Related Experiment Video
Updated: Jul 11, 2025
06:08
Automatic Identification of Dendritic Branches and their Orientation
Published on: September 17, 2021
2.0K
AimSeg: A machine-learning-aided tool for axon, inner tongue and myelin segmentation
Pau Carrillo-Barberà1,2,3,4, Ana Maria Rondelli5,6, Jose Manuel Morante-Redolat1,2,3
1Centro de Investigación Biomédica en Red sobre Enfermedades Neurodegenerativas (CIBERNED), Universitat de València, Valencia, Spain.
Plos Computational Biology
|November 17, 2023
Summary
AimSeg is a new bioimage analysis tool that automates the segmentation of axons and myelin from electron microscopy images. This tool accurately quanties uncompacted myelin, improving myelin analysis and reducing processing time.
Area of Science:
- Neuroscience
- Biotechnology
- Medical Imaging
Background:
- Electron microscopy (EM) is crucial for analyzing myelinated nerve fibers in the central and peripheral nervous system.
- The g-ratio, a standard measure of myelin quality, is traditionally manually calculated from EM images, a process that is time-consuming and lacks reproducibility.
- Existing methods often overlook the inner tongue, an uncompacted myelin region vital for myelin growth.
Purpose of the Study:
- To develop AimSeg, a novel bioimage analysis tool for segmenting axons, inner tongues, and myelin sheaths.
- To improve the accuracy and efficiency of myelin analysis by incorporating inner tongue quantification.
- To provide a user-friendly platform for both automated and assisted segmentation of neural tissue EM images.
Main Methods:
- Developed AimSeg, a bioimage analysis software implemented in Fiji, utilizing machine learning classifiers trained in ilastik.
- Trained machine learning classifiers on transmission EM (TEM) images of healthy and remyelinating nerve tissue.
- Validated AimSeg's segmentation performance on TEM data from various sample types, comparing automated and user-assisted modes.
Main Results:
- AimSeg demonstrated good performance in segmenting axons, inner tongues, and myelin from TEM images.
- User-assisted segmentation with AimSeg showed potential for further accuracy improvements with minimal user input.
- The tool significantly reduced analysis time compared to traditional manual measurements.
- AimSeg facilitates the creation of large, high-quality datasets for training advanced deep learning models.
Conclusions:
- AimSeg offers a unique, efficient, and reproducible method for analyzing myelin structure and quality using EM images.
- The tool's ability to quantify uncompacted myelin provides novel insights into myelin growth and development.
- AimSeg streamlines neurobiological research by reducing manual labor and enhancing data quality for myelin studies.
