Influence of measurement and processing conditions of resting state fMRI on functional connectivity

Resting state functional MRI（rsfMRI）is widely used to examine brain disorders. Various measurement conditions are used for rsfMRI because of the current lack of parameter standardization. Functional connectivity measured by rsfMRI might be used for the diagnosis of brain disorders, such as dementia, autism, depression, and schizophrenia. However, physical and physiological noises affect functional connectivity signals. To remove noises, signal filtering is routinely performed before connectivity analysis. Here, we examined the effects of filtering, particularly of bandwidth, and measurement duration, on connectivity artifacts. We found that filtering bandwidth produced artifactual connectivity, with narrow bandwidths generating more artifacts. Therefore, to reduce connectivity artifacts produced by filtering of rsfMRI signals in the preprocessing procedure, the threshold of correlation values for functional connectivity analysis should be defined according not only to the number of volumes and measurement duration but also to the bandwidth used in preprocessing.