Rapid Estimation of Entire Brain Strain Using Deep Learning Models.

Area of Science:

• Biomechanics
• Computational Neuroscience
• Machine Learning

Background:

• Head impacts can cause brain deformation linked to mild traumatic brain injury (mTBI).
• Current finite element (FE) models accurately simulate brain deformation but are computationally intensive, limiting clinical application.
• Accelerating brain deformation calculation is crucial for real-time monitoring and large-scale analysis.

Purpose of the Study:

• To develop and validate a deep learning model for rapid and accurate calculation of brain deformation from head impacts.
• To overcome the computational limitations of traditional FE models for clinical and research applications.

Main Methods:

• A five-layer deep neural network with feature engineering was developed.
• The model was trained and tested on a dataset of 2511 head impacts, combining FE simulations and real-world data from sports.
• The model calculates maximum principal strain (Green Lagrange) for all brain elements.

Main Results:

• The deep learning model computes brain strain in under 0.001 seconds.
• Achieved an average root mean squared error of 0.022 with a standard deviation of 0.001 over twenty model repeats.
• Demonstrated high accuracy in calculating brain strain across various head impacts.

Conclusions:

• The proposed deep learning model offers a fast and accurate method for estimating brain strain from head impacts.
• This approach has potential for real-time clinical applications and efficient analysis of large head impact datasets.
• Further improvements are possible by incorporating diverse head impact data for broader applicability.