chip setup. Sim4Life’s multiphysics solvers for acoustics, electromagnetism, and induced heating, or optionally a kWave solver plug-in for elastic wave simulation, subsequently determine the spatial distribution of acoustic pressure and radiation force, associated stress and tissue heating, as well as the lead-field transfer matrix for neural sensing. Mechanical and thermal exposure serve as input to biophysical models of hypothesized interaction mechanisms to predict fiber recruitment. Using reciprocity, axonal activity is directly linked to the corresponding compound action potential signals (CAP) and compared with in vivo and ex vivo (stimulating explanted rat spinal root) CAP measurements to identify signatures of relevant interaction mechanisms. This end-to-end pipeline is fully realized on the o²S²PARC platform for collaborative and FAIR computational modeling, ensuring reproducibility and accessibility of the results.
By advancing mechanistic understanding and guiding protocol optimization, our pipeline supports the development of effective FUS-based neuromodulation strategies for both research and clinical applications.