Neuron type classification in rat brain based on integrative convolutional and tree-based recurrent neural networks

Zhang, Tielin; Zeng, Yi; Zhang, Yue; Zhang, Xinhe; Shi, Mengting; Tang, Likai; Zhang, Duzhen

doi:10.1038/s41598-021-86780-4

Cited by 27 publications

(12 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Far fewer morphological classification studies have also included glia, and they typically did not focus on directly comparing neurons to glia. For example, Leyh et al (2021) classified different types of microglia in healthy and diseased mouse model, while Zhang et al (2021) added glia as a separate phenotype in a multiclass neuron type categorization task using convolutional neural networks. Recognizing the morphological signatures that distinguish glia from neurons is an important yet unfulfilled step.…”

Section: Discussionmentioning

confidence: 99%

Machine Learning Classification Reveals Robust Morphometric Biomarker of Glial and Neuronal Arbors

Akram

Wei

Ascoli

2022

Preprint

View full text Add to dashboard Cite

Neurons and glia are the two main cell classes in the nervous systems of most animals. Although functionally distinct, neurons and glia are both characterized by multiple branching arbors stemming from the cell bodies. Glial processes are generally known to form smaller trees than neuronal dendrites. However, the full extent of morphological differences between neurons and glia in multiple species and brain regions has not yet been characterized, nor is it known whether these cells can be reliably distinguished based on geometric features alone. Here, we show that multiple supervised learning algorithms (K-nearest neighbor, random forest, and support vector machine) deployed on a large database of morphological reconstructions can systematically classify neuronal and glial arbors with nearly perfect accuracy and precision. Moreover, we report multiple morphometric properties, both size-related and size-independent, that differ substantially between these cell types. In particular, we newly identify an individual morphometric measurement, Average Branch Euclidean Length (ABEL) that can robustly separate neurons from glia across multiple animal models, a broad diversity of experimental conditions, and anatomical areas, with the notable exception of the cerebellum. We discuss the practical utility and physiological interpretation of this discovery.

show abstract

Section: Discussionmentioning

confidence: 99%

Machine Learning Classification Reveals Robust Morphometric Biomarker of Glial and Neuronal Arbors

Akram

Wei

Ascoli

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Notably, even for well-studied neurons, the terms “class” and “type” are used rather interchangeably, without universally sharp defining criteria. These definitions are sometimes elusive in vertebrate cell types too, and may affect cell classifications as discussed below ( Tasic et al, 2018 ; Zhang et al, 2021 ).…”

Section: Cell Classifications In the Mapped Nervous System Of ...mentioning

confidence: 99%

Open Frontiers in Neural Cell Type Investigations; Lessons From Caenorhabditis elegans and Beyond, Toward a Multimodal Integration

Rapti

2022

Front. Neurosci.

View full text Add to dashboard Cite

Nervous system cells, the building blocks of circuits, have been studied with ever-progressing resolution, yet neural circuits appear still resistant to schemes of reductionist classification. Due to their sheer numbers, complexity and diversity, their systematic study requires concrete classifications that can serve reduced dimensionality, reproducibility, and information integration. Conventional hierarchical schemes transformed through the history of neuroscience by prioritizing criteria of morphology, (electro)physiological activity, molecular content, and circuit function, influenced by prevailing methodologies of the time. Since the molecular biology revolution and the recent advents in transcriptomics, molecular profiling gains ground toward the classification of neurons and glial cell types. Yet, transcriptomics entails technical challenges and more importantly uncovers unforeseen spatiotemporal heterogeneity, in complex and simpler nervous systems. Cells change states dynamically in space and time, in response to stimuli or throughout their developmental trajectory. Mapping cell type and state heterogeneity uncovers uncharted terrains in neurons and especially in glial cell biology, that remains understudied in many aspects. Examining neurons and glial cells from the perspectives of molecular neuroscience, physiology, development and evolution highlights the advantage of multifaceted classification schemes. Among the amalgam of models contributing to neuroscience research, Caenorhabditis elegans combines nervous system anatomy, lineage, connectivity and molecular content, all mapped at single-cell resolution, and can provide valuable insights for the workflow and challenges of the multimodal integration of cell type features. This review reflects on concepts and practices of neuron and glial cells classification and how research, in C. elegans and beyond, guides nervous system experimentation through integrated multidimensional schemes. It highlights underlying principles, emerging themes, and open frontiers in the study of nervous system development, regulatory logic and evolution. It proposes unified platforms to allow integrated annotation of large-scale datasets, gene-function studies, published or unpublished findings and community feedback. Neuroscience is moving fast toward interdisciplinary, high-throughput approaches for combined mapping of the morphology, physiology, connectivity, molecular function, and the integration of information in multifaceted schemes. A closer look in mapped neural circuits and understudied terrains offers insights for the best implementation of these approaches.

show abstract

“…Learning representations from raw morphological data has only been explored in one further study so far. Zhang et al (2021) recently trained LSTMs on morphological reconstruction files coupled with convolutional networks on density maps to successfully classify several types of rat neurons. Their model does not allow for the generation of new data, however.…”

Section: Related Workmentioning

confidence: 99%

MorphVAE: Generating Neural Morphologies from 3D-Walks using a Variational Autoencoder with Spherical Latent Space

Laturnus

Berens

2021

Preprint

View full text Add to dashboard Cite

For the past century, the anatomy of a neuron has been considered one of its defining features: The shape of a neuron`s dendrites and axon fundamentally determines what other neurons it can connect to. These neurites have been described using mathematical tools e.g. in the context of cell type classification, but generative models of these structures have only rarely been proposed and are often computationally inefficient. Here we propose MORPHVAE, a sequence-to-sequence variational autoencoder with spherical latent space as a generative model for neural morphologies. The model operates on walks within the tree structure of a neuron and can incorporate expert annotations on a subset of the data using semi-supervised learning. We develop our model on artificially generated toy data and evaluate its performance on dendrites of excitatory cells and axons of inhibitory cells of mouse motor cortex (M1) and dendrites of retinal ganglion cells. We show that the learned latent feature space allows for better cell type discrimination than other commonly used features. By sampling new walks from the latent space we can easily construct new morphologies with a specified degree of similarity to their reference neuron, providing an efficient generative model for neural morphologies.

show abstract

Neuron type classification in rat brain based on integrative convolutional and tree-based recurrent neural networks

Cited by 27 publications

References 34 publications

Machine Learning Classification Reveals Robust Morphometric Biomarker of Glial and Neuronal Arbors

Machine Learning Classification Reveals Robust Morphometric Biomarker of Glial and Neuronal Arbors

Open Frontiers in Neural Cell Type Investigations; Lessons From Caenorhabditis elegans and Beyond, Toward a Multimodal Integration

MorphVAE: Generating Neural Morphologies from 3D-Walks using a Variational Autoencoder with Spherical Latent Space

Contact Info

Product

Resources

About