Small-worldness of functional networks in Alzheimer's disease

Abstract

To construct the functional networks of human brains by resting-state functional magnetic resonance imaging (fMRI) and examine whether or not the small-world property of functional brain networks changes in patients with Alzheimer disease (AD). A total of 33 AD patients and 20 healthy old volunteers were recruited. Their cognitive functions were evaluated by the mini-mental state examination (MMSE) and Mattis dementia rating scale (DRS). The resting-state BOLD-fMRI data were acquired and preprocessed. Then the correlation coefficient of every pair of 90 regions was calculated and correlation matrix Z (N×N, N = 90) generated by Fisher Z transformation. The small-world property of functional brain networks was tested for AD patients and normal elders according to the definition of small-worldness. The changes of relevant parameters in AD patients were examined by two sample t-tests. Behavioral results: the MMSE scores of AD group and normal controls (NOR) were 20.60 ± 2.30 and 28.20 ± 1.80 respectively. The DRS scores of AD and NOR groups were 96.00 ± 10.82 and 123.22 ± 13.74 respectively. The MMSE and DRS scores were statistically different between two groups. Calculation of small-world parameters: Within the range of 0.1 ≤ Sparsity ≤ 0.4, both groups satisfied the small-world property. However, the clustering coefficient Cp and the average shortest path Lp of AD group were significantly higher than those of NOR group at each threshold (P < 0.05). The functional brain networks in AD patients still have the property of small-world. But the levels of Cp and Lp are higher in AD patients than those in normal controls. It suggests that the capacity of information transmission in functional brain networks and the ability of information integration in different brain regions in AD patients are impaired. This finding is of great importance in elucidating the pathological mechanisms of AD from the viewpoint of networks.