MRI-FREE PERSONALIZATION OF NON-INVASIVE BRAIN STIMULATION PARAMETERS USING PROXY HEAD MODEL GROUPS

Non-invasive brain stimulation (NIBS) modulates neural activity by generating an intracranial electric field (dose) and is used as a therapeutic approach for neuropsychiatric disorders. Because the delivered dose depends strongly on inter-individual anatomical variability, computational modeling is widely used to estimate and optimize stimulation. Standard computational workflows rely on participant MRI to build individualized head models and to select stimulation parameters that maximize the on-target electric field while minimizing off-target stimulation. However, MRI-based personalization is limited by cost and accessibility. Here, we propose an MRI-less optimization framework based on representative anatomical head models (proxy models). Proxy models are selected based on subject characteristics (e.g., age) and used to optimize stimulation at the group level. Using open MRI datasets to generate young and elderly proxy models (each of 50 participants), we optimized to maximize on-target versus off-target electric-field magnitude across 50 test subjects. We evaluated this methodology in two NIBS modalities: electrode montage optimization for transcranial electrical stimulation targeting deep brain tissue and coil placement optimization for transcranial magnetic stimulation targeting the hand motor area. Representative proxy models achieved targeting performance comparable to individualized optimization, with differences on the order of 15% in electric-field intensity and focality, with particularly better performance in older adults. In conclusion, these results demonstrate the feasibility of MRI-less NIBS optimization and suggest that combining subject characteristics with age-appropriate optimized configurations may improve prediction accuracy, particularly in older populations.