Basic Properties Of Site Effect For Amplitude Of Low Frequency Fluctuation In Resting-state FMRI Signal

Amplitude of low frequency fluctuation(ALFF)has been widely used in resting-state functional Magnetic Resonance Imaging(fMRI)to investigate the local spontaneous activity of human brain.To increase statistical power and improve the generalization of conclusion,more and more researchers focus on multi-site studies.It is necessary to perform the global mean normalization(GMN)for ALFF in order to compare the result which come from different sites.However,mathematically,the GMN is not so perfect,it may diminish the true underlying effects and introduce the artificial results.What's worse,the rationality and the necessity of GMN is still unclear.Recently,some research has shown that it can remarkably reduce the degree of variation of ALFF among different sites,but the reason remains unknown.In this paper,we used 1000 Functional Connectomes Project(FCP)to reveal the basis feature of site effect of ALFF,then,try to reveal the mechanism of the GMN to reduce the site effect of ALFF.In the first study,we revealed the resource of site effect of ALFF from spatial perspective.As shown in the result,there are apparent region-specific distribution for site effect of ALFF.In particular,the contribution of white matter is much larger than gray matter in the brain.However,the tendency between white matter and gray matter of different sites are basically consistent.This suggests that the site effect for ALFF of white matter and gray matter maybe have the same origin.In the second study,we used fractional amplitude of low-frequency fluctuation(fALFF)to find the resource of site effect for ALFF from frequency perspective.The site effect for ALFF is not low-frequency specific but it reflects the character of whole frequency band.Finally,we found that the site effect of amplitude measurement in rest state fMRI are closely related to the feature of blood oxygen level dependent(BOLD)in the non-functional regions,which provides evidence and insights for developing signal level de-noise methods and standardization in the future.