Robust neural circuit reconstruction from serial electron microscopy with convolutional recurrent networks

TL;DR

This paper introduces a new neural network architecture that improves the generalization of neuron segmentation across different brain tissues, addressing limitations of existing methods in connectomics.

Contribution

The paper presents a novel convolutional recurrent neural network that enhances generalization in neuron reconstruction, and establishes new state-of-the-art results on a challenging, tissue-agnostic connectomics benchmark.

Findings

Current methods perform poorly on tissue-agnostic tasks

The proposed network outperforms existing approaches on the STAR challenge

The architecture advances automated, generalizable connectomics analysis

Abstract

Recent successes in deep learning have started to impact neuroscience. Of particular significance are claims that current segmentation algorithms achieve "super-human" accuracy in an area known as connectomics. However, as we will show, these algorithms do not effectively generalize beyond the particular source and brain tissues used for training -- severely limiting their usability by the broader neuroscience community. To fill this gap, we describe a novel connectomics challenge for source- and tissue-agnostic reconstruction of neurons (STAR), which favors broad generalization over fitting specific datasets. We first demonstrate that current state-of-the-art approaches to neuron segmentation perform poorly on the challenge. We further describe a novel convolutional recurrent neural network module that combines short-range horizontal connections within a processing stage and long-range top-down connections between stages. The resulting architecture establishes the state of the art on the STAR challenge and represents a significant step towards widely usable and fully-automated connectomics analysis.