RoboEM: automated 3D flight tracing for synaptic-resolution connectomics
Problem
Reconstructing neuronal networks from 3D electron microscopy data requires costly manual proofreading, especially for thin axons.
Why it matters
Manual proofreading makes mapping brain wiring at synaptic scale prohibitively expensive and slow.
What the AI did
RoboEM, an AI-based self-steering 3D 'flight' system, navigates along neurites using only 3D-EM data, improving automated segmentation and replacing manual proofreading.
Results
Substantially improved state-of-the-art segmentations on mouse and human cortex; ~400-fold lower annotation cost than manual error correction
Impact
Applied to mouse and human cortical data, it can replace manual proofreading for complex connectomic analyses.
Record
| Venue | Nature Methods |
|---|---|
| Published | 2024 |
| Work type | article |
| Authors | Martin Schmidt, Alessandro Motta, Meike Sievers, Moritz N. Helmstaedter |
| Citations | 30 (OpenAlex, at import time) |
| References | 92 |
| Model or method | RoboEM (AI-based self-steering system) |
| Organization | not stated in source |
| Primary source | https://doi.org/10.1038/s41592-024-02226-5 |
| Open access | hybrid |
| Keywords | connectomics, electron microscopy, automated tracing, proofreading, axons |
Importance rating
significant — thousands to millions affected
Abstract states manual proofreading is required for many connectomic analyses; RoboEM yields ~400-fold lower annotation cost on mouse and human cortex, addressing a foundational neuroscience bottleneck.