Biomimetic IGA neuron growth modeling with neurite morphometric features and CNN-based prediction

Abstract

Neuron growth is a complex, multi-stage process that neurons undergo to develop sophisticated morphologies and interwoven neurite networks. Recent experimental research advances have enabled us to examine the effects of various neuron growth factors and seek potential causes for neurodegenerative diseases, such as Alzheimer’s disease, Parkinson’s disease, and amyotrophic lateral sclerosis. A computational tool that studies the neuron growth process could shed crucial insights on the effects of various factors and potentially help find a cure for neurodegeneration. However, there is a lack of computational tools to accurately and realistically simulate the neuron growth process within reasonable time frames. Bio-phenomenon-based models ignore potential neuron growth factors and cannot generate realistic results, and bio-physics-based models require extensive, high-order governing equations that are computationally expensive. In this paper, we incorporate experimental neurite features into a phase field method-based neuron growth model using an isogeometric analysis collocation (IGA-C) approach. Based on a semi-automated quantitative analysis of neurite morphology, we obtain relative turning angle, average tortuosity, neurite endpoints, average segment length, and the total length of neurites. We use the total neurite length to determine the evolving days in vitro (DIV) and select corresponding neurite features to drive and constrain the neuron growth. This approach archives biomimetic neuron growth patterns with automatic growth stage transitions by incorporating corresponding DIV neurite morphometric data based on the total neurite length of the evolving neurite morphology. Furthermore, we built a convolutional neural network (CNN) to significantly reduce associated computational costs for predicting complex neurite growth patterns. Our CNN model adopts a customized convolutional autoencoder as the backbone that takes neuron growth simulation initializations and target iteration as the input and predicts the corresponding neurite patterns. This approach achieves high prediction accuracy (97.77%) while taking 7 orders of magnitude less computational times when compared with our IGA-C neuron growth solver.