Reliability of electric field simulations in different age groups: impact of data quality metrics.

Abstract

Electric field simulations based on individualized Magnetic Resonance Imaging (MRI)-derived head models are increasingly used to optimize non-invasive brain stimulation (NIBS) protocols, enabling individualized dose adjustments to achieve a desired current flow at cortical targets. However, the quality of structural MRI data-affected by motion artefacts, tissue contrast, and scanner-related noise-may influence the accuracy of these simulations, an issue possibly exacerbated in older adults who often exhibit lower image quality. If reduced image quality compromises electric field estimation, it could limit the feasibility of individualized dosing in aging populations. In this study, we examined whether standardized image quality metrics-entropy focus criterion (EFC), signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), spatial resolution (FWHM), and intensity non-uniformity (INU)-systematically relate to the magnitude of simulated electric fields in young and older adults. We analysed MRI and simulation data of a focal C3 montage from 106 healthy adults, 47 young adults (mean age: 24.8, age range: 20-35 years) and 59 older adults (mean age: 69.5 years, age range: 60-79) using SimNIBS for computational modelling of the electric fields and MRIQC for image quality assessment. Structural equation modelling was used to quantify direct and indirect effects of age group on electric field magnitude, with image quality as a mediating factor. Head volume was included in extended models to control for anatomical variation. Our findings demonstrate that MR image quality is associated with simulated electric field magnitude, with higher EFC, lower SNR, lower CNR, and higher INU associated with reduced field estimates, and older adults showing generally lower simulated electric fields compared to younger adults. While image quality accounted for some of the age-related differences in electric field strength, group differences remained even after controlling for EFC, SNR and head volume. Critically, electric field simulations remained sufficiently reliable despite lower image quality in older adults, supporting their use for individualized dose adjustment across the lifespan.