Te-Wei Kao1, Yung-Chin Hsu2, and Wen-Yih Isaac Tseng1,3
1Institute of Medical Device and Imaging, National Taiwan University College of Medicine, Taipei, Taiwan, 2AcroViz Technology, Inc., Taipei, Taiwan, 3Molecular Imaging Center, National Taiwan University, Taipei, Taiwan
Synopsis
Previous
studies did not clearly characterize normal ageing patterns of white matter tracts
across lifespan. Here, we performed tract-specific automatic analysis over the
whole brain to measure the diffusion indices on 610 healthy participants recruited
from Cam-CAN. We used quadratic or linear models to plot the curves of diffusion
indices against age, and calculated the average values and slopes of the curves
in 10 subsystems classified from 76 major tracts. Our study characterized
different ageing patterns corresponding to different subsystems.
Introduction
Ageing
is a major risk factor for common neurodegenerative disease, and so
comprehensive assessment of the human brain structures across lifespan is important
for understanding normal aging1. Several studies on brain aging
suggest that diffusion tensor imaging (DTI) can help us advance our knowledge
of the brain2. However, previous papers only studied a few major
white matter tracts, and the problem of uneven age distribution of the
population often occurred. For these reasons, the past findings have been
inconsistent in characterizing the trajectory of age effects3. Here,
we used whole-brain tract-specific automatic analysis of DTI to investigate the
microstructural properties of 10 subsystems classified from 76 white matter
tracts. We performed analysis on participants recruited from the Cambridge
Centre for Ageing and Neuroscience (Cam-CAN) databank whose age distributed uniformly
across lifespan.Methods
Participants:
649
healthy participants were recruited from Cam-CAN. After screening, usable data were obtained from 610 people: age
range=18-88 years; mean=53.99 years; SD=18.34 years; female=309 and
male=301.
MRI
data acquisition:
The
MRI data were acquired on a 3T MRI system (TIM Trio, Siemens, Erlangen) with a
32-channel phased array coil. For Cam-CAN protocol3, T1-weighted
imaging utilized a 3D magnetization-prepared rapid gradient echo (MPRAGE) pulse
sequence. The imaging parameters of MPRAGE were TR/TE/TI= 2250/2.99/900ms,
flip angle=9°, FOV=256×240×192mm3, and spatial resolution=1×1×1mm3. DTI utilized a pulsed gradient twice-refocused spin-echo
diffusion echo planar imaging sequence. Two-shell
DTI acquisition scheme used 30 diffusion-encoding gradients directed to each
shell; the b-values of the two shells were 1000 and 2000 s/mm2. Three
images acquired at b-value=0 served as the baseline images. The imaging
parameters of DTI were TR/TE=9100/104ms, FOV=192×192mm2,
matrix size=96×96, and slice thickness=2mm.
Data
analysis:
From
DTI data, we reconstructed five diffusion indices, i.e. generalized fractional anisotropy
(GFA), fractional anisotropy (FA), mean diffusivity (MD), axial diffusivity (AD), and radial diffusivity (RD). We used tract-based automatic analysis (TBAA) to
obtain a 2D connectogram for each DTI dataset, which provided diffusion index
profiles of 76 white matter tract bundles4. We also used TBAA to
sample the tissue probability maps, and corrected the probability of partial volume
effects of the cerebrospinal fluid (PVECSF). The quadratic or linear
model was built to plot the curve patterns and calculate the average values
across lifespan (AV) and slopes of the five diffusion indices and the probability
of PVECSF. The curve patterns and resulting metrics were compared in
10 subsystems classified from the 76 white matter tracts (Table 1)(Figure
1).
Results
The
limbic system (AFLimbic) and cortical-cortical (AFCortical)
association fiber system had similar curve patterns in GFA and FA, but the
former had higher AV and steeper slopes in MD, AD, RD, and the probability of
PVECSF. For the projection fiber system, the corticospinal tract (PFCST)
had the highest AV and flat slopes in GFA and FA accompanied with the lowest AV
and flat slope in RD. There were slight differences in AV and slopes between the
frontal-striatum (PFFS) and thalamic radiation (PFTR)
projection fibers. Notably, a trend of the slopes from steep to flat was found
in frontal (CFFrontal), parietal (CFParietal), occipital
(CFOccipital), and temporal (CFTemporal) callosal fiber
system in GFA and FA. As to the anterior and posterior commissures (CFCom),
the unique positive slopes in GFA and FA were noted (Table 2)(Figure 2).Discussion
The
decrease of GFA and FA, and increase of MD, AD, and RD are a reflection of the
microstructural integrity changes in loss of myelin, axonal fibers, and
increase in extracellular space2. Most of white matter tract bundles
had the similar curve patterns that the integrity declined across lifespan. In
the PFCST, however, GFA and FA had relatively higher values and flatter
slopes, indicating slower degradation. In the AFLimbic, MD, AD, RD
and PVECSF increased in tandem. The finding implies strong effects
of PVECSF on AFLimbic, consistent with previous studies5.
Moreover, our results were the first to uncover the process of sequential
decline in CFFrontal, CFParietal, CFOccipital,
and CFTemporal, which was opposite to the posterior-to-anterior shift in aging
hypothesis in functional neuroimaging studies6. In addition, we found the
unique positive slopes of GFA and FA in CFCom, suggesting maintained
integrity of the anterior and posterior commissures. In summary, GFA and FA were
suitable to represent different types of curve patterns. A limitation of the
study was the imperfect correction of PVECSF, but the GFA
and FA were almost unaffected and stable, implying that our observations were
still valid.Conclusion
Our
study revealed the characteristic normal ageing patterns of microstructural
properties in 10 subsystems. The information may be helpful in identifying
accelerated ageing in neurodegenerative disease. Acknowledgements
All
data were provided by Cam-CAN, subject to a data transfer agreement. We thank
the Cam-CAN team (http://www.cam-can.org/), which was crucial in recruiting
participants, developing the protocol, and overseeing data management. References
- Cox SR, et al. Ageing
and brain white matter structure in 3,513 UK Biobank participants. Nat Commun.
2016;7:13629.
- Moseley M. Diffusion
tensor imaging and aging - a review. NMR Biomed. 2002;15(7-8): 553-560.
- Shafto MA, et al. The
Cambridge Centre for Ageing and Neuroscience (Cam-CAN) study protocol: a
cross-sectional, lifespan, multidisciplinary examination of healthy cognitive
ageing. BMC Neurol. 2014;14:204.
- Chen, Yu-Jen, et al.
Automatic whole brain tract-based analysis using predefined tracts in a
diffusion spectrum imaging template and an accurate registration strategy. Human
brain mapp. 2015;36(9):3441-3458.
- Luis Concha L, et al.
Diffusion tensor tractography of the Limbic System. AJNR Am J Neuroradiol.
2005;26(9):2267-2274.
- Davis SW, et al. Que
PASA? The posterior-anterior shift in aging. Cereb Cortex. 2008;18(5):1201-1209.