Modeling Dynamic Functional Connectivity with Latent Factor Gaussian Processes

Abstract

Dynamic functional connectivity, as measured by the time-varying covariance of neurological signals, is believed to play an important role in many aspects of cognition. While many methods have been proposed, reliably establishing the presence and characteristics of brain connectivity is challenging due to the high dimensionality and noisiness of neuroimaging data. We present a latent factor Gaussian process model which addresses these challenges by learning a parsimonious representation of connectivity dynamics. The proposed model naturally allows for inference and visualization of time-varying connectivity. As an illustration of the scientific utility of the model, application to a data set of rat local field potential activity recorded during a complex non-spatial memory task provides evidence of stimuli differentiation.

References in corpus (5)

• Rates of contraction of posterior distributions based on Gaussian process priors
• Kernel Interpolation for Scalable Structured Gaussian Processes (KISS-GP)
• Reproducing kernel Hilbert spaces of Gaussian priors
• Estimating Time-Varying Effective Connectivity in High-Dimensional fMRI Data Using Regime-Switching Factor Models
• Nonparametric Modeling of Dynamic Functional Connectivity in fMRI Data

Cited by in corpus (1)

• Deep Representations for Time-varying Brain Datasets