Enabling High Throughput Mechanical Analysis of the Human Brain
Abstract
Research ProblemThe Office of Naval Research s Naval Force Health Protection has an interest in developing Digital Twins and enabling Digital Engineering. Ideally, it would be possible to synthesize data from medical scans, the medical record, and sensors to create a subject-specific in silico model of every service person, analogous to efforts underway in physical Navy platforms. Then, it would be possible to conduct subject-specific simulation for applications such as tracking the onset of injury (e.g., TBI) or designing optimal protective gear (e.g., advanced helmet technology). However, at present, subject-specific simulations are prohibitively costly for real world implementation. Specifically, it is very labor intensive (e.g., one trained person working for months or years) to convert a subject scan (e.g., a brain MRI acquired in the clinical setting) to a finite element analysis simulation where a suiteof subject specific simulations can be performed. To date, the ONR Naval Force Health Protection Program and the PANTHER project have made significant strides in subject specific finite element analysis, as evidence by several recent publications. However, the methods used to create these simulations are very labor intensive, and require many manual and partially automated steps for success. Here, we propose an alternative synergistic paradigm where we prioritize throughput over fidelity.ObjectivesThe goal of this work isto develop methods for high throughput simulation of brain tissue. There are three planned phases:1. Phase 1 will be to create a fully automated high throughput simulation pipeline where input MRI scans will be converted directly to Finite Element Analysis (FEA) simulations without any human intervention.2. Phase 2 will be to incrementally increase the fidelity of the model while maintaining the automated high throughput paradigm.3. Phase 3 will be to converge on a pragmatic fidelity for high throughput analysis.Through these overarching objectives, we will be able to meaningfully step towards high throughput simulation of the brain.Technical ApproachWe have established a Python-based GitHub workflow that enables easy collaboration across universities and easy onboarding of new contributors to the project. This collaborative project is named #autotwin# which is shorthand for an #automated workflow from Image to Mesh for the Human Digital Twin. All parties have been working on the GitHub autotwin organization, which contains both public (published under open source licenses) and private (ongoing projects and notes) resources for achieving our goals. The team at BostonUniversity is responsible for the pixel repository which is focused on the image segmentation pipeline.Anticipated OutcomeThe anticipated outcome of this work is a pipeline to automatically analyze MRI scans with the finite element method. We anticipate that the pipeline will be generally applicable to other simulation types, in particular other simulationsrelevant to developing human DigitalTwins.Impact on DoD CapabilitiesIdeally, this work should transition directly to the Office of Naval Research s Naval Force Health Protection Digital Twin initiatives. Successful completion of this work will be a game changing building block in that direction.Approved for Public Release
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 15, 2023
- Source ID
- N000142312450
Entities
People
- Emma Lejeune
Organizations
- Boston University
- Office of Naval Research
- United States Navy