Factors Contributing to Short-Term Structural Variability in a Longitudinal MRI Dataset

When planning longitudinal magnetic resonance imaging (MRI) studies, it is advisable to consider various (confounding) factors that could influence brain structural changes over time. The goal of this study was to identify factors that contribute to intraindividual variability of brain structure within a short period of time. We employed multilevel sparse partial least squares regression to investigate the changes in regional gray matter volume in the longitudinal Day2day MRI dataset. The findings suggest that the changes in regional GM volume estimations were primarily driven by image quality, while the outdoor temperature and time since baseline appeared as the main predictors of volumetric changes in insular and diencephalic brain regions. We additionally investigated factors associated with variability in image quality. The findings underscore the importance of maintaining adequate participant arousal during scanning.