Rich club disturbances of the human connectome from subjective cognitive decline to Alzheimer's disease

Abstract

Alzheimer's disease (AD) has a preclinical phase that can last for decades prior to clinical dementia onset. Subjective cognitive decline (SCD) is regarded as the last preclinical AD stage prior to the development of amnestic mild cognitive decline (aMCI) and AD dementia (d-AD). The analysis of brain structural networks based on diffusion tensor imaging (DTI) has identified the so-called 'rich club', a set of cortical regions highly connected to each other, with other regions referred to as peripheral. It has been reported that rich club architecture is affected by regional atrophy and connectivity, which are reduced in patients with aMCI and d-AD. Methods: We recruited 62 normal controls, 47 SCD patients, 60 aMCI patients and 55 d-AD patients and collected DTI data to analyze rich-club organization. Results: We demonstrated that rich club organization was disrupted, with reduced structural connectivity among rich club nodes, in aMCI and d-AD patients but remained stable in SCD patients. In addition, SCD, aMCI and d-AD patients showed similar patterns of disrupted peripheral regions and reduced connectivity involving these regions, suggesting that peripheral regions might contribute to cognitive decline and that disruptions here could be regarded as an early marker of SCD. This organization could provide the fundamental structural architecture for complex cognitive functions and explain the low prevalence of cognitive problems in SCD patients. Conclusions: These findings reveal a disrupted pattern of the AD connectome that starts in peripheral regions and then hierarchically propagates to rich club regions, when patients show clinical symptoms. This pattern provides evidence that disruptions in rich club organization are a key factor in the progression of AD that can dynamically reflect the progression of AD, thus representing a potential biomarker for early diagnosis.

Keywords: Alzheimer's disease; DTI; graph theory; rich club; subjective cognitive decline.

Figure 1

Rich club functions of FABIRC-weighted group networks. The figures show (A) rich club coefficients and (B) normalized rich club coefficients for a range of ks. The graph shows the association between the mean (standard error) Ø_norm as a function of node degree (k) for each of the groups. The differences between NC and patient groups emerge as the node degree increases (N=224). *P<0.05, ** P<0.01, *** P<0.001. Normalized rich club coefficients were larger than 1, suggesting rich club organization in all groups.

Figure 2

Rich club regions of all groups. (A) Rich club members (red nodes) across all healthy and patient groups (N=224). (B) A simplified example of the three classes of connections: rich club connections, linking two rich club nodes; feeder connections, linking one rich club node to one peripheral node; and local connections, linking two peripheral nodes.

Figure 3

Rich club organization during disease progression. Group differences in rich club network properties are displayed. Bar graphs display the mean (standard error) age- and gender-adjusted connectivity strengths for (A) rich club, (B) feeder and (C) local (N=224). *P<0.05, ** P<0.01, *** P<0.001. Scatter plots show the relationship between the feeder connectivity strength and AVLT-delayed recall score (age-, gender- and education-corrected, after Bonferroni corrections for the number of cognitive test variables) for the (D) NC group, (E) SCD group, (F) aMCI group, and (G) d-AD group. The solid lines show the best-fitting linear regression line and the 95% confidence intervals.

Figure 4

The aberrant connections in diagnostic groups relative to NC: (A) SCD patients and NC, (B) aMCI patients and NC, and (C) d-AD patients and NC. Red edges indicate affected rich club connections, purple edges indicate affected feeder connections, and green edges indicate affected local connections. The classification of rich club nodes and non-rich club nodes is depicted by the inner ring (gray palette, with black squares indicating rich club nodes and gray ones indicating non-rich club nodes). (D) Proportion (%, y-axis) of significantly altered connections (100% × observed/expected) illustrated by rich club, feeder and local edges.

Figure 5

Group differences in network topological metrics during disease progression. Bar graphs display the mean (standard error) of the age- and gender-corrected metrics of (A) strength, (B) clustering coefficient, (C) characteristic path length, (D) normalized clustering coefficient and (E) normalized characteristic path length (N=224). *P<0.05, ** P<0.01, *** P<0.001.

Figure 6

Whole-brain structural connectivity of nodes with the highest number of aberrant connections in SCD individuals relative to NC. (A-C) Nodes in the blue box (CAU.L, CAU.R, ORBmid.L) were those with the highest number of aberrant connections in SCD individuals relative to NC. The connections displayed are those that connect with the CAU.L, CAU.R, and ORBmid.L. The bar graphs display the mean (standard error) of the age- and gender-corrected nodal efficiency values of (D) CAU.L, (E) ORBmid.L and (F) CAU.R for each group.

Figure 7

The scatter plots illustrate the significant associations between CAU.L nodal efficiency and AVLT-delayed recall scores controlled for age, gender and education after Bonferroni corrections for the number of cognitive test variables in (A) NC, (B) SCD, (C) aMCI, (D) d-AD groups. The solid lines show the best-fitting linear regression line and the 95% confidence intervals.

Figure 8

The scatter plots illustrate the significant associations between CAU.L nodal efficiency and AVLT-recognition scores controlled for age, gender and education after Bonferroni corrections for the number of cognitive test variables in (A) NC, (B) SCD, (C) aMCI, (D) d-AD groups. The solid lines show the best-fitting linear regression line and the 95% confidence intervals.