Longitudinal neuroimaging data plays an important role in mapping the neurodevelopmental profiles of major neuropsychiatric and neurodegenerative disorders and the normal brain. The generation of such developmental maps is critical for the prevention, diagnosis, and treatment of many brain-related diseases. The aim of this paper is to develop a spatio-temporal Gaussian process (STGP) framework to accurately delineate the developmental trajectories of brain structure and function, while achieving better prediction by explicitly incorporating the spatial and temporal features of longitudinal neuroimaging data. Our STGP integrates a functional principal component model (FPCA) and a partition parametric space-time covariance model to capture the medium-to-large and small-to-medium spatio-temporal dependence structures, respectively. We develop a three-stage efficient estimation procedure as well as a predictive method based on a kriging technique. The STGP has two key novelties: (1) it can efficiently use a small number of parameters to capture complex non-stationary and non-separable spatio-temporal dependence structures, and (2) it can accurately predict spatio-temporal changes. We illustrate the STGP by using simulated data sets and two real data analyses-longitudinal positron emission tomography data from the Alzheimer's Disease Neuroimaging Initiative and longitudinal lateral ventricle surface data from a longitudinal study of early brain development.
Hyun JW, Li Y, Huang C, Styner M, Lin W, Zhu H. STGP: Spatio-temporal Gaussian process models for longitudinal neuroimaging data. NeuroImage 134:550-562, 2016. NIHMSID: NIHMS778171 PubMed [journal] PMID: 27103140, PMCID: PMC4912881.