Abstract: Neuroimaging has mainly depended on cross-sectional data to study the brain through the lifespan, but these studies can only relate brain differences to intersubject differences in age. Only longitudinal neuroimaging studies can infer on age-induced changes in the brain, crucial for studies of developmental and aging. In this talk I will provide an overview of longitudinal modelling techniques for neuroimaging data, elaborating on why the standard neuroimaging tools (SPM & FSL) are generally not up to the task, and highlight a series of work in my group on modelling longitudinal neuroimaging data. Our most basic approach is a GEE with independence working covariance and robust standard errors, which provides a very fast and practical approach for which we have created a user-friendly implementation (SwE). While standard statistical tools (e.g. R's lme4/nlme) are computationally efficient and robust, they only fit one outcome at a time. We have developed a highly optimised linear mixed effects (LME) implementation that exploits vectorised computation so that all voxels are simultaneously updated at each iteration; combined with HPC (SGE/SLURM) integration, this approach makes fitting LME's possible for arbitrarily large datasets (BigLMM). We have also developed an approach that is faster than LME but more statistically efficient than GEE: by combining a covariance regression that avoids iteration and variance parameter discretization that maximises vectorisation, we have a flexible, fast and sensitive LME implementation (FEMA). Finally, for longitudinal binary images, e.g. lesion masks of white matter hyperintensities, we propose a relative-risk regression to support user's preference for relative risk (RR) units instead of odds-ratios; since a conventional RR regression with log-link and binomial variance function can be very unstable, we use a GEE approach with log-link, identity variance function, and unknown dispersion parameter along with a penalty to avoid infinite parameter estimates. This suite of work is a small indication of the rich methodological opportunities for the growing volume of longitudinal neuroimaging studies.