SPatial REgression Analysis of Diffusion tensor imaging (SPREAD) for longitudinal progression of neurodegenerative disease in individual subjects

Objectives: To develop a novel statistical method for analysis of longitudinal DTI data in individual subjects. Materials and Methods: The proposed SPatial REgression Analysis of Diffusion tensor imaging (SPREAD) method incorporates a spatial regression fitting of DTI data among neighboring voxels and a resampling method among data at different times. Both numerical simulations and real DTI data from healthy volunteers and multiple sclerosis (MS) patients were used in the study to evaluate this method. Results: Statistical inference based on SPREAD was shown to perform well through both group comparisons among simulated DTI data of individuals (especially when the group size is smaller than 5) and longitudinal comparisons of human DTI data within the same individual. Conclusions: When pathological changes of neurodegenerative diseases are heterogeneous in a population, SPREAD provides a unique way to assess abnormality during disease progression at the individual level. Consequently, it has the potential to shed light on how the brain has changed as a result of disease or injury.