OSCR

Machine learning-based combination of the central vein sign, cortical lesions and paramagnetic rim lesions: a web-based tool for the diagnosis of multiple sclerosis.

Code ↔ Paper

7 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 7 matches
  1. [1] § Materials and methods › Machine learning framework and evaluation protocol ↔ step_1_find_best_algorithm_for_each_combination_of_features.py, lines 64–114 · score 0.85 · neighbours classifier, decision tree, logistic regression, random forest, boosting, balanced accuracy
  2. [2] § Materials and methods › Machine learning framework and evaluation protocol ↔ interpretability_plots.py, lines 52–76 · score 0.62 · logistic regression, random forest, tree, SVC, XGB, classification
  3. [3] § Materials and methods › Machine learning framework and evaluation protocol ↔ step_6_retrain_models_on_full_training_set.py, lines 50–109 · score 0.57 · F1 score, full training, retrained, balanced accuracy, precision, sensitivity
  4. [4] § Results ↔ step_1_find_best_algorithm_for_each_combination_of_features.py, lines 64–114 · score 0.56 · random forest classifier, logistic regression, best model, prediction
  5. [5] § Results ↔ interpretability_plots.py, lines 207–253 · score 0.55 · random forest classifier, logistic regression, SHAP, prediction, model
  6. [6] § Materials and methods › Machine learning framework and evaluation protocol ↔ step_2_find_subset_of_algorithm_combination_pairs_outperforming_dis.py, lines 159–247 · score 0.53 · simplified variable, algorithm combination pairs, CL1, PRL1, DIS, balanced accuracy
  7. [7] § Results ↔ step_2_find_subset_of_algorithm_combination_pairs_outperforming_dis.py, lines 159–247 · score 0.51 · outperformed DIS, algorithm combination pairs, balanced accuracy, simplified, variables, CVS

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 · 188 lines · 7.4 KB · no license · 2 matches

  1. import pandas as pd
  2. import json
  3. import numpy as np
  5. from sklearn.model_selection import train_test_split
  6. from sklearn.preprocessing import StandardScaler
  7. from sklearn.metrics import accuracy_score, confusion_matrix, classification_report, balanced_accuracy_score, f1_score
  8. from sklearn.linear_model import LogisticRegression
  9. from sklearn.svm import SVC
  10. from sklearn.ensemble import RandomForestClassifier
  11. from sklearn.model_selection import GridSearchCV
  12. from sklearn.pipeline import Pipeline
  13. from sklearn.neighbors import KNeighborsClassifier
  14. from sklearn.tree import DecisionTreeClassifier
  15. from xgboost import XGBClassifier
  16. from sklearn.ensemble import AdaBoostClassifier
  17. from itertools import chain, combinations
  18. import warnings
  19. from utils import *
  20. warnings.filterwarnings("ignore", category=FutureWarning)
  21. from sklearn.model_selection import cross_val_predict, StratifiedKFold, KFold
  22. from sklearn.metrics import balanced_accuracy_score, confusion_matrix
  23. import numpy as np
  25. global n_model
  26. n_model = 0
  27. np.random.seed(RANDOM_SEED)
  30. def powerset(iterable):
  31. "powerset([1,2,3]) --> (1,) (2,) (3,) (1,2) (1,3) (2,3) (1,2,3)"
  32. s = list(iterable)
  33. return chain.from_iterable(combinations(s, r) for r in range(1, len(s)+1))
  36. def find_best_threshold(y_true, y_pred_proba):
  37. """
  38. Find the best threshold for the given predictions
  39. :param y_true: GT labels
  40. :param y_pred_proba: predicted probabilities
  41. :return: best threshold
  42. """
  43. thresholds = np.linspace(0, 1, 100)
  44. best_threshold = None
  45. best_balanced_accuracy = -1
  46. best_sensitivity = -1
  48. for threshold in thresholds:
  49. y_pred = (y_pred_proba[:, 1] >= threshold).astype(int)
  50. tn, fp, fn, tp = confusion_matrix(y_true, y_pred).ravel()
  51. sensitivity = tp / (tp + fn)
  52. specificity = tn / (tn + fp)
  53. balanced_accuracy = (sensitivity + specificity) / 2
  55. if balanced_accuracy > best_balanced_accuracy or \
  56. (balanced_accuracy == best_balanced_accuracy and sensitivity > best_sensitivity):
  57. best_threshold = threshold
  58. best_balanced_accuracy = balanced_accuracy
  59. best_sensitivity = sensitivity
  61. return best_threshold
  64. def cross_validate_best_model(X, y):
  65. models = [
  66. LogisticRegression(),
  67. SVC(probability=True),
  68. RandomForestClassifier(),
  69. KNeighborsClassifier(),
  70. DecisionTreeClassifier(),
  71. XGBClassifier(),
  72. AdaBoostClassifier(),
  73. ]
  75. best_model = None
  76. best_balanced_accuracy = -1
  77. associated_probability_threshold = None
  79. for model_class in models:
  80. skf = StratifiedKFold(n_splits=10, shuffle=True, random_state=RANDOM_SEED)
  81. # skf = KFold(n_splits=10, shuffle=True, random_state=42)
  82. balanced_accuracies = []
  83. probability_thresholds = []
  85. for train_index, test_index in skf.split(X, y):
  86. X_train, X_test = X.iloc[train_index], X.iloc[test_index]
  87. y_train, y_test = y[X_train.index], y[X_test.index]
  89. if USE_PIPELINE:
  90. model = Pipeline([
  91. ("scaler", StandardScaler()),
  92. ("model", model_class)
  93. ])
  94. else:
  95. model = model_class
  97. model.fit(X_train, y_train)
  98. y_pred_proba = model.predict_proba(X_test)
  99. best_threshold = find_best_threshold(y_test, y_pred_proba)
  100. y_pred = (y_pred_proba[:, 1] >= best_threshold).astype(int)
  101. balanced_accuracies.append(balanced_accuracy_score(y_test, y_pred))
  102. probability_thresholds.append(best_threshold)
  104. avg_balanced_accuracy = np.mean(balanced_accuracies)
  106. if avg_balanced_accuracy > best_balanced_accuracy:
  107. best_model = model_class
  108. best_balanced_accuracy = avg_balanced_accuracy
  109. associated_probability_threshold = np.mean(probability_thresholds)
  111. best_model = str(best_model).split("(")[0]
  112. best_model = 'XGBClassifier()' if "XGBClassifier" in str(best_model) else best_model
  114. return best_model, associated_probability_threshold
  117. def find_best_algorithm_for_specific_combination_of_features(X, y, features):
  118. global n_model
  119. n_model += 1
  120. feats = [f for f in features if f not in ("Select3*-v2NA", "Select6*-v2NA")]
  121. if "Select3*-v2NA" in features:
  122. feats += ["Select3*-v2NA_0.0", "Select3*-v2NA_1.0", "Select3*-v2NA_NA"]
  123. if "Select6*-v2NA" in features:
  124. feats += ["Select6*-v2NA_0.0", "Select6*-v2NA_1.0", "Select6*-v2NA_NA"]
  125. this_X = X[feats].dropna()
  126. this_y = y[this_X.index]
  128. # # #reindex
  129. # # this_X = this_X.reset_index(drop=True)
  130. # # this_y = this_y.reset_index(drop=True)
  131. # this_X, this_y = prepare_data(this_X, this_y)
  133. best_model, best_threshold = cross_validate_best_model(this_X, this_y)
  135. return best_model, best_threshold
  138. def make_combinations(excluded_features):
  139. all_features = [f for f in ALL_FEATURES if f not in excluded_features]
  140. all_combinations = powerset(all_features)
  141. n_combinations = 2 ** len(all_features) - 1
  142. print(f"Number of combinations: {n_combinations}")
  144. # remove all combinations where both Select3*-v2NA and % perivenular les are present
  145. all_combinations = [c for c in all_combinations if not ("Select3*-v2NA" in c and "% perivenular les" in c)]
  146. # remove all combinations where both Select6*-v2NA and % perivenular les are present
  147. all_combinations = [c for c in all_combinations if not ("Select6*-v2NA" in c and "% perivenular les" in c)]
  148. # remove all combinations where both PRL1 and number_PRL are present
  149. all_combinations = [c for c in all_combinations if not ("PRL1" in c and "number_PRL" in c)]
  150. # remove all combinations where both CL1 and CL-count-updated are present
  151. all_combinations = [c for c in all_combinations if not ("CL1" in c and "CL-count-updated" in c)]
  152. # remove all combinations where both Select3*-v2NA and Select6*-v2NA are present
  153. all_combinations = [c for c in all_combinations if not ("Select3*-v2NA" in c and "Select6*-v2NA" in c)]
  155. print(f"Number of combinations after filtering: {len(all_combinations)}")
  157. return all_combinations
  160. def find_best_algorithm_for_each_combination(X, y):
  161. excluded_features = ("age", "sex", "Filippi", "OCB_presence")
  162. all_features = [f for f in ALL_FEATURES if f not in excluded_features]
  163. all_combinations = make_combinations(excluded_features)
  164. best_model_per_combination = []
  166. for i, combination in enumerate(all_combinations):
  167. print(f"Progress: {i + 1}/{len(all_combinations)}")
  168. best_model, best_threshold = find_best_algorithm_for_specific_combination_of_features(X, y, combination)
  169. print(f"Best model for combination {combination}: {best_model}, best threshold: {best_threshold}\n")
  171. feature_presence = {f: f in list(combination) for f in sorted(all_features, key=lambda x: x.upper())}
  172. infos = {
  173. "model": best_model,
  174. "threshold": best_threshold,
  175. }
  176. infos.update(feature_presence)
  177. best_model_per_combination.append(infos)
  179. best_model_per_combination = pd.DataFrame(best_model_per_combination)
  181. return best_model_per_combination
  184. if __name__ == "__main__":
  185. X, y = get_full_dataset(TRAINING_DATA_PATH, dropna=False, return_X_y=True, prepare=True)
  186. best_models = find_best_algorithm_for_each_combination(X, y)
  187. best_models.to_csv(BEST_ALGORITHM_PER_COMBINATION_OF_FEATURES_PATH, index=False)
  188. print("Done.")

step_1_find_best_algorithm_for_each_combination_of_features.py at commit 9551af5, no license · at the source

Overview

Authors: Maxence Wynen1,2, Colin Vanden Bulcke1,2,3, Serena Borrelli2,4, Pedro M Gordaliza5,6, Anna Stölting2, François Guisset2, Clément Cordier1, Maria Sofia Martire7, Agnese Tamanti8, Benoit Macq1, Pascal Sati9, Massimo Filippi7,10,11, Massimiliano Calabrese8, Martina Absinta12,13, Daniel S Reich14, Meritxell Bach Cuadra5,6, Pietro Maggi2,3
ORCID iDs: Maxence Wynen, Colin Vanden Bulcke, Serena Borrelli, Pedro M Gordaliza, Anna Stölting, François Guisset, Pascal Sati, Massimo Filippi, Daniel S Reich
14 affiliations
  1. ICTEAM Institute, Université Catholique de Louvain, 1348 Louvain-la-Neuve, Belgium
  2. Neuroinflammation Imaging Lab (NIL), Université Catholique de Louvain, 1200 Brussels, Belgium
  3. Department of Neurology, Cliniques Universitaires Saint-Luc (CUSL), Université Catholique de Louvain, 1200 Brussels, Belgium
  4. Department of Neurology, Hôpital Erasme, Hôpital Universitaire de Bruxelles, Université Libre de Bruxelles, 1070 Brussels, Belgium
  5. CIBM Center for Biomedical Imaging, CH-1015 Lausanne, Switzerland
  6. Radiology Department, Lausanne University Hospital (CHUV) and University of Lausanne, CH-1011 Lausanne, Switzerland
  7. Neurology Unit, IRCCS San Raffaele Hospital, 20132 Milan, Italy
  8. Department of Neurosciences and Biomedicine and Movement, The Multiple Sclerosis Center of University Hospital of Verona, 37129 Verona, Italy
  9. Department of Neurology, Cedars-Sinai Medical Center, 90048 Los Angeles, CA, USA
  10. Vita-Salute San Raffaele University, Milan, Italy
  11. Neuroimaging Research Unit, Division of Neuroscience, IRCCS San Raffaele Scientific Institute, 20132 Milan, Italy
  12. Experimental Neuropathology Lab, IRCCS Humanitas Research Institute, 20132 Milan, Italy
  13. Department of Biomedical Sciences, Humanitas University, 20072 Milan, Italy
  14. Translational Neuroradiology Section, National Institute of Neurological Disorders and Stroke (NINDS), National Institutes of Health (NIH), 20892 Bethesda, MD, USA
Institutions: UCLouvain (Belgium); Cliniques Universitaires Saint-Luc (Belgium); Université Libre de Bruxelles (Belgium); Erasmus Hospital (Belgium); Centre Hospitalier Universitaire Vaudois (Switzerland); Centre d'Imagerie BioMedicale (Switzerland); University of Lausanne (Switzerland); IRCCS Ospedale San Raffaele (Italy); University of Verona (Italy); Azienda Ospedaliera Universitaria Integrata Verona (Italy); Cedars-Sinai Medical Center (United States); Vita-Salute San Raffaele University (Italy); Humanitas University (Italy); National Institutes of Health (United States); National Institute of Neurological Disorders and Stroke (United States)
Journal: Brain communications, volume 8, issue 2, article fcag079
Dates: received 28 May 2025; accepted 9 March 2026; published online 11 March 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1093/braincomms/fcag079 · PMID 41884595 · PMCID PMC13010066 · OpenAlex W7135084735
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: structural MRI / diffusion (modality), multiple sclerosis (population), clinical / translational (subfield)
Methods: Connectivity, Statistics, Machine learning
Keywords: multiple sclerosis, machine learning, diagnosis, MRI, computer-aided diagnosis
Topic: Multiple Sclerosis Research Studies (Pathology and Forensic Medicine, Medicine), according to OpenAlex
Funding: National Multiple Sclerosis Society; Conrad N. Hilton Foundation; Fondazione Italiana Sclerosi Multipla; International Progressive MS Alliance; Waalse Gewest (2010235); École Polytechnique Fédérale de Lausanne; Fonds De La Recherche Scientifique - FNRS; Fondazione Cariplo; Fédération Wallonie-Bruxelles; Koning Boudewijnstichting; Hôpitaux Universitaires de Genève; Université de Genève; Université de Lausanne; Centre d'Imagerie BioMédicale; Centre Hospitalier Universitaire Vaudois; Fondazione Regionale per la Ricerca Biomedica; National Institutes of Health; Université Catholique de Louvain; Cliniques Universitaires Saint-Luc
Citations: cited by 1 paper (Europe PMC); 37 references in the paper

Abstract

Multiple sclerosis diagnostic criteria lack optimal specificity, leading to potential misdiagnosis. Advanced magnetic resonance imaging (MRI) biomarkers like the central vein sign, cortical lesions and paramagnetic rim lesions are highly specific to multiple sclerosis and could potentially improve diagnostic accuracy. In this study, we applied machine learning techniques to a retrospective, multicentric dataset of 322 multiple sclerosis/multiple sclerosis-mimic (204/118) and 84 prodromal multiple sclerosis/non-multiple sclerosis (43/41) adult patients, incorporating the central vein sign, cortical lesions and paramagnetic rim lesions. We compared (5 × 2 cross-validation combined F-test) the diagnostic performance of 71 machine learning models, each corresponding to a distinct combination of full-count or simplified biomarker inputs, against the baseline dissemination in space McDonald criteria. The aim was to evaluate the multiple sclerosis diagnostic power of combining these biomarkers in an MRI-only diagnostic framework. 51 of the 71 models significantly outperformed the dissemination in space criterion (P < 0.05), with balanced accuracy improvements up to 13.0% (confidence interval: [+10.5; +17.0]). The best overall model (random forest, using full-count assessments) achieved 95.7% (confidence interval: [93.2; 99.7]) balanced accuracy; the best simplified model (logistic regression, using only simplified assessments) reached 94.7% with no significant difference with the former (P = 0.29). Notably, 12/51 high-performing models used only simplified assessments. To further investigate the models’ generalizability, external validation on two out-of-distribution test sets using bootstrapping (1000 resamples) confirmed these results and highlighted a more robust generalization for the best model using solely simplified biomarkers. On the first external test set (n = 37, Verona), the simplified model achieved 97.2% balanced accuracy, while the full-count model reached 93.3% (versus 83.3% for baseline). On the second test set (n = 84, prodromal cases), the simplified model achieved 92.6% (versus 60.1% for baseline) showing competitive performance against the full-count model (93.9%). Both models improved all key performance metrics—balanced accuracy, sensitivity, specificity, precision and F1 score—over the baseline on both test sets (all P < 0.0001). Within a non-invasive MRI-only diagnostic framework, these results show that the incorporation of advanced imaging biomarkers into the multiple sclerosis-MRI diagnostic criteria significantly enhances the diagnostic accuracy—a statement holding true even when using simplified central vein sign, cortical lesions and paramagnetic rim lesions assessments. The study also provides a publicly available online diagnostic tool, facilitating further interaction, validation and clinical support (https://www.msdiagnostictool.org).

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 7 matches between paragraphs and lines of code.

maxencewynen/MS-Diagnostic-Tool-UCLouvain

License: none: the authors keep all their rights
State: the link answers, verified on 30 September 2026
Evidence: files inventoried
Commit: 9551af5d43f011334c8dd8734b34c3424022cf73, 25 September 2024
Languages: Python (9)
Size: 9 files, 9 scripts
Software Heritage: not archived
Found in: “Data availability”
Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: NumPy (6 files), scikit-learn (6 files), XGBoost (6 files), pandas (5 files), SciPy (2 files), Matplotlib (1 file), seaborn (1 file), SHAP (1 file)
Availability: 1 check, the latest on 30 September 2026: the link answers
  • 30 September 2026: the link answers
9 files
At the source: github.com/maxencewynen/MS-Diagnostic-Tool-UCLouvain

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;
  • 9 scripts, each with its path and the digest of its content;
  • 7 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

Qualified researchers can access the individual patient data of this study upon reasonable request and material transfer agreement between institutes. The trained models of the study are available at https://www.msdiagnostictool.org/. Source code used in the paper is available at https://github.com/maxencewynen/MS-Diagnostic-Tool-UCLouvain.

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, 30 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 17 authors, 5 keywords, 19 funders, 25 references.

Cite

This paper

Wynen, M., Vanden Bulcke, C., Borrelli, S., Gordaliza, P. M., Stölting, A., Guisset, F., Cordier, C., Martire, M. S., Tamanti, A., Macq, B., Sati, P., Filippi, M., Calabrese, M., Absinta, M., Reich, D. S., Bach Cuadra, M., & Maggi, P. (2026). Machine learning-based combination of the central vein sign, cortical lesions and paramagnetic rim lesions: a web-based tool for the diagnosis of multiple sclerosis. Brain communications, 8(2), fcag079. https://doi.org/10.1093/braincomms/fcag079

BibTeX

@article{wynen2026machine,
author = {Wynen, Maxence and Vanden Bulcke, Colin and Borrelli, Serena and Gordaliza, Pedro M and Stölting, Anna and Guisset, François and Cordier, Clément and Martire, Maria Sofia and Tamanti, Agnese and Macq, Benoit and Sati, Pascal and Filippi, Massimo and Calabrese, Massimiliano and Absinta, Martina and Reich, Daniel S and Bach Cuadra, Meritxell and Maggi, Pietro},
title = {{Machine learning-based combination of the central vein sign, cortical lesions and paramagnetic rim lesions: a web-based tool for the diagnosis of multiple sclerosis}},
journal = {Brain communications},
year = {2026},
month = mar,
volume = {8},
number = {2},
pages = {fcag079},
publisher = {Oxford University Press},
issn = {2632-1297},
doi = {10.1093/braincomms/fcag079},
url = {https://doi.org/10.1093/braincomms/fcag079},
pmid = {41884595},
pmcid = {PMC13010066}
}

RIS

TY - JOUR
AU - Wynen, Maxence
AU - Vanden Bulcke, Colin
AU - Borrelli, Serena
AU - Gordaliza, Pedro M
AU - Stölting, Anna
AU - Guisset, François
AU - Cordier, Clément
AU - Martire, Maria Sofia
AU - Tamanti, Agnese
AU - Macq, Benoit
AU - Sati, Pascal
AU - Filippi, Massimo
AU - Calabrese, Massimiliano
AU - Absinta, Martina
AU - Reich, Daniel S
AU - Bach Cuadra, Meritxell
AU - Maggi, Pietro
TI - Machine learning-based combination of the central vein sign, cortical lesions and paramagnetic rim lesions: a web-based tool for the diagnosis of multiple sclerosis
T2 - Brain communications
J2 - Brain Commun
PY - 2026
DA - 2026/03/11
VL - 8
IS - 2
SP - fcag079
SN - 2632-1297
PB - Oxford University Press
DO - 10.1093/braincomms/fcag079
UR - https://doi.org/10.1093/braincomms/fcag079
LA - en
ER -

CSL-JSON

{
"id": "10.1093/braincomms/fcag079",
"type": "article-journal",
"title": "Machine learning-based combination of the central vein sign, cortical lesions and paramagnetic rim lesions: a web-based tool for the diagnosis of multiple sclerosis",
"container-title": "Brain communications",
"author": [
{
"family": "Wynen",
"given": "Maxence"
},
{
"family": "Vanden Bulcke",
"given": "Colin"
},
{
"family": "Borrelli",
"given": "Serena"
},
{
"family": "Gordaliza",
"given": "Pedro M"
},
{
"family": "Stölting",
"given": "Anna"
},
{
"family": "Guisset",
"given": "François"
},
{
"family": "Cordier",
"given": "Clément"
},
{
"family": "Martire",
"given": "Maria Sofia"
},
{
"family": "Tamanti",
"given": "Agnese"
},
{
"family": "Macq",
"given": "Benoit"
},
{
"family": "Sati",
"given": "Pascal"
},
{
"family": "Filippi",
"given": "Massimo"
},
{
"family": "Calabrese",
"given": "Massimiliano"
},
{
"family": "Absinta",
"given": "Martina"
},
{
"family": "Reich",
"given": "Daniel S"
},
{
"family": "Bach Cuadra",
"given": "Meritxell"
},
{
"family": "Maggi",
"given": "Pietro"
}
],
"container-title-short": "Brain Commun",
"volume": "8",
"issue": "2",
"page": "fcag079",
"DOI": "10.1093/braincomms/fcag079",
"PMID": "41884595",
"PMCID": "PMC13010066",
"ISSN": "2632-1297",
"publisher": "Oxford University Press",
"URL": "https://doi.org/10.1093/braincomms/fcag079",
"language": "en",
"issued": {
"date-parts": [
[
2026,
3,
11
]
]
}
}

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:10.1016/j.nicl.2026.104007 [code]
A comparative study of deep learning for cortical lesion MRI segmentation with explainability analysis in multiple sclerosis.
Journal: NeuroImage. Clinical
In common: multiple sclerosis, structural MRI / diffusion, 4 references, 2 authors
[2] doi:10.64898/2026.07.15.26357954 [code]
Portable Ultra-Low Field MRI Deep-Learning Algorithms for White Matter Lesion Segmentation Improve Accuracy and Reflect Clinical Disability in Multiple Sclerosis
Journal: medRxiv (preprint)
In common: seaborn, scikit-learn, pandas, 3 other tools, multiple sclerosis, structural MRI / diffusion, clinical / translational, 1 reference, author Daniel S. Reich
[3] doi:10.1136/jnnp-2025-335884 [code]
Diffusivity anisotropy signature of slowly expanding lesions predicts progression independent of relapse activity in multiple sclerosis.
Journal: Journal of neurology, neurosurgery, and psychiatry
In common: seaborn, scikit-learn, pandas, 3 other tools, multiple sclerosis, structural MRI / diffusion, clinical / translational, 3 references
[4] doi:10.1186/s12880-026-02481-2 [code]
Deep learning-based neuroanatomical profiling reveals population-specific brain changes in multiple sclerosis: a large-scale Middle Eastern study.
Journal: BMC medical imaging
In common: seaborn, scikit-learn, pandas, 3 other tools, multiple sclerosis, structural MRI / diffusion, clinical / translational, 3 references
[5] doi:10.7554/elife.109717 [code]
Retrosplenial cortex enables context-dependent goal-directed sensorimotor transformation.
Journal: eLife
In common: SHAP, XGBoost, seaborn, 5 other tools, 1 reference
[6] doi:10.1073/pnas.2516601123 [code]
Unveiling the glymphatic system's role in brain aging: A comprehensive biomarker and modifiable intervention target.
Journal: Proceedings of the National Academy of Sciences of the United States of America
In common: SHAP, XGBoost, seaborn, 5 other tools, structural MRI / diffusion, clinical / translational
[7] doi:10.1038/s41467-026-71555-0 [code]
A deep representation learning model to predict response to vagus nerve stimulation.
Journal: Nature communications
In common: SHAP, XGBoost, seaborn, 5 other tools, structural MRI / diffusion, clinical / translational
[8] doi:10.21203/rs.3.rs-9914920/v1 [code]
Prediction of cognitive performance by demographics, sleep, and brain morphometry: machine learning findings from ENIGMA-Sleep Working Group
Journal: Research Square (preprint)
In common: SHAP, XGBoost, seaborn, 5 other tools, structural MRI / diffusion
[9] doi:10.1016/j.nicl.2026.104012 [code]
Structural-functional multilayer brain network properties and outcome of combined repetitive transcranial magnetic stimulation and psychotherapy for obsessive-compulsive disorder.
Journal: NeuroImage. Clinical
In common: SHAP, XGBoost, seaborn, 5 other tools, structural MRI / diffusion
[10] doi:10.1016/j.isci.2026.115329 [code]
Brain metastases converge on shared geometric architecture and transcriptomic landscape yet remain distinct from gliomas.
Journal: iScience
In common: SHAP, XGBoost, seaborn, 5 other tools, clinical / translational

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.