White matter microstructural aging depends on developmental order, fibre calibre and vascularization
Poster No:
2168
Submission Type:
Abstract Submission
Authors:
Tyler Robinson1, Yutong Sun1, Jordan Chad1, Paul Chang2, J. Jean Chen3
Institutions:
1Rotman Research Institute, Baycrest, Toronto, Ontario, 2Rotman Reserach Institute, Baycrest, Toronto, Ontario, 3Baycrest Health Sciences, Toronto, Ontario
First Author:
Co-Author(s):
Introduction:
To understand the consistently observed spatial distribution of white-matter (WM) aging, developmentally driven theories of retrogenesis have gained traction, positing that the order WM development predicts declines(5,8,9). The validity of such theories remains uncertain, in part due to lack of clarity on the definition of developmental order(6,3,8). An alternative theory is termed "gain-predicts-loss"(8), which posits that the rate of development predicts rate of declines, likewise for which several definitions may apply(7,9). Our recent findings also suggest that WM degeneration may vary by physiological parameters such as perfusion(7). Here, using the HCP-A data, we address the question of whether WM degeneration is determined by development trajectory or physiological state.
Methods:
535 data sets from healthy subjects (300 female, aged 36-100) were drawn from the HCP-A study (OMB Control# 0925-0667)(2), including diffusion-weighted MRI (93 directions at b=1500s/mm2) at 1.5mm isotropic resolution. We corrected for eddy-current and susceptibility-related distortions via EDDY, and kurtosis-corrected fractional anisotropy (FA) and mean diffusivity (MD) maps were derived using Dipy's DKI tool. FreeSurfer TRACULA (v 7.2.0)(10) was used to provide 10 bilateral tract segmentations in subject space. All DTI values were normalized to the youngest 50 subjects to produce FAperc and MDperc values. We defined 6 models to help explain DTI age effects:
Last-in-first-out models;
(1) "Last-appearing-first-out" (LAFO): ordered by time of prenatal emergence(6)
(2) "Last-myelinated-first-out" (LMFO): ordered by time of peak R1 ("myelination") at birth(3)
Gain-predicts-loss models;
(3) "Slowest-peak-first-out" (SPFO-MD): ordered by rate of MD change from adolescence to adult peak(8)
(4) "Slowest-myelinated-first-out (SMFO): ordered by rate of R1 change from adolescence to adult peak(8)
Physiological models:
(5) "Axon-diameter-determined" (ADD): ordered by mean axon diameter at birth(4)
(6) "Vascular-diameter-determined" (VDD): ordered by mean macrovascular density(1)
The predictiveness of these models was assessed using multivariate regression in R (version 4.1.1), with sex as a covariate of no interest.
Last-in-first-out models;
(1) "Last-appearing-first-out" (LAFO): ordered by time of prenatal emergence(6)
(2) "Last-myelinated-first-out" (LMFO): ordered by time of peak R1 ("myelination") at birth(3)
Gain-predicts-loss models;
(3) "Slowest-peak-first-out" (SPFO-MD): ordered by rate of MD change from adolescence to adult peak(8)
(4) "Slowest-myelinated-first-out (SMFO): ordered by rate of R1 change from adolescence to adult peak(8)
Physiological models:
(5) "Axon-diameter-determined" (ADD): ordered by mean axon diameter at birth(4)
(6) "Vascular-diameter-determined" (VDD): ordered by mean macrovascular density(1)
The predictiveness of these models was assessed using multivariate regression in R (version 4.1.1), with sex as a covariate of no interest.
Results:
Age-related declines were observed in MDperc for all ten tracts, and in FAperc for seven of ten tracts (Fmajor, Fminor, ATR, CAB, CCG, ILF, and SLFP).
Both last-in-first-out models demonstrated significant associations with at least one measure of microstructural integrity (Fig. 1). In the LAFO model, tract development order was positively associated with mean FAperc but not MDperc. In LMFO, R1 order was inversely associated with FAperc, while MDperc was positively associated. In rate-of-decline comparisons, significant interactions by tract order on association between FAperc and age were observed in the LMFO ordering significantly associated with FAperc rate of decline (Fig. 2). No significant contributions were identified by the gain-predicts-loss models (SPFO-MD and SMFO).
Both physiological determinant models demonstrated significant associations with both FAperc and MDperc (Fig. 1). FAperc was positively associated with axon diameter and inversely related to macro-vascular density order. Inversely, MDperc was positively associated with vascular density and negatively associated with axonal diameter. In rate-of-decline comparisons, the rate of FAperc decline correlated positively with axonal diameter and negatively with vascular density (Fig. 2). The reverse was true for MDperc.
Both last-in-first-out models demonstrated significant associations with at least one measure of microstructural integrity (Fig. 1). In the LAFO model, tract development order was positively associated with mean FAperc but not MDperc. In LMFO, R1 order was inversely associated with FAperc, while MDperc was positively associated. In rate-of-decline comparisons, significant interactions by tract order on association between FAperc and age were observed in the LMFO ordering significantly associated with FAperc rate of decline (Fig. 2). No significant contributions were identified by the gain-predicts-loss models (SPFO-MD and SMFO).
Both physiological determinant models demonstrated significant associations with both FAperc and MDperc (Fig. 1). FAperc was positively associated with axon diameter and inversely related to macro-vascular density order. Inversely, MDperc was positively associated with vascular density and negatively associated with axonal diameter. In rate-of-decline comparisons, the rate of FAperc decline correlated positively with axonal diameter and negatively with vascular density (Fig. 2). The reverse was true for MDperc.
Conclusions:
Of the many alternative definitions of "last-in-first-out", order of myelination seems most promising, i.e. fastest loss in tracts last to reach peak myelination. Importantly, our results also suggest that tracts with the largest fibre diameters and lowest vascular densities to be more preserved in aging. This also implies that larger fibre diameters are associated with lower vascular density, a finding to be further investigated.
Lifespan Development:
Aging 2
Modeling and Analysis Methods:
Diffusion MRI Modeling and Analysis
Neuroanatomy, Physiology, Metabolism and Neurotransmission:
White Matter Anatomy, Fiber Pathways and Connectivity 1
Keywords:
Aging
Development
White Matter
Other - Microstructure; Retrogenesis; DTI
1|2Indicates the priority used for review
Provide references using author date format
2. Bookheimer, S. Y. (2019). The Lifespan Human Connectome Project in Aging: An overview. NeuroImage, 185, 335–348.
3. Grotheer, M. (2022). White matter myelination during early infancy is linked to spatial gradients and myelin content at birth. Nature Communications, 13(1), 997.
4. Huang, S. Y. (2020). High-gradient diffusion MRI reveals distinct estimates of axon diameter index within different white matter tracts in the in vivo human brain. Brain Structure & Function, 225(4), 1277–1291.
5. Kiely, M. (2022). Insights into human cerebral white matter maturation and degeneration across the adult lifespan. NeuroImage, 247, 118727.
6. Ouyang, Y. (2021). Analysis of Age-Related White Matter Microstructures Based on Diffusion Tensor Imaging. Frontiers in Aging Neuroscience, 13, 664911.
7. Robinson, T. D. (2023). In search of a unifying theory of white matter aging: associations with age in perfusion detectable in advance of microstructure. In bioRxiv (p. 2023.06.30.547294). https://doi.org/10.1101/2023.06.30.547294
8. Slater, D. A. (2019). Evolution of white matter tract microstructure across the life span. Human Brain Mapping, 40(7), 2252–2268.
9. Yeatman, J. D.(2014). Lifespan maturation and degeneration of human brain white matter. Nature Communications, 5, 4932.
10. Yendiki, A. (2011). Automated probabilistic reconstruction of white-matter pathways in health and disease using an atlas of the underlying anatomy. Frontiers in Neuroinformatics, 5, 23.