Modeling Dynamic Functional Connectivity with Latent Factor Gaussian Processes

TL;DR

This paper introduces a latent factor Gaussian process model that effectively captures and visualizes dynamic brain connectivity from noisy neuroimaging data, demonstrated through rat brain activity during a memory task.

Contribution

The paper proposes a novel latent factor Gaussian process model that improves inference and visualization of time-varying connectivity in high-dimensional neuroimaging data.

Findings

Evidence of stimuli differentiation in rat brain activity

Effective modeling of high-dimensional, noisy neuroimaging data

Enhanced inference of dynamic connectivity patterns

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.