Dynamic Functional Connectivity Analysis And Its Application In Subjective Cognitive Decline Based On Resting-state FMRI

Functional connectivity(FC)analysis based on resting-state functional magnetic resonance imaging(rs-fMRI)has been widely used to study the mechanism and to diagnose mental and neurological diseases.In recent years,researchers have found that dynamic FC(dFC)can capture a series of dynamic states of the brain and reveal the complex characteristics of brain networks in a more comprehensive way.Many clinical studies have shown that dFC can provide new evidence for the early diagnosis of brain diseases.But current methods for identification of brain dynamic states only consider the FC dynamics,and they overlook the local spontaneous brain activity dynamics,such as the fractional amplitude of low frequency fluctuation(fALFF).Therefore,we have little knowledge about the relationship between dynamic patterns of FC and local spontaneous neural activity and it still remains unclear whether such a relationship is linked to brain diseases.This paper proposed a new method for analyzing the dynamic brain states by considering dynamic patterns in both FC and local spontaneous activity.First,the method uses dFC and dynamic fALFF(dfALFF)as features to characterize two kinds of dynamic states: dFC states and dfALFF states.Second,a new feature,co-occurrence frequency,is proposed to describe the occurrence rate of one pair of dfALFF state and dFC state among all time points for each participant.Furthermore,we employed ensemble clustering techniques to combine dFC and dfALFF for identifying a new type of brain dynamic state,joint dFC-dfALFF state.The proposed new method was applied to a clinical rs-fMRI dataset to search for differences of dynamic brain states in patients with subjective cognitive decline(SCD)and mild cognitive impairment(MCI).SCD and MCI are early stages of Alzheimer’s disease(AD),so the identification of neural biomarkers of SCD and MCI are crucial for early diagnosis of these pre-AD diseases.Results showed that,(1)the occurrence frequency of a default mode network(DMN)-dominated dFC state was significantly different between the healthy elderly and SCD as well as between the healthy elderly and MCI,(2)the co-occurrence frequency of this DMNdominated dFC state and an dfALFF state was significantly different between SCD and MCI as well as between the healthy elderly and MCI,(3)the occurrence frequency of a DMNdominated joint dFC-dfALFF state was significantly different between the healthy elderly and SCD as well as between the healthy elderly and MCI.These results collectively suggested that DMN plays a key role in the development of AD,and SCD can be considered as a transition stage between normal aging and MCI.In conclusion,we proposed a new framework to study dynamic brain states by taking into account both regional synchronous neural activity(dFC)and local spontaneous neural activity(dfALFF),providing a more comprehensive description of brain dynamics.This new framework could be used to analyze abnormal brain dynamics in patients with mental and neurological diseases.As a consequence,it can improve the understanding of the neural mechanism of brain diseases and can provide new neural markers for diagnosis of brain diseases.