Using fixel‐based analysis to understand the role of crossing fibers in age‐related white matter decline

Aditi Sathe, Niranjana Shashikumar, Elizabeth E. Moore, Kimberly R. Pechman, Kurt G. Schilling, Bennett A. Landman, Katherine A. Gifford, Andrew R. Bender, Timothy J. Hohman, Angela L. Jefferson, Derek B. ArcherAlzheimer 2024 · DOI

TL;DR

This study introduces Fixel-Based Analysis (FBA) as a more precise method to study age-related white matter decline, moving beyond traditional metrics. FBA reveals that declining fiber density, particularly in areas with complex fiber crossings, is a key driver of age-related white matter changes, offering new insights into neurodegeneration.

What Problem Does This Paper Address?

Aging is associated with significant changes in white matter, the brain's communication network. Traditional methods like Diffusion Tensor Imaging (DTI) have been used to study these changes, but they often struggle to accurately characterize the complex microstructural alterations, especially in regions where multiple white matter tracts intersect (crossing fibers). These limitations mean we might be missing crucial details about how white matter degenerates with age, potentially hindering our understanding of age-related cognitive decline and neurodegenerative diseases. This paper aims to address this by employing Fixel-Based Analysis (FBA), a technique designed to provide higher resolution and specificity in these complex white matter regions.

Methods

The researchers utilized Fixel-Based Analysis (FBA), a novel approach that offers greater precision in quantifying white matter microstructure, particularly in areas with crossing fibers. This method was applied to diffusion MRI data preprocessed using the PreQual pipeline. Key FBA metrics calculated included:

  • Fiber Density (FD): A measure of the number of fibers within a voxel.
  • Fiber Bundle Cross-section (FC): A measure of the size of the fiber bundles.
  • Combined Effect (FDC): The combined impact of FD and FC.
  • Complexity (CX): A measure of the multi-fiber organization within a voxel.

To account for the varying complexity of fiber crossings, the study employed the TractSeg Bottleneck atlas. This atlas categorizes white matter regions into low, medium, and high bottleneck domains based on the number of converging white matter bundles. FBA metrics were then quantified within these specific regions. Linear regression models were used to assess the association between age and FBA metrics, controlling for several covariates including age, sex, education, race/ethnicity, APOE ε4 status, and cognitive status.

Datasets

The study used cross-sectional data from the Vanderbilt Memory & Aging Project cohort, comprising 325 participants. The average age at baseline was 72.9 ± 7.3 years, with 40% of participants diagnosed with mild cognitive impairment (MCI). No new datasets were introduced.

Key Insights and Ideas

  • Beyond Traditional Metrics: FBA offers a significant improvement over traditional DTI metrics by providing more granular insights into white matter microstructure, especially in complex fiber regions.
  • Fiber Density is Key: The decline in white matter integrity with age appears to be primarily driven by a reduction in fiber density (FD), rather than changes in the cross-sectional area of the fiber bundles (FC).
  • Bottlenecks Matter: Regions with a higher number of crossing fibers (high and medium bottleneck ROIs) show more pronounced age-related declines in white matter integrity, particularly in the primary fiber bundles. This suggests that these complex intersection points are more vulnerable to aging.
  • Cingulum Bundle Vulnerability: The cingulum bundle, a tract known to be affected by aging, also exhibits differential declines in FD and FC with age, further highlighting the utility of FBA in identifying specific tract vulnerabilities.

Key Results

  • Age-Related Decline: Fiber density (FD), fiber bundle cross-section (FC), and their combined effect (FDC) all showed significant declines with increasing age.
  • Primary Fiber Bundle Impact: The primary fiber bundle experienced the most substantial reductions in FDC with age (p<0.0001).
  • FD Drives Decline: This FDC reduction was predominantly due to changes in FD, while FC remained relatively stable.
  • Bottleneck Region Differences: Age-related alterations in the primary fiber bundle were most prominent in the high and medium bottleneck ROIs compared to low bottleneck ROIs (p<0.0001).
  • Cingulum Bundle Findings: FBA of the cingulum bundle revealed differential declines in FD and FC over time, with specific patterns observed across the low, medium, and high bottleneck domains.

Limitations and Future Work

  • Cross-Sectional Design: The study used cross-sectional data, which limits the ability to establish causality and track individual changes over time. Longitudinal studies are needed to confirm these findings.
  • Generalizability: While the cohort is from a specific research project, future work should aim to replicate these findings in larger, more diverse populations to ensure generalizability.
  • Tract-Specific Neurodegeneration: The authors suggest that this research lays the groundwork for future investigations into tract-specific patterns of neurodegeneration in aging. This could involve applying FBA to study specific neurological conditions and their impact on distinct white matter pathways.
  • Understanding Mechanisms: Further research is needed to elucidate the underlying biological mechanisms driving the observed declines in fiber density, particularly in complex fiber regions.

Summary updated

Abstract

Abstract Background Aging is linked to significant white matter abnormalities, which are often studied using traditional diffusion tensor imaging (DTI) metrics; however, these traditional metrics have limited sensitivity/specificity to neurobiological characteristics. Here, we use fixel‐based analysis (FBA) – an approach with more precision in areas of crossing fibers – to study age‐related white matter microstructural decline. Method This study uses cross‐sectional data from the Vanderbilt Memory &amp; Aging Project cohort [n=325, age at baseline: 72.9 ± 7.3 years, 40% mild cognitive impairment (MCI)]. Diffusion MRI was preprocessed using the PreQual pipeline and MRtrix was used to calculate fiber density (FD), fiber‐bundle cross‐section (FC), their combined effect (FDC), and complexity (CX) of the multi‐fiber organization within a voxel. To account for heterogeneity in the number of crossing fibers, the TractSeg Bottleneck atlas was split into distinct low/medium/high regions‐of‐interest (ROIs) based on the number of white matter bundles converging within a single voxel. Fixel‐based metrics were then quantified in each ROI. We conducted a linear regression to determine the association between age and all FBA metrics, covarying for age, sex, education, race/ethnicity, apolipoprotein‐e4 status, and cognitive status. Result FD, FC, and FDC all show significant decline with age, with the primary fiber bundle displaying the greatest reductions in FDC (p&lt;0.0001). This decrease is predominantly driven by changes in FD, while FC remains relatively consistent over time. Alterations to the primary fiber bundle are the most prominent in the high and medium bottleneck ROIs as opposed to low (p&lt;0.0001). FBA of the cingulum bundle (a tract vulnerable to abnormal aging) also showed differential decline in FD and FC over time. Figures 1‐3 provide illustrations of notable findings. The low/medium/high bottleneck ROIs correlate exceedingly well with global CX, supporting the use of the Bottleneck atlas to interpret structural white matter differences. Conclusion This study provides novel insights into the micro‐ and macro‐structural differences underlying age‐related white matter alterations. The findings underscore the value of FBA in characterizing age‐related white matter degeneration, particularly in regions susceptible to bottleneck effects. This research also lays the foundation for future investigations into tract‐specific patterns of neurodegeneration in aging.
BibTeX
@inproceedings{sathe2024using,
  author = {Aditi Sathe and Niranjana Shashikumar and Elizabeth E. Moore and Kimberly R. Pechman and Kurt G. Schilling and Bennett A. Landman and Katherine A. Gifford and Andrew R. Bender and Timothy J. Hohman and Angela L. Jefferson and Derek B. Archer},
  title = {{Using fixel‐based analysis to understand the role of crossing fibers in age‐related white matter decline}},
  booktitle = {Alzheimer},
  year = {2024},
  doi = {10.1002/alz.090876},
}
Figure