Causal connections in the brain: from functional to causal connectivity

This chapter provides an introduction to causal discovery analysis (CDA) for applied neuroimaging researchers, and a reference for best practices in applying CDA to functional magnetic resonance imaging (fMRI) data. The discovery of correlated fluctuations in fMRI has changed how we think about cognition and has spurred an understanding of the brain as an interconnected network. Functional connectivity relies on undirected measures of connectivity, however, and the field is increasingly aware of the need for measures of brain causal (directed) connectivity. Here, we discuss the application of a class of algorithms designed for statistically assessing causality, CDA, to brain data. CDA can be applied to other types of brain data as well, but this chapter focuses on functional neuroimaging data. We discuss the formal background of CDA methods, challenges inherent in applying these methods to fMRI data, and previous applications of these methods in the literature.