Pseudo resting-state by task-evoked functional connectivity.

Functional connectomics investigates how brain regions are functionally associated. Studies in literature seek to infer a “pseudo-resting” state from task data, however it should be tested whether resting-state-like connectivity can be inferred by task-fMRI data. This work investigates several preprocessings of task-evoked connectivity performing classification experiments, to test their ability to reproduce a real “resting-state” connectivity and their impact on a clinical context, namely the comparison of Typical Readers and Developmental Dyslexics. Our results suggest that a task-free “pseudo-resting” connectivity cannot be inferred from task-fMRI data and that signal preprocessing does not influence the way the classification rules are inferred.