Structural Network Topology Reveals Higher Brain Resilience in Individuals with Preclinical Alzheimer's Disease

Abstract

Introduction:

The diagnosis of Alzheimer's disease (AD) requires the presence of amyloid and tau pathology, but it remains unclear how they affect the structural network in the pre-clinical stage. We aimed to assess differences in topological properties in cognitively normal (CN) individuals with varying levels of amyloid and tau pathology, as well as their association with AD pathology burden.

Methods:

A total of 68 CN individuals were included and stratified by normal/abnormal (−/+) amyloid (A) and tau (T) status based on positron emission tomography results, yielding three groups: A−T− (n = 19), A+T− (n = 28), and A+T+ (n = 21). Topological properties were measured from structural connectivity. Group differences and correlations with A and T were evaluated.

Results:

Compared with the A−T− group, the A+T+ group exhibited changes in the structural network topology. At the global level, higher assortativity was shown in the A+T+ group and was correlated with greater tau burden (r = 0.29, p = 0.02), while no difference in global efficiency was found across the three groups. At the local level, the A+T+ group showed disrupted topological properties in the left hippocampus compared with the A−T− group, characterized by lower local efficiency (p < 0.01) and a lower clustering coefficient (p = 0.014).

Conclusions:

The increased linkage in the higher level architecture of the white matter network reflected by assortativity may indicate increased brain resilience in the early pathological state. Our results encourage further investigation of the topological properties of the structural network in pre-clinical AD.

Impact statement

The present work explored the topological patterns of the brain structural network in subjects with pre-clinical Alzheimer's disease (AD) and their correlation with the two AD hallmarks, amyloid and tau. Results showed an increased linkage in the higher level architecture of the white matter network measured by assortativity, while global efficiency remained unchanged. The study provides a potential graph-based biomarker of brain connectivity for early identification of AD, and the results encourage further investigation of structural network properties.