Session: 06-01-02: Injury and Damage Biomechanics II - Organ and Tissue Injury Biomechanics 2

Paper Number: 145023

145023 - Data-Driven Depiction of Aging Related Physiological Volume Shrinkage in Brain White Matter: An Image Processing Based 3d Micromechanical Model 

Aging in human brain is characterized by diminishing cognitive ability which may be caused by either healthy or pathological aging mechanisms. Physiological aging leads to degenerative neurons and oligodendrocytes. These lead to manifestation of morphological changes (cerebral atrophy). Magnetic Resonance Elastography (MRE)/ Magnetic Resonance Imaging (MRI) scans have enabled non-interventional measurements of altering morphology (i.e., volume and micro-architecture changes) and characterize brain’s mechanical properties. These mechanical properties would encompass brain’s stifffness, softness and friction. MREs are very potent in detecting wave patterns that underscore atrophic changes in brain micro-architectures as a function of age (i.e., over an individual’s lifetime). Advanced imaging techniques like MRE/MRI are influenced by localized atrophic changes. This has motivated lot of research to accurately quantify volumetric degeneration (shrinkage of brain matter) and diminishing mechanical properties (shear moduli).

In this paper, image processing tools are deployed on processed MRE/MRI healthy aging patient scans to formulate a mathematical model to describe age dependent brain white matter (BWM) shrinkage. This research is a novel attempt to develop high-order aging brain computational models to describe age dependent softening in brain matter. The image processing code built in-house using NIBABEL libraries to read MRE/MRI scans and data science (regression), statistical analysis and optimization schema are then leveraged to explore various mathematical models to describe BWM volume shrinkage as a function of age. This volume shrinkage function will then be used in formulating finite element (FE) codes to define volume fractions (VF) of axons, myelin, and extra-cellular matrix (ECM) phases in a triphasic-modeled BWM representative volume element (RVE). The analysis and development of FEM codes is beyond the scope of the presented results. In this paper, the focus is to determine trends in volume shrinkage from the processed and analyzed MRE/MRI scans obtained for six patients in the age group of 27 to 65 years. Each MRE/MRI patient scan comprises of 60 layers/slices to make up the whole brain scan. Image processing techniques are then leveraged to carry out pixel based volumetric analysis on stacked image slices to arrive at an average volume shrinkage percentage b/w compared patient scans. Each of the six patient scans are compared against each other to obtain a 6 by 6 matrix volume shrinkage grid (upper triangular matrix) recording averaged volume shrinkage percentage entries.

Lastly, Grid search driven regularized curve fitting and central tendency measures (statistical analysis) functions are coded in python to derive best fit mathematical model of BWM volume as a function of age. volume fraction of axon and glia phases in brain white matter (BWM) representative volume element (RVE) 3D FEM setup, subjected to uniaxial tensile load. Results from mathematical modeling suggested that  n=3 (cubic) order model best described the volume shrinkage as a function of age (time). The results were compared to central tendency statistical analysis and both approaches conformed that a cubic (high-order) function could be apropos to describe non-linear changes in brain volume over age. The findings from this study would be critical ingredient in subsequent aging brain study. The derived functions will be used for defining VF parameters and micro-architectures for full-fledged Ogden hyper-elastic (HE) finite element models RVEs to depict BWM softening with age.

 

Presenting Author: Assimina Pelegri Rutgers, The State University on New Jersey

Presenting Author Biography: Dr. Assimina A. Pelegri is Professor and Chair of the Mechanical and Aerospace Engineering Department at Rutgers. Her research interests include composite materials modeling and characterization spanning from super-hard materials’ response on dynamic loads to modeling neuron degeneration for Alzheimer’s and traumatic brain injury using hybrid numerical/AI approaches. She is the recipient of Outstanding Faculty - Rutgers SOE, Rutgers Mentor of the Year, Outstanding Young Engineering Alumni - Georgia Institute of Technology, ASME Gold Medal- PTS, Amelia Earhart Award - Zonta International, and Fulbright Awards. She holds Ph.D. and MS degrees in aerospace engineering from Georgia Institute of Technology.

Authors:

Mohit Agarwal Rutgers, The State University of New Jersey
John G. Georgiadis Illinois Institute of Technology
Assimina Pelegri Rutgers, The State University on New Jersey