|
Abstract:
|
In many cases, functional magnetic resonance imaging (fMRI) data is collected in a longitudinal manner and it is of interest to measure or account for a trend in the functional connectivity across time. In this work, we build a longitudinal functional connectivity model using a variance components approach. First, for all subject's visits, we account for the autocorrelation inherent in the fMRI time series data using a non-parametric technique. Second, iterated weighted least squares is used to estimate the between subject, and the within subject between visit variances. Finally, we estimate the connectivity network and longitudinal trend(s) using least squares. Our novel method seeks to account for the within subject dependence across multiple visits while restricting the number of parameters in order to make the method computationally feasible and stable. Model performance is examined in a series of simulations and through an application to longitudinal resting-state fMRI data from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database.
|
