Deep learning of material transport in complex neurite networks

Angran Li,Amir Barati Farimani,Yongjie Jessica Zhang

doi:10.1038/s41598-021-90724-3

Angran Li, Amir Barati Farimani + Show 1 more

Open Access

https://doi.org/10.1038/s41598-021-90724-3

Copy DOI

Journal: Scientific Reports	Publication Date: May 28, 2021
Citations: 11	License type: open-access

Affiliation: Carnegie Mellon University

Abstract

Neurons exhibit complex geometry in their branched networks of neurites which is essential to the function of individual neuron but also brings challenges to transport a wide variety of essential materials throughout their neurite networks for their survival and function. While numerical methods like isogeometric analysis (IGA) have been used for modeling the material transport process via solving partial differential equations (PDEs), they require long computation time and huge computation resources to ensure accurate geometry representation and solution, thus limit their biomedical application. Here we present a graph neural network (GNN)-based deep learning model to learn the IGA-based material transport simulation and provide fast material concentration prediction within neurite networks of any topology. Given input boundary conditions and geometry configurations, the well-trained model can predict the dynamical concentration change during the transport process with an average error less than 10% and 120 sim 330 times faster compared to IGA simulations. The effectiveness of the proposed model is demonstrated within several complex neurite networks.

Highlights

Neurons exhibit complex geometry in their branched networks of neurites which is essential to the function of individual neuron and brings challenges to transport a wide variety of essential materials throughout their neurite networks for their survival and function
The aim of our graph neural network (GNN) model is to learn from the material transport simulation data and predict the transport process in any neurite network
We introduce a graph representation of the neurite network and build the GNN model based on the graph network (GN) framework[36]

Summary

Introduction

Neurons exhibit complex geometry in their branched networks of neurites which is essential to the function of individual neuron and brings challenges to transport a wide variety of essential materials throughout their neurite networks for their survival and function. Recent studies show that molecular motors play fundamental roles in intracellular transport to carry the material and move directionally along the cytoskeletal structure like microtubules or actin filaments[8,9,10] Motivated by these findings, different mathematical models based on partial differential equations (PDEs) have been proposed to help understand transport mechanisms and pathology of neuron diseases. Li et al developed an encoder-decoder based convolutional neural network (CNN) to directly predict concentration distribution of a reaction–diffusion system, bypassing the expensive FEM calculation process[29] While these works manage to learn the underlying physical models for prediction, they are limited to handle the problem in relatively simple geometry with Euclidean data (e.g. structured grid) available for training. GNNs were applied to predict the drag force associated with the laminar flow around airfoils[34] or the property of crystalline materials[35], understand the interaction behind physics scenes[36], solve PDEs by modeling spatial transformation functions[37], or learn particle-based simulation in complex physical s ystems[38]

Methods

Results

Conclusion