Investigating the Impact of Genetic Background on Brain Dynamic Functional Connectivity Through Machine Learning: A Twins Study

Functional magnetic resonance imaging (fMRI) is a popular approach for understanding the functional connectivity of human brain. Recently, dynamic functional connectivity has been used to analyze connectivity variations on resting state fMRI. Here, we use task based fMRI (using the Poffenberger Paradigm) data collected in mono- and dizygotic twin pairs. The task is to examine if the two groups of twins can be discriminated by using the dynamic connectivity, so to prove that genetic background has an effect on functional connectivity. To this aim, we have explored the dynamic connectivity patterns of task-relevant and task-orthogonal sub-networks using graph Laplacian representation in combination with a metric defined on the space of covariance matrices to compute the similarity between twins' dynamics in the mental state. Linear SVMs with an unsupervised feature selection (Laplacian Score) were then used to discriminate the two classes of twins.