MAGNETIC RESONANCE IMAGING BASED FUNCTIONAL CONNECTIVITY METHODS

Abstract

Functional connectivity analyses based on functional Magnetic Resonance Imaging (fMRI) data have gained an important place in brain research. There are alternative functional connectivity estimation approaches, which, despite the similarity of the overall results, produce significant differences in their details. For effective use of the functional connectivity metrics, the strengths and weaknesses of various approaches need to be well understood. While the seed-based functional connectivity analyses based on the selection of those anatomic regions of interest derived from the literature represent a stronger approach for hypothesis testing, the independent component analysis (ICA) as a data-driven approach provides an unbiased evaluation possibility for exploratory data analysis. Another difference between the methods is related to group analyses in terms of registering individual brains to a common template or implementing anatomical definitions on the spatial coordinates of individual brains. While the latter increases the success in studies on pathologies that lead to large-scale brain deformations, the former may be advantageous for deriving normative results from large data sets. Lastly, volume vs surface-based approaches for the definition of cortical anatomy in the individual space also significantly affect the results of functional connectivity analyses. In this review, functional connectivity estimation methods will be compared by evaluating them using these three perspectives.