The abstract is read at the publisher; the summary is JClub's.
A deep-learning pipeline enabled complete 3D mapping of 27,880 spiral ganglion neurons in one adult human cochlea, with 97% accuracy versus manual counts.
- Why it matters: Prior human temporal-bone studies estimated spiral ganglion neuron numbers from only every 10th tissue section, limiting distribution and tonotopic mapping.
- What they did: The authors digitized 421 consecutive 20 µm celloidin-embedded H&E sections from one adult cochlea, trained UNet++ models on manually labeled nuclei, and compared automated counts to histopathologist manual counts.
- The result: Automated detection counted SGNs in 191 sections with 97% accuracy, and omitting sections to every 10th preserved median inferred total count accuracy of 98%, enabling full 3D visualization with tonotopic and basal-to-apical distributions.
The findingWhy it mattersWhat they didThe result
AI summary of the abstract · may contain errors · not medical advice · read the paper.