G-Ratio and Conduction Velocity are Associated with Aging but Differentially with Cognitive Decline
Poster No:
1211
Submission Type:
Abstract Submission
Authors:
Benjamin Newman1, Jason Druzgal1, John Van Horn1
Institutions:
1University of Virginia, Charlottesville, VA
First Author:
Co-Author(s):
Introduction:
Recent efforts to quantify brain cellular microstructure have developed metrics of brain structure with implications for neuronal function. Termed aggregate g-ratio and aggregate conduction velocity, these metrics incorporate measures of myelin and intra-axonal volume. However it is not currently known how these metrics compare to age in an aging cohort nor how they predict cognition within this group. White matter hyperintensities (WMH) are a marked appearance of the aging brain, and should provide an ideal validation for the loss of myelin and subsequent functional disruption, estimated by these metrics.
Methods:
Participants: 86 subjects were recruited for baseline neuroimaging with an age range between 58-81 (mean = 68.2 ± 5.84 S.D.) years old. There were 27 male and 58 female participants with average ages of 66.70 ± 4.97 S.D. and 68.87 ± 6.04 S.D., respectively.
Image Acquisition: T1-weighted images were acquired using the ADNI3 designed MP-RAGE sequence with an isotropic voxel size 1.0×1.0×1.0mm3, TE=2980ms and TR=2300ms. Diffusion-weighted images were acquired with an isotropic voxel size of 1.7×1.7×1.7mm3, TE=70ms and TR=2900ms; 10 b=0 images and 64 gradient directions were collected at both b=1500s/mm2 and b=3000s/mm2.
Behavioral and Cognitive Metrics: All subjects participated in the Trail Marking Task A & B, a measure of executive function where subjects are timed to completion of the task. B is considered more cognitively demanding as subjects must switch between letters and numbers in sequence.
Image Processing: As described in more detail in previous work3 images were preprocessed with MRtrix34, FSL5, and Freesurfer6 to calculate voxel-wise aggregate g-ratio and aggregate conduction velocity with the mean b=0 images being used in lieu of T2-weighted images. WMH were automatically segmented out from the T1w images using Freesurfer. The mean value of these metrics was measured within each of the 48 regions of the JHU WM Atlas as a singular whole white matter ROI. Within subjects T-tests were used to evaluate the differences between differences within WMH and WM. Linear models tested the association between the trails tasks and cellular microstructure with controls for subject age and sex.
Image Acquisition: T1-weighted images were acquired using the ADNI3 designed MP-RAGE sequence with an isotropic voxel size 1.0×1.0×1.0mm3, TE=2980ms and TR=2300ms. Diffusion-weighted images were acquired with an isotropic voxel size of 1.7×1.7×1.7mm3, TE=70ms and TR=2900ms; 10 b=0 images and 64 gradient directions were collected at both b=1500s/mm2 and b=3000s/mm2.
Behavioral and Cognitive Metrics: All subjects participated in the Trail Marking Task A & B, a measure of executive function where subjects are timed to completion of the task. B is considered more cognitively demanding as subjects must switch between letters and numbers in sequence.
Image Processing: As described in more detail in previous work3 images were preprocessed with MRtrix34, FSL5, and Freesurfer6 to calculate voxel-wise aggregate g-ratio and aggregate conduction velocity with the mean b=0 images being used in lieu of T2-weighted images. WMH were automatically segmented out from the T1w images using Freesurfer. The mean value of these metrics was measured within each of the 48 regions of the JHU WM Atlas as a singular whole white matter ROI. Within subjects T-tests were used to evaluate the differences between differences within WMH and WM. Linear models tested the association between the trails tasks and cellular microstructure with controls for subject age and sex.
Results:
Aggregate g-ratio was significantly reduced in WMH compared to the whole WM skeleton (T84=-13.898, p<0.001; Fig. 1). Aggregate conduction velocity was also significantly reduced in WMH compared to the whole WM skeleton (T84=-26.633, p<0.001; Fig. 1). However, the microstructure of the WMH was not significantly associated with age (g-ratio F2,84=0.00086, p=0.59 n.s.; conduction velocity F2,84=-0.0023, p=0.277 n.s.) nor performance on either Trail Marking Tasks (Trail A, g-ratio: F2,84=0.304, p=0.983 n.s.; Trail A, conduction velocity: F2,84=-6.439, p=0.54 n.s.; Trail B, g-ratio: F2,84=26.35, p=0.551 n.s.; Trail B, conduction velocity: F2,84=-2.513, p=0.938 n.s. Fig. 2). In the whole WM skeleton though Trail A & B was significantly associated with aggregate conduction velocity (Trail A: F2,84=-41.03, p<0.05; Trail B: F2,84=-176.080, p<0.01) but not aggregate g-ratio (Trail A: F2,84=-34.45, p=0.379; Trail B: F2,84=-30.66, p=0.797 n.s.). Both aggregate conduction velocity and aggregate g-ratio were observed to have a negative relationship with age in the WM skeleton (F2,84=-0.0023, p<0.05; F2,84=-0.0015, p<0.05, respectively).
·Figure 2: Chart showing variation between microstructural metrics and Trails performance in WMH and whole WM skeleton.
·Figure 1: Visual comparison of a single subject's aggregate g-ratio and aggregate conduction velocity maps, MWH are far more visible in posterior regions as measured by aggregate conduction velocity.
Conclusions:
These results demonstrate that reductions in aggregate conduction velocity within the normal appearing WM skeleton, are associated with poorer cognitive performance. Aggregate g-ratio showed similar patterns of WMH damage and age-related decline, but did not have a similar relationship with cognitive performance. These results suggest that not only are aggregate conduction velocity and g-ratio measuring different microstructural features, but that the loss of larger diameter axons may drive cognitive decline in aging.
Higher Cognitive Functions:
Executive Function, Cognitive Control and Decision Making 2
Lifespan Development:
Aging 1
Keywords:
Aging
Cognition
Demyelinating
Myelin
White Matter
WHITE MATTER IMAGING - DTI, HARDI, DSI, ETC
1|2Indicates the priority used for review
Provide references using author date format
2. Tournier JD, Smith R, Raffelt D, et al. MRtrix3: A fast, flexible and open software framework for medical image processing and visualisation. NeuroImage. 2019;202:116137.
3. Jenkinson M, Beckmann CF, Behrens TE, Woolrich MW, Smith SM. Fsl. Neuroimage. 2012;62(2):782-790.
4. Fischl B. FreeSurfer. Neuroimage. 2012;62(2):774-781.