OSCR

Autoencoders for unsupervised analysis of rat myeloarchitecture.

Code ↔ Paper

6 matches between paragraphs of the paper and lines of its authors' code, computed by the harvester (lexical-v1). Click a colored paragraph or line to see its counterpart.

The 6 matches
  1. [1] § Materials and methods › Feature extraction › Autoencoders (AE) ↔ AE/architectures/AE_model128.py, lines 65–106 · score 0.59 · ReLU, sigmoid, kernel, stride, decoder, linear
  2. [2] § Materials and methods › Feature extraction › Autoencoders (AE) ↔ AE/architectures/AE_model256.py, lines 67–110 · score 0.59 · ReLU, sigmoid, kernel, stride, decoder, linear
  3. [3] § Materials and methods › Feature extraction › Autoencoders (AE) ↔ AE/scripts/train_AE_model128_part1.py, lines 171–186 · score 0.56 · MSE loss function, Adam, patience, optimizer, epoch, trained
  4. [4] § Materials and methods › Feature extraction › Autoencoders (AE) ↔ AE/scripts/train_AE_model256_part1.py, lines 172–187 · score 0.56 · MSE loss function, Adam, patience, optimizer, epoch, trained
  5. [5] § Materials and methods › Feature analysis › Unsupervised clustering ↔ scripts_analysis/featureExtraction_clustering/D/color_dictionary_27_3_cov.py, lines 10–71 · score 0.56 · symmetric KL divergence, assignment, covariance, spherical, match, clusters
  6. [6] § Materials and methods › Feature analysis › Unsupervised clustering ↔ scripts_analysis/featureExtraction_clustering/D/colorDictionary_4animals.py, lines 10–71 · score 0.55 · symmetric KL divergence, assignment, covariance, spherical, match, clusters

Paper

Loaded from Europe PMC by your browser, not stored by OSCR: doi.org · Europe PMC

The paper is loaded when this pane is shown.

The authors' code

Python · 126 lines · 3.7 KB · MIT · 1 match

  1. import torch.nn as nn
  3. KERNEL_SIZE = 3
  4. STRIDE = 2
  5. PADDING = 1
  6. A = 8
  7. B = 16
  8. C = 32
  9. D = 64
  10. E = 128
  11. F = 256
  12. G = 512
  14. LS = 256
  16. # Define the Autoencoder class
  17. class Encoder(nn.Module):
  18. # input shape in tensor form: (1, 128, 128), i.e., patches of size 128x128
  20. def __init__(self, output_dim=LS):
  21. super(Encoder, self).__init__()
  23. self.output_dim = output_dim
  25. # Encoder
  26. self.encoder = nn.Sequential(
  28. nn.Conv2d(1, A, KERNEL_SIZE, STRIDE, PADDING),
  29. nn.BatchNorm2d(A),
  30. nn.ReLU(),
  31. nn.Conv2d(A, B, KERNEL_SIZE, STRIDE, PADDING),
  32. nn.BatchNorm2d(B),
  33. nn.ReLU(),
  34. nn.Conv2d(B, C, KERNEL_SIZE, STRIDE, PADDING),
  35. nn.BatchNorm2d(C),
  36. nn.ReLU(),
  37. nn.Conv2d(C, D, KERNEL_SIZE, STRIDE, PADDING),
  38. nn.BatchNorm2d(D),
  39. nn.ReLU(),
  40. nn.Conv2d(D, E, KERNEL_SIZE, STRIDE, PADDING),
  41. nn.BatchNorm2d(E),
  42. nn.ReLU(),
  43. nn.Conv2d(E, F, KERNEL_SIZE, STRIDE, PADDING),
  44. nn.BatchNorm2d(F),
  45. nn.ReLU(),
  46. nn.Conv2d(F, G, KERNEL_SIZE, STRIDE, PADDING),
  47. nn.BatchNorm2d(G),
  48. nn.ReLU(),
  49. nn.Flatten(),
  50. nn.Linear(G, LS),
  51. # nn.ReLU(),
  52. # nn.Linear(F, LS)
  54. )
  56. def forward(self, x):
  57. x = x.float()
  58. x = self.encoder(x)
  59. return x
  61. # ---------------------------------------------------------------------------------------------------------------------
  62. # ---------------------------------------------------------------------------------------------------------------------
  65. class Decoder(nn.Module):
  67. def __init__(self, input_dim=LS):
  68. super(Decoder, self).__init__()
  70. self.input_dim = input_dim
  72. self.linear1 = nn.Linear(LS, G)
  73. self.relu = nn.ReLU()
  74. # self.linear2 = nn.Linear(F, G)
  77. # Decoder
  78. self.decoder = nn.Sequential(
  79. nn.ReLU(),
  80. nn.ConvTranspose2d(G, F, KERNEL_SIZE, STRIDE, PADDING, output_padding=1),
  81. nn.ReLU(),
  82. nn.ConvTranspose2d(F, E, KERNEL_SIZE, STRIDE, PADDING, output_padding=1),
  83. nn.ReLU(),
  84. nn.ConvTranspose2d(E, D, KERNEL_SIZE, STRIDE, PADDING, output_padding=1),
  85. nn.ReLU(),
  86. nn.ConvTranspose2d(D, C, KERNEL_SIZE, STRIDE, PADDING, output_padding=1),
  87. nn.ReLU(),
  88. nn.ConvTranspose2d(C, B, KERNEL_SIZE, STRIDE, PADDING, output_padding=1),
  89. nn.ReLU(),
  90. nn.ConvTranspose2d(B, A, KERNEL_SIZE, STRIDE, PADDING, output_padding=1),
  91. nn.ReLU(),
  92. nn.ConvTranspose2d(A, 1, KERNEL_SIZE, STRIDE, PADDING, output_padding=1),
  93. nn.Sigmoid() # ensures that the output values are in range (0,1)
  94. )
  95. # self.output = nn.Conv2d(1, 1, kernel_size=1, stride=1) # no tenc clar que fa aixo
  97. def forward(self, x):
  98. x = x.float()
  99. x = self.linear1(x)
  100. # x = self.relu(x)
  101. # x = self.linear2(x)
  102. # reshape 3d tensor to 4d tensor
  103. x = x.reshape(x.shape[0], G, 1, 1)
  104. x = self.decoder(x)
  105. # return self.output(x)
  106. return x
  109. # ---------------------------------------------------------------------------------------------------------------------
  110. # ---------------------------------------------------------------------------------------------------------------------
  113. # putting them together:
  114. class AutoEncoder(nn.Module):
  116. def __init__(self):
  117. super(AutoEncoder, self).__init__()
  118. self.encoder = Encoder(output_dim=LS)
  119. self.decoder = Decoder(input_dim=LS)
  121. def forward(self, x):
  122. return self.decoder(self.encoder(x))
  125. def instance():
  126. return AutoEncoder()

AE_model128.py at commit ac529b0, under MIT · at the source

Overview

Authors: Melina Estela1, Raimo A Salo1, Isabel San Martín Molina1, Omar Narvaez1, Ville Kolehmainen2, Jussi Tohka1, Alejandra Sierra1
  1. A.I. Virtanen Institute for Molecular Sciences, University of Eastern Finland, Kuopio, Finland
  2. Department of Applied Physics, University of Eastern Finland, Kuopio, Finland
Institutions: University of Eastern Finland (Finland)
Journal: Brain structure & function, volume 231, issue 8, article 141
Dates: received 2 April 2026; accepted 17 July 2026; published online 26 September 2026; in print 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1007/s00429-026-03166-w · PMID 42799916 · PMCID PMC13616018 · OpenAlex W7214462535
Open access: hybrid, a free copy (OpenAlex)
Status: code verified
Categories: histology / microscopy (modality), rat (organism), traumatic brain injury (population), cellular / molecular (subfield)
Methods: Connectivity, Smoothing, state filtering, decompositions, Machine learning, Statistics
Keywords: Histological image analysis, Convolutional autoencoders, PCA, DINOv3, Tissue clustering, Myelin, Traumatic brain injury, Computational pathology
MeSH: Brain*, Image Processing, Computer-Assisted*, Myelin Sheath*, Unsupervised Machine Learning*, White Matter*, Animals, Autoencoder, Brain Injuries, Traumatic, Cluster Analysis, Clustering Algorithms, Gray Matter, Male, Principal Component Analysis, Rats (* major topic)
Topic: Advanced Neuroimaging Techniques and Applications (Radiology, Nuclear Medicine and Imaging, Medicine), according to OpenAlex
Funding: Itä-Suomen Yliopisto; European Commission; Kuopion Yliopistollinen Sairaala; China Scholarship Council; Biocenter Finland
Citations: not cited yet (Europe PMC); 48 references in the paper

Abstract

Quantitative assessment of brain histology is often constrained by predefined feature sets and labor-intensive manual annotations. To overcome these limitations, we employed unsupervised deep learning to automatically extract and quantify tissue organizational patterns from myelin-stained rat brain sections without the need for prior labeling. We evaluated nonlinear convolutional autoencoders (AEs) against linear principal component analysis (PCA) and DINOv3, a self-supervised foundation model applied without task-specific fine-tuning, for feature representation, followed by clustering with Gaussian mixture models. Compared to both PCA and DINOv3, AEs better preserved fine axonal architecture and produced more consistent and interpretable tissue clusters across hierarchical levels. The resulting clusters revealed anatomically meaningful tissue organization, including different white matter densities and grey matter subregions. When applied to tissue from sham and mild traumatic brain injury animals, AE-derived features also captured pathology-related alterations, such as white matter loss and injury-specific microstructural changes. These findings suggest that task-specific unsupervised training on domain-specific histological data can automatically characterize tissue organization at multiple scales and detect pathological changes, offering a scalable, annotation-free approach to computational neuropathology.

Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1007/s00429-026-03166-w.

Reproduced under the paper's license (CC BY), from the paper cited above.

Repository

Its files are read in the Code ↔ Paper reader above, with 6 matches between paragraphs and lines of code.

UEF-Multiscale-Imaging/AE-and-GMM-myelin-stained-tissue

License: MIT
State: the link answers, verified on 28 September 2026
Evidence: files inventoried
Commit: ac529b0afa16fa6971c09b207408305d5100f806, 10 February 2026
Languages: Python (50), Shell (1)
Size: 66 files, 51 scripts
Software Heritage: not archived
Found in: “Data availability”
Holds: README, license file
Not found: CITATION.cff, environment file, tests, continuous integration, documentation
Tools: NumPy (40 files), Pillow (34 files), Matplotlib (21 files), PyTorch (13 files), OpenCV (10 files), scikit-image (8 files), scikit-learn (4 files), SciPy (4 files)
Availability: 1 check, the latest on 28 September 2026: the link answers
  • 28 September 2026: the link answers
53 files
At the source: github.com/UEF-Multiscale-Imaging/AE-and-GMM-myelin-stained-tissue

The paper's code and data availability statement is in the Data section.

Tracing map

Proposed by the machine: these links were found in the paper and verified at the source, without human review. The map will receive a Zenodo DOI once one of the paper's authors has validated it with their ORCID.

What the map holds:

  • 1 repository of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 51 scripts, each with its path and the digest of its content;
  • 6 matches between paragraphs of the paper and lines of the code (method lexical-v1);
  • neither the text of the paper nor the code itself.

Its JSON (tracing-map.json) is deposited on Zenodo with its DOI once the map is validated.

Data

No dataset and no data link were found in the paper.

Data Availability Statement

Our source code is available on GitHub (https://github.com/UEF-Multiscale-Imaging/AE-and-GMM-myelin-stained-tissue). Data used in this paper will be shared by the lead contact, Alejandra Sierra upon request. Requests are typically answered within 2 weeks.

During the preparation of this work, the authors used Claude to help with refining language, clarity and organizing sections. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Reproduced under the paper's license (CC BY), from the paper cited above.

Versions

The history of this record: each version stored by the harvester or made by a correction of its authors or of the maintainers of its code, and what changed in its facts. The texts of the paper (its abstract, its availability statements) are not part of it; versions that changed only those are not listed.

Version 1, 28 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 7 authors, 8 keywords, 14 MeSH terms, 5 funders, 36 references.

Cite

This paper

Estela, M., Salo, R. A., San Martín Molina, I., Narvaez, O., Kolehmainen, V., Tohka, J., & Sierra, A. (2026). Autoencoders for unsupervised analysis of rat myeloarchitecture. Brain structure & function, 231(8), 141. https://doi.org/10.1007/s00429-026-03166-w

BibTeX

@article{estela2026autoencoders,
author = {Estela, Melina and Salo, Raimo A and San Martín Molina, Isabel and Narvaez, Omar and Kolehmainen, Ville and Tohka, Jussi and Sierra, Alejandra},
title = {{Autoencoders for unsupervised analysis of rat myeloarchitecture}},
journal = {Brain structure \& function},
year = {2026},
month = sep,
volume = {231},
number = {8},
pages = {141},
publisher = {Springer Science+Business Media},
issn = {1863-2653},
doi = {10.1007/s00429-026-03166-w},
url = {https://doi.org/10.1007/s00429-026-03166-w},
pmid = {42799916},
pmcid = {PMC13616018}
}

RIS

TY - JOUR
AU - Estela, Melina
AU - Salo, Raimo A
AU - San Martín Molina, Isabel
AU - Narvaez, Omar
AU - Kolehmainen, Ville
AU - Tohka, Jussi
AU - Sierra, Alejandra
TI - Autoencoders for unsupervised analysis of rat myeloarchitecture
T2 - Brain structure & function
J2 - Brain Struct Funct
PY - 2026
DA - 2026/09/26
VL - 231
IS - 8
SP - 141
SN - 1863-2653
PB - Springer Science+Business Media
DO - 10.1007/s00429-026-03166-w
UR - https://doi.org/10.1007/s00429-026-03166-w
LA - en
ER -

CSL-JSON

{
"id": "10.1007/s00429-026-03166-w",
"type": "article-journal",
"title": "Autoencoders for unsupervised analysis of rat myeloarchitecture",
"container-title": "Brain structure & function",
"author": [
{
"family": "Estela",
"given": "Melina"
},
{
"family": "Salo",
"given": "Raimo A"
},
{
"family": "San Martín Molina",
"given": "Isabel"
},
{
"family": "Narvaez",
"given": "Omar"
},
{
"family": "Kolehmainen",
"given": "Ville"
},
{
"family": "Tohka",
"given": "Jussi"
},
{
"family": "Sierra",
"given": "Alejandra"
}
],
"container-title-short": "Brain Struct Funct",
"volume": "231",
"issue": "8",
"page": "141",
"DOI": "10.1007/s00429-026-03166-w",
"PMID": "42799916",
"PMCID": "PMC13616018",
"ISSN": "1863-2653",
"publisher": "Springer Science+Business Media",
"URL": "https://doi.org/10.1007/s00429-026-03166-w",
"language": "en",
"issued": {
"date-parts": [
[
2026,
9,
26
]
]
}
}

The tracing map gets a citation of its own once an author has validated it and it has a DOI.

Similar papers

The papers with a page that share the most with this one: the tools found in their code, their categories, datasets, cited references and authors, the rarest counting most.

[1] doi:10.1093/jnen/nlaf152 [code]
Clinical and pathologic correlations of machine learning quantification of Aβ deposits across 3 brain regions of decedents with Alzheimer disease.
Journal: Journal of neuropathology and experimental neurology
In common: OpenCV, scikit-image, Pillow, 5 other tools, 1 reference
[2] doi:10.1016/j.patter.2026.101538 [code]
A multi-modal foundation model for brain disease diagnosis and medical imaging.
Journal: Patterns (New York, N.Y.)
In common: OpenCV, scikit-image, Pillow, 5 other tools, 1 reference
[3] doi:10.1038/s41598-026-57519-w [code]
Automated segmentation of neurons and spinal cord structures in immunofluorescence images using SpineDL.
Journal: Scientific reports
In common: OpenCV, scikit-image, Pillow, 5 other tools, histology / microscopy, 1 reference
[4] doi:10.1371/journal.pcbi.1013499 [code]
VesiclePy: A machine learning vesicle analysis toolbox for volume electron microscopy.
Journal: PLoS computational biology
In common: OpenCV, scikit-image, Pillow, 5 other tools, histology / microscopy, 1 reference
[5] doi:10.1016/j.isci.2026.116168 [code]
See the small lesions: Frequency-guided spatial debiasing GAN for multimodal medical image fusion.
Journal: iScience
In common: OpenCV, scikit-image, Pillow, 5 other tools, 1 reference
[6] doi:10.1371/journal.pcbi.1014263 [code]
MIRAGE: Robust multi-modal architectures translate fMRI-to-image models from vision to mental imagery.
Journal: PLoS computational biology
In common: OpenCV, scikit-image, Pillow, 5 other tools, 1 reference
[7] doi:10.1371/journal.pcbi.1014571 [code]
SynAPSeg: A novel dataset and image analysis framework for deep learning-based synapse detection and quantification.
Journal: PLoS computational biology
In common: OpenCV, scikit-image, Pillow, 5 other tools, 1 reference
[8] doi:10.1002/alz.71649 [code]
Postmortem brain MRI reveals differential associations of subcortical and limbic volumes with cortical thinning and neurodegenerative pathologies.
Journal: Alzheimer's & dementia : the journal of the Alzheimer's Association
In common: OpenCV, scikit-image, Pillow, 5 other tools, histology / microscopy, cellular / molecular
[9] doi:10.1038/s41598-026-50475-5 [code]
Segmentation and classification of hippocampal subregions using multi-task generative adversarial networks.
Journal: Scientific reports
In common: OpenCV, Pillow, PyTorch, 4 other tools, histology / microscopy, 1 reference
[10] doi:10.1016/j.isci.2026.116206 [code]
Gut distension evokes rapid neural dynamics in vagal and hindbrain populations of larval zebrafish.
Journal: iScience
In common: OpenCV, scikit-image, Pillow, 5 other tools, 1 reference

Contribute

The authors of this paper can claim it, correct its record and validate its tracing map, and the maintainers of its code (its owner, or a public member of its organization) correct what it says of their repository; anyone signed in can ask for its removal. Every request goes to OSCR's own machine, which answers it; your account page follows them.

Sign in with ORCID to claim this paper as one of its authors, correct its record or validate its tracing map: when the paper's metadata lists your ORCID iD, you are recognized at once. Maintainers of its code: sign in with GitHub, then claim the repository on your account page.

Request its removal

To ask OSCR to remove this record, the copies of its authors' scripts or its tracing map, use the removal request page: signed in, you say who you are, what to remove and why, then review and confirm the request. Published rules decide every request (how).

Discussion, reproductions, activity

Discussion: questions and error reports about this paper and its code, from signed-in readers and its authors. It opens with sign-in.

Reproductions: reports from readers who ran the authors' code: what they reproduced, with which environment, commit and data. It opens with sign-in.

Activity: what happens around this paper: new versions of its record, its map's validation, discussions and reproductions. It opens with sign-in.