Alterations in Dynamic Functional Connectivity in Individuals With Subjective Cognitive Decline

Abstract

Purpose: To investigate the dynamic functional connectivity (DFC) and static parameters of graph theory in individuals with subjective cognitive decline (SCD) and the associations of DFC and topological properties with cognitive performance. Methods: Thirty-three control subjects and 32 SCD individuals were enrolled in this study, and neuropsychological evaluations and resting-state functional magnetic resonance imaging scanning were performed. Thirty-three components were selected by group independent component analysis to construct 7 functional networks. Based on the sliding window approach and k-means clustering, distinct DFC states were identified. We calculated the temporal properties of fractional windows in each state, the mean dwell time in each state, and the number of transitions between each pair of DFC states. The global and local static parameters were assessed by graph theory analysis. The differences in DFC and topological metrics, and the associations of the altered neuroimaging measures with cognitive performance were assessed. Results: The whole cohort demonstrated 4 distinct connectivity states. Compared to the control group, the SCD group showed increased fractional windows and an increased mean dwell time in state 4, characterized by hypoconnectivity both within and between networks. The SCD group also showed decreased fractional windows and a decreased mean dwell time in state 2, dominated by hyperconnectivity within and between the auditory, visual and somatomotor networks. The number of transitions between state 1 and state 2, between state 2 and state 3, and between state 2 and state 4 was significantly reduced in the SCD group compared to the control group. No significant differences in global or local topological metrics were observed. The altered DFC properties showed significant correlations with cognitive performance. Conclusion: Our findings indicated DFC network reconfiguration in the SCD stage, which may underlie the early cognitive decline in SCD subjects and serve as sensitive neuroimaging biomarkers for the preclinical detection of individuals with incipient Alzheimer's disease.

Keywords: dynamic functional connectivity; fractional windows; graph theory; independent component analysis; subjective cognitive decline.

Figure 1

Independent components (n = 33) identified by group independent component analysis. (A) Independent component spatial maps divided into seven functional networks. (B) Group averaged static functional connectivity matrix between pairs of independent components. BG, basal ganglia; AUD, auditory; VIS, visual; SMN, sensorimotor; CEN, cognitive executive; DMN, default mode; CB, cerebellar.

Figure 2

The four states identified by k-means clustering analysis and the corresponding cluster centroids. The total number and percentage of the reoccurrence times of each state are listed above each cluster (A), and the 5% strongest connections of each state are shown (B). BG, basal ganglia; AUD, auditory; VIS, visual; SMN, sensorimotor; CEN, cognitive executive; DMN, default mode; CB, cerebellar.

Figure 3

The four dynamic functional connectivity patterns of the two groups. (A) The centroid matrices for the normal controls. (B) The centroid matrices for the subjective cognitive decline participants.

Figure 4

Temporal properties of dynamic functional connectivity states between the two groups. (A) Fractional windows in each state. (B) Mean dwell time in each state. (C) The number of transitions between pairs of states. The parameters of each individual in the normal control (NC) and subjective cognitive decline (SCD) groups are presented in blue and khaki dots respectively. The black lines indicate the mean values, and the light gray rectangles cover the data within one standard error above and below the mean. *p < 0.05; **p < 0.01.