Links between gray-matter stiffness, age and memory using Magnetic Resonance Elastography
Poster No:
1169
Submission Type:
Abstract Submission
Authors:
Grégoria Kalpouzos1, Zuzana Istvanfyova1,2, Göran Hagman1,2, Rebecca Ericsson1, Ronja Löfström1, Farshad Falahati1, Christoffer Olsson3, Rodrigo Moreno3, Jonas Persson1,4
Institutions:
1Karolinska Institutet, Stockholm, Sweden, 2Karolinska University Hospital, Stockholm, Sweden, 3KTH Royal Institute of Technology, Stockholm, Sweden, 4Örebro University, Örebro, Sweden
First Author:
Co-Author(s):
Zuzana Istvanfyova
Karolinska Institutet|Karolinska University Hospital
Stockholm, Sweden|Stockholm, Sweden
Karolinska Institutet|Karolinska University Hospital
Stockholm, Sweden|Stockholm, Sweden
Göran Hagman
Karolinska Institutet|Karolinska University Hospital
Stockholm, Sweden|Stockholm, Sweden
Karolinska Institutet|Karolinska University Hospital
Stockholm, Sweden|Stockholm, Sweden
Introduction:
Magnetic Resonance Elastography (MRE) applied to the brain has been recently developed to assess the biomechanical properties of the neural tissue such as stiffness (Yin et al 2018). Little is known about age effects on cerebral stiffness, and whether it can explain interindividual variability in cognitive performance. The few MRE studies in normal aging suggested a local reduction of cerebral stiffness and an association between hippocampal stiffness and memory (Coelho & Sousa 2022; Delgorio et al 2022). Our aims were to (1) characterize gray-matter stiffness as a function of age voxel-by-voxel, and (2) investigate age-independent associations between stiffness and visuospatial learning. Besides, we tested whether adding gray-matter volume or density (GMv, GMd) contributed to the associations between stiffness, age and cognition.
Methods:
Twenty-six healthy volunteers (age range: 26-79 years old, 12 women) underwent magnetic resonance imaging and a battery of cognitive tests. MRI was performed on a 3.0T Philips Ingenia scanner, equipped for MRE with a pump connected to the scanner that sends waves through vibrations to a pillow on which the head lies, and a validated pulse sequence together with a direct inversion algorithm (Mayo Clinic, USA) that estimates the shear stiffness (i.e., resistance of a material to a shear deformation) and damping ratio (i.e., resistance of a material to oscillations). We here used the reconstructed stiffness maps whose values were expressed in pascals, and the T1-weighted images. Visuospatial learning was assessed using the Hagman test (Holleman et al 2022). The subjects were presented for 45 seconds with a 3x3 cells matrix, in which different symbols in a particular rotation and color were displayed. The instructions were to memorize the content and draw it in an empty matrix immediately after. The same test was repeated 30 minutes later to test learning. The T1-weighted images were segmented in Statistical Parametric Mapping (SPM12), and further processed using DARTEL (Ashburner 2007) to obtain spatially normalized maps of GMv (modulated images) and GMd (unmodulated images) into MNI space. The stiffness maps, coregistered to the T1s, were normalized using the flow fields from DARTEL. Images were smoothed with an 8-mm kernel. A GM mask was applied to all voxel-based analyses. To address aim 1, voxel-based regressions were performed between the stiffness maps and age at p FWE-corrected < .05. To address aim 2, voxel-based correlations were performed between the stiffness maps and performance at Hagman trial-2 at p < .001 (uncorrected), controlling for age and Hagman trial-1 score. Using MarsBaR toolbox, mean stiffness was extracted from the significant MRE-clusters and applied to the GMv and GMd maps to further extract volume and density for follow-up stepwise regression analyses.
Results:
With older age, lower stiffness was found in inferior frontal regions (BA 44/45, 47), premotor and somatosensory regions (BA 6, 3), lateral temporal areas (BA 21/22, 42) and insula (Fig 1A). These associations remained largely unaltered when adding GMv or GMd to the models (e.g., left frontal cluster: R2 model 1=0.792, R2 model 2 (adding GMd)=0.797, p for R2 change=.44). Controlling for age and Hagman test trial-1, better visuospatial learning was related to higher stiffness in posterior cingulate/retrosplenial area, frontal cortex, insula, striatum, and posterior hippocampus (Fig 1B). These associations remained largely significant when GMd or GMv were added to the models (e.g., for retrosplenial cluster: R2 model 1=0.560, R2 model 2=0.560, p for R2 change = .95).
·Figure 1. Associations between (A) age and cerebral stiffness, and between (B) stiffness and visuospatial learning performance
Conclusions:
Reduced GM stiffness may contribute to brain aging independently of atrophy. GM stiffness may also contribute to interindividual differences in cognition, as higher stiffness in key-regions for visuospatial processing and memory was associated with better visuospatial learning independently of age.
Learning and Memory:
Learning and Memory Other
Lifespan Development:
Aging 1
Novel Imaging Acquisition Methods:
Anatomical MRI
Multi-Modal Imaging 2
Imaging Methods Other
Keywords:
Aging
Cognition
Learning
Memory
Morphometrics
MRI
STRUCTURAL MRI
Other - MR Elastography
1|2Indicates the priority used for review
Provide references using author date format
Coelho, A. (2022), "Magnetic resonance elastography of the ageing brain in normal and demented populations: A systematic review", Human Brain Mapping, vol. 43, no. 13, pp. 4207-4218
Delgorio, P.L. (2022), "Structure-Function Dissociations of Human Hippocampal Subfield Stiffness and Memory Performance", Journal of Neuroscience, vol. 42, no. 42, pp. 7957-7968
Holleman, J. (2022), "Cortisol, cognition and Alzheimer's disease biomarkers among memory clinic patients", BMJ neurololy open, vol. 4, no. 2, pp. e00034
Yin, Z. (2018), "Stiffness and Beyond: What MR Elastography Can Tell Us About Brain Structure and Function Under Physiologic and Pathologic Conditions", Topics in Magnetic Resonance Imaging, vol. 27, no. 5, pp. 305-318.