Focused ultrasound is increasingly adopted as a non-invasive neuromodulation modality for brain circuits.
However, its impact on axons, crucial for applications involving peripheral neuropathies and long-range brain connectivity, remains poorly characterized. In particular, current evidence of direct effects is inconsistent, with both excitatory and inhibitory outcomes observed under comparable stimulation conditions, and with the underlying biophysical mechanisms still unclear. This limits both mechanistic interpretability and translation to clinical application.
This thesis presents a set of complementary studies with the common goal of characterizing and understanding the potential of focused ultrasound stimulation for peripheral and spinal circuits. These studies are positioned along different stages of the translational pathway, ranging from biophysical computational modeling to experimental investigations in animal models, and ultimately to human applications. Across these stages, the models progressively increase physiological realism while necessarily trading experimental controllability.
The thesis opens by positioning ultrasound neuromodulation within the broader landscape of neuromodulation technologies.
The physical foundations of ultrasound stimulation are then established and the current state of the field is reviewed, identifying key challenges and open questions, with a specific focus on the interaction between ultrasound and axonal structures.
The second chapter presents a computational biophysical framework to model intramembrane cavitation in multi-compartmental, morphologically realistic neuronal representations and uses it to investigate the response of myelinated and unmyelinated axons to ultrasound fields. A further step consists of building an end-to-end in silico pipeline to predict ultrasound-evoked nerve activity in a realistic environment.
The third part of the thesis focuses on experimental models to investigate the direct effects of ultrasound on axonal structures. A microfabricated nerve-on-a-chip platform enables controlled sonication and multi-site recordings from isolated nerve fascicles, while an in vivo rodent model provides physiological realism and motor outputs. This work characterizes ultrasound-evoked compound action potentials and muscular responses, quantifies parameter-dependent effects, and compares ultrasound and electrical stimulation in terms of recruitment and temporal dynamics.
The fourth chapter moves toward translation to sensory feedback applications, focusing on human stimulation paradigms. This part systematically explores the stimulation parameter space, defines the evoked sensation, and models the perception thresholds and intensity.
The thesis concludes with a discussion of the main contributions, ongoing research, and future directions, highlighting how biophysical modeling and controlled experimental studies at different translational levels can complement each other and support th
