Multicenter reliability of electric field simulations: Evidence from a traveling-subjects study

Abstract

Multicenter neuroimaging studies face challenges from scanner-related measurement biases that can obscure biological effects. Traveling-subject (TS) designs allow direct quantification of such biases by scanning the same participants across multiple sites. While reliability of structural, functional, and diffusion MRI has been established, the reliability of electric field simulations for transcranial brain stimulation has not been examined. We assessed inter-scanner and scan-rescan reliability of simulated electric field magnitudes in 10 participants scanned twice on each of five 3T Siemens scanners, using SimNIBS v4.1 to simulate focal tDCS across seven cortical and cerebellar targets. ICC values indicated good to excellent inter-scanner (0.86-0.97) and scan-rescan (0.94-0.98) reliability, with between-scanner variance not exceeding measurement error. Intra-individual segmentation variability substantially explained within-person differences in field magnitudes, whereas image quality did not predict segmentation variability. These results demonstrate that individualized electric field simulations are robust to scanner-related variation, supporting their use in multicenter brain stimulation research.