Mapping Structural Connectome Patterns in Alzheimer's Disease

Tengfei Li, X. Wang, Martín G. Cole, Zhengwu Zhang, Hongtu ZhuAlzheimer 2025 · DOI

TL;DR

This study investigates how changes in brain connectivity, specifically fractional anisotropy (FA), mediate the impact of amyloid-beta (Aβ) buildup on cognitive decline in Alzheimer's Disease (AD). They found that Aβ's influence on cognitive impairment is significantly channeled through alterations in structural connectivity within key brain networks, particularly the default mode network (DMN) and frontoparietal network (FPN), even in the early stages of the disease.

What Problem Does This Paper Address?

Alzheimer's Disease (AD) is characterized by the accumulation of amyloid-beta (Aβ) plaques and neurofibrillary tangles, leading to progressive cognitive decline. While the link between Aβ and cognitive impairment is established, the precise mechanisms by which this occurs are still being unraveled. Specifically, how the physical structure of brain connections (structural connectivity) acts as an intermediary in this process remains an active area of research. Understanding these mediating pathways is crucial for developing targeted interventions that can disrupt the cascade from Aβ pathology to cognitive dysfunction. This paper aims to map these structural connectome patterns and quantify their role in mediating the relationship between Aβ and cognitive decline in AD.

Methods

The researchers employed diffusion tensor imaging (DTI) tractography to reconstruct white matter fiber bundles in the brain. They utilized the Tractoflow and TractDL pipelines to extract central fibers within identified bundles. For each fiber bundle, they computed the mean fractional anisotropy (FA), a measure reflecting the integrity and directionality of white matter. Linear mixed models were used to assess the associations between Aβ levels, FA, and cognitive scores (measured by the Mini Mental State Examination - MMSE), while controlling for various confounding factors like age, sex, and education. Furthermore, they conducted mediation analyses to quantify the extent to which FA changes explain the relationship between Aβ and MMSE scores. To explore early AD effects, they also examined these associations in cognitively normal (CN) subjects.

Datasets

The study utilized data from the Alzheimer's Disease Neuroimaging Initiative (ADNI2) dataset. This dataset includes a cohort of subjects with AD, mild cognitive impairment (MCI), and cognitively normal (CN) individuals, along with their corresponding MRI scans, Aβ levels (from cerebrospinal fluid - CSF), and cognitive assessments. The data was matched by visits to ensure consistency.

Key Insights and Ideas

  • Connectivity as a Crucial Mediator: The most significant insight is that Aβ's detrimental effect on cognition is not direct but is substantially mediated by changes in brain structural connectivity. This highlights the importance of the physical integrity of neural pathways in the AD cascade.
  • Early Susceptibility of DMN and FPN: The default mode network (DMN) and frontoparietal network (FPN) appear to be particularly vulnerable to Aβ's influence in the early stages of AD. This suggests that disruptions in these networks, which are critical for various cognitive functions, may be early biomarkers or drivers of cognitive decline.
  • Visual Network Involvement: The visual network also plays a role in mediating the effects of Aβ on cognition, indicating that even sensory processing pathways can be impacted by the disease pathology.

Key Results

  • The study identified 16 fiber bundles showing significant mediation effects between Aβ and MMSE scores.
  • These mediated pathways were primarily located within visual networks, the DMN, the FPN, and connections between the DMN and FPN, as well as DMN and sensorimotor networks.
  • Aβ was positively associated with FA in six DMN or FPN-related fibers during early AD phases (in cognitively normal subjects), suggesting that even before significant cognitive impairment, Aβ can impact the structural integrity of these networks.
  • The total indirect effect of Aβ on cognitive decline through FA changes was 0.080 (95% CI [0.025, 0.150]), accounting for approximately 21.6% of the total effect.

Limitations and Future Work

  • Cross-sectional Nature: While the study uses longitudinal data (multiple visits), the mediation analysis itself is presented in a way that suggests a cross-sectional interpretation of the relationships. Future work could explore the temporal dynamics of these mediation pathways more explicitly.
  • Specific Fiber Identification: While the study identifies specific fiber bundles, the precise functional implications of alterations in these exact bundles could be further elaborated.
  • Generalizability: The findings are based on the ADNI dataset. Future research could aim to replicate these findings in independent cohorts to enhance generalizability.
  • Therapeutic Implications: The identified mediating pathways offer potential targets for therapeutic interventions. Future work could explore how interventions aimed at preserving or restoring the integrity of these specific networks might mitigate cognitive decline in AD.

Summary updated

Abstract

Aβ's impact on cognitive decline was partly mediated (21.6%) by FA changes-primarily within visual, DMN, FPN, and DMN-FPN networks. DMN and FPN connectivity appears especially susceptible in early AD phases.
BibTeX
@inproceedings{li2025mapping,
  author = {Tengfei Li and X. Wang and Martín G. Cole and Zhengwu Zhang and Hongtu Zhu},
  title = {{Mapping Structural Connectome Patterns in Alzheimer's Disease}},
  booktitle = {Alzheimer},
  year = {2025},
  doi = {10.1002/alz70855\_104137},
}
Figure