Within-subject template estimation for unbiased longitudinal image analysis

► We introduce unbiased longitudinal processing of brain MRI of several time points. ► We demonstrate that inerpolation asymmetries are not the only source of bias. ► We create a robust within-subject template to initialize all time points. ► Reliability is significantly increased, while over-regularization is avoided. ► Precision allows for smaller sample sizes in clinical trials to assess biomarkers.