SpineTool 2.0: A Deep Learning Platform with an Interactive User Interface for Dendritic Spine Segmentation and Morphological Analysis

Topic: Disease mechanisms in lab and animal studies

Quantitative analysis of dendritic spine morphology is central to understanding synaptic function and its disruption in neurological disease, yet reliable measurement depends on accurate segmentation of spines from fluorescence microscopy images — a task still limited by faint or discontinuous spine necks and by tools that require extensive manual correction. Here we present SpineTool 2.0, an integrated software platform that combines a cascade of deep learning-based segmentation models four three-dimensional U-Net networks performing dendrite segmentation, neck reconstruction, resolution enhancement, and spine segmentation with an interactive, Napari-based user interface that provides three-dimensional visualization together with a comprehensive set of manual correction, editing, and project-management tools. The platform supports both fully automatic (AI-based) and semi-automatic workflows across heterogeneous in vitro, ex vivo, and in vivo confocal and two-photon datasets without dataset-specific retraining. Benchmarked against existing dendritic segmentation tools, SpineTool 2.0 achieves markedly higher accuracy in both dendritic shaft and individual spine segmentation, while substantially reducing the loss of spine necks and the merging of adjacent spines that limit current approaches. Applied to in vitro primary hippocampal neurons from a low-dose amyloid-β model of Alzheimer’s disease, SpineTool 2.0 detected a larger and more complete population of dendritic spines than the original SpineTool and revealed statistically significant morphological changes under Aβ exposure that had previously gone undetected. These results show that improvements in automatic segmentation accuracy can directly alter the biological conclusions drawn from morphometric analysis, underscoring the need for accurate, correctable segmentation tools in synaptic biology. SpineTool 2.0, together with its training and test datasets, is released as open-source software.

Key findings

• Applied to in vitro primary hippocampal neurons from a low-dose amyloid-β model of Alzheimer’s disease, SpineTool 2.0 detected a larger and more complete population of dendritic spines than the original SpineTool and revealed statistically significant morphological changes under Aβ exposure that had previously gone undetected.
• These results show that improvements in automatic segmentation accuracy can directly alter the biological conclusions drawn from morphometric analysis, underscoring the need for accurate, correctable segmentation tools in synaptic biology.

Study details

• Evidence strengthLow

Read the original

• DOI: 10.21203/rs.3.rs-10996498/v1
• Original source →

Free copy unknown: no free-copy check has an answer for this record yet. That does not mean it is paywalled.

@article{antonova2026-e8cd5270,
  author  = {Antonova, D. and Gurenko, S. and Pegova, M. and Ustinova, A. and Shteynberg, G. and Chukanov, V. and Pchitskaya, E.},
  title   = {{SpineTool 2.0: A Deep Learning Platform with an Interactive User Interface for Dendritic Spine Segmentation and Morphological Analysis}},
  journal = {Research Square},
  year    = {2026},
  doi     = {10.21203/rs.3.rs-10996498/v1},
  url     = {https://doi.org/10.21203/rs.3.rs-10996498/v1}
}

Related studies in Disease mechanisms

• Artificial intelligence in neuropathology: deep learning-based assessment of tauopathy Laboratory investigation; a journal of technical methods and pathology · 2019-02-15 · usefulness score 38/100
• Direct evidence for dendritic spine compensation and regeneration in Alzheimer's disease models Alzheimer's & dementia : the journal of the Alzheimer's Association · 2025-10 · usefulness score 46/100
• Amyloid-β Causes NMDA Receptor Dysfunction and Dendritic Spine Loss through mGluR1 and AKAP150-Anchored Calcineurin Signaling The Journal of neuroscience : the official journal of the Society for Neuroscience · 2024-09-11 · usefulness score 43/100
• α-Synuclein triggers cofilin pathology and dendritic spine impairment via a PrPC-CCR5 dependent pathway Cell death & disease · 2024-04-13 · usefulness score 43/100
• Pin1 mediates Aβ42-induced dendritic spine loss Science signaling · 2018-03-20 · usefulness score 39/100
• Dendritic spine pathology and thrombospondin-1 deficits in Down syndrome Free radical biology & medicine · 2017-09-28 · usefulness score 39/100

Similar studies

• Synaptic Compensatory Plasticity in Alzheimer's Disease 2023 · Disease mechanisms
• Interpretable classification of Alzheimer's disease pathologies with a convolutional neural network pipeline 2019 · Digital health tools
• Overcoming segmentation bottlenecks in brain tissue for spatial proteomic inference at population scale 2026 · Genetics
• Dendritic spine abnormalities in amyloid precursor protein transgenic mice demonstrated by gene transfer and intravital multiphoton microscopy 2005 · Disease mechanisms
• Dendritic spines: Revisiting the physiological role 2019 · Disease mechanisms
• Generalizing Amyloid β Pathology from Postmortem Human Brains with Vision Foundation Models for Lesion Quantification 2026 · Vascular contributions

More on Disease mechanisms

• A review that covers this ground Interface between tauopathies and synucleinopathies: a tale of two proteins Annals of neurology · 2006-03-01 · usefulness score 40/100
• Newest in this subject Aβ42 oligomers disrupt calcium homeostasis via Piezo1 and induce neurotoxicity in differentiated SH-SY5Y cells Research Square · 2026-10-29 · Preprint — not yet peer-reviewed · usefulness score 16/100
• Newest in this subject Neuronal Hyperactivity Induces Tau Phosphorylation Selectively in Excitatory Pyramidal Neurons Which Contributes to GluA2 Regulation bioRxiv · 2026-10-28 · Preprint — not yet peer-reviewed · usefulness score 16/100

See all studies in Disease mechanisms →