OSCR

Data-driven subtyping and staging of ALS: A multicenter, longitudinal, deformation-based morphometry study.

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] § Methods › Image preprocessing ↔ PELICAN.sh, lines 117–164 · score 0.77 · volume_pol, Pre processing, PELICAN, ANTs, denoising, template
  2. [2] § Methods › Image preprocessing ↔ app_cicLngPipeline_ants_asym.sh, lines 61–115 · score 0.77 · volume_pol, Pre processing, ANTs, denoising, pipeline, template
  3. [3] § Methods › Subtype and stage inference modeling ↔ sim/simrun.py, lines 303–354 · score 0.75 · sample log likelihood, Cross Validation Information, Model selection, CVIC, subtypes
  4. [4] § Results › SuStaIn input features and optimal number of subtypes ↔ sim/simrun.py, lines 303–354 · score 0.72 · positional variance diagrams, cross validation information, log likelihood, CVIC, SuStaIn model, PVDs
  5. [5] § Results › SuStaIn input features and optimal number of subtypes ↔ pySuStaIn/AbstractSustain.py, lines 402–523 · score 0.58 · positional variance diagrams, cross validation, SuStaIn model, PVDs, likelihood, sequence
  6. [6] § Results ↔ sim/simrun.py, lines 246–301 · score 0.52 · positional variance diagrams, SuStaIn model, PVDs, inferred, sequence, subtype

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 · 358 lines · 17 KB · MIT · 3 matches

  1. ###
  2. # pySuStaIn: a Python implementation of the Subtype and Stage Inference (SuStaIn) algorithm
  3. #
  4. # If you use pySuStaIn, please cite the following core papers:
  5. # 1. The original SuStaIn paper: https://doi.org/10.1038/s41467-018-05892-0
  6. # 2. The pySuStaIn software paper: https://doi.org/10.1016/j.softx.2021.100811
  7. #
  8. # Please also cite the corresponding progression pattern model you use:
  9. # 1. The piece-wise linear z-score model (i.e. ZscoreSustain): https://doi.org/10.1038/s41467-018-05892-0
  10. # 2. The event-based model (i.e. MixtureSustain): https://doi.org/10.1016/j.neuroimage.2012.01.062
  11. # with Gaussian mixture modeling (i.e. 'mixture_gmm'): https://doi.org/10.1093/brain/awu176
  12. # or kernel density estimation (i.e. 'mixture_kde'): https://doi.org/10.1002/alz.12083
  13. # 3. The model for discrete ordinal data (i.e. OrdinalSustain): https://doi.org/10.3389/frai.2021.613261
  14. #
  15. # Thanks a lot for supporting this project.
  16. #
  17. # Authors: Peter Wijeratne ([email hidden]) and Leon Aksman ([email hidden])
  18. # Contributors: Arman Eshaghi ([email hidden]), Alex Young ([email hidden]), Cameron Shand ([email hidden])
  19. ###
  20. import numpy as np
  21. from matplotlib import pyplot as plt
  22. from matplotlib import cbook as cbook
  24. import os
  26. import pandas as pd
  28. from simfuncs import *
  30. from kde_ebm.mixture_model import fit_all_gmm_models, fit_all_kde_models
  31. from kde_ebm import plotting
  33. import warnings
  34. warnings.filterwarnings("ignore",category=cbook.mplDeprecation)
  36. from pySuStaIn.ZscoreSustain import ZscoreSustain
  37. from pySuStaIn.MixtureSustain import MixtureSustain
  38. from pySuStaIn.MixedTypeSustain import MixedTypeSustain
  39. import simfuncs_mixed
  41. import sklearn.model_selection
  43. import pylab
  45. def main():
  47. #***************** parameters for generating the ground-truth subtypes
  48. N = 5 # number of biomarkers
  49. M = 800 # number of observations ( e.g. subjects )
  50. N_S_ground_truth = 3 # number of ground truth subtypes
  52. # the fractions of the total number of subjects (M) belonging to each subtype
  53. ground_truth_fractions = np.array([0.5, 0.30, 0.20])
  55. #create some generic biomarker names
  56. BiomarkerNames = ['Biomarker ' + str(i) for i in range(N)]
  58. #***************** parameters for SuStaIn-based inference of subtypes
  59. use_parallel_startpoints = True
  61. # number of starting points
  62. N_startpoints = 25
  63. # maximum number of inferred subtypes - note that this could differ from N_S_ground_truth
  64. N_S_max = 2
  65. N_iterations_MCMC = int(1e4) #Generally recommend either 1e5 or 1e6 (the latter may be slow though) in practice
  67. #labels for plotting are biomarker names
  68. SuStaInLabels = BiomarkerNames
  70. # cross-validation
  71. validate = True
  72. N_folds = 3 #Set low to speed things up here, but generally recommend 10 in practice
  74. # either 'mixture_GMM' or 'mixture_KDE' or 'zscore' or 'mixed'
  75. sustainType = 'mixed'
  77. assert sustainType in ("mixture_GMM", "mixture_KDE", "zscore", "mixed"), \
  78. "sustainType should be either mixture_GMM, mixture_KDE, zscore, or mixed"
  80. #****************** generate the ground-truth sequences and groud-truth data (i.e. subjects' biomarker measures)
  81. dataset_name = 'sim'
  82. output_folder = os.path.join(os.getcwd(), dataset_name + '_' + sustainType)
  83. if not os.path.isdir(output_folder):
  84. os.mkdir(output_folder)
  86. if sustainType == 'mixture_GMM' or sustainType == "mixture_KDE":
  88. np.random.seed(5)
  90. ground_truth_subj_ids = list(np.arange(1, M+1).astype('str'))
  91. ground_truth_sequences = generate_random_mixture_sustain_model(N, N_S_ground_truth)
  93. ground_truth_subtypes = np.random.choice(range(N_S_ground_truth), M, replace=True, p=ground_truth_fractions).astype(int)
  95. N_stages = N
  97. ground_truth_stages_control = np.zeros((int(np.round(M * 0.25)), 1))
  98. ground_truth_stages_other = np.random.randint(1, N_stages+1, (int(np.round(M * 0.75)), 1))
  99. ground_truth_stages = np.vstack((ground_truth_stages_control, ground_truth_stages_other)).astype(int)
  101. data, data_denoised = generate_data_mixture_sustain(ground_truth_subtypes, ground_truth_stages, ground_truth_sequences, sustainType)
  103. # choose which subjects will be cases and which will be controls
  104. MIN_CASE_STAGE = np.round((N + 1) * 0.8)
  105. index_case = np.where(ground_truth_stages >= MIN_CASE_STAGE)[0]
  106. index_control = np.where(ground_truth_stages == 0)[0]
  108. labels = 2 * np.ones(data.shape[0], dtype=int) # 2 - intermediate value, not used in mixture model fitting
  109. labels[index_case] = 1 # 1 - cases
  110. labels[index_control] = 0 # 0 - controls
  112. data_case_control = data[labels != 2, :]
  113. labels_case_control = labels[labels != 2]
  115. if sustainType == "mixture_GMM":
  116. mixtures = fit_all_gmm_models(data, labels)
  117. elif sustainType == "mixture_KDE":
  118. mixtures = fit_all_kde_models(data, labels)
  120. fig, ax = plotting.mixture_model_grid(data_case_control, labels_case_control, mixtures, SuStaInLabels)#, plotting_font_size=20)
  121. fig.show()
  122. fig.savefig(os.path.join(output_folder, 'kde_fits.png'))
  124. L_yes = np.zeros(data.shape)
  125. L_no = np.zeros(data.shape)
  126. for i in range(N):
  127. if sustainType == "mixture_GMM":
  128. L_no[:, i], L_yes[:, i] = mixtures[i].pdf(None, data[:, i])
  129. elif sustainType == "mixture_KDE":
  130. L_no[:, i], L_yes[:, i] = mixtures[i].pdf(data[:, i].reshape(-1, 1))
  132. sustain = MixtureSustain(L_yes, L_no, SuStaInLabels, N_startpoints, N_S_max, N_iterations_MCMC, output_folder, dataset_name, use_parallel_startpoints)
  134. elif sustainType == 'zscore':
  136. np.random.seed(10)
  138. Z_vals = np.array([[1, 2, 3]] * N) # define the Z-score based events for each biomarker
  139. Z_max = np.array([5] * N) # maximum z-score for each biomarker
  141. ground_truth_subj_ids = list(np.arange(1, M+1).astype('str'))
  143. # generate the ground truth sequence for each subtype
  144. ground_truth_sequences = generate_random_Zscore_sustain_model(Z_vals, N_S_ground_truth)
  146. # randomly generate the ground truth subtype and stage assignment for every one of the M subjects
  147. ground_truth_subtypes = np.random.choice(range(N_S_ground_truth), M, replace=True, p=ground_truth_fractions).astype(int)
  149. N_stages = np.sum(Z_vals > 0) + 1
  151. #1/4 of all subjects are assigned stage zero, the rest a random number from 1 to number of stages
  152. ground_truth_stages_control = np.zeros((int(np.round(M * 0.25)), 1))
  153. ground_truth_stages_other = np.random.randint(1, N_stages+1, (int(np.round(M * 0.75)), 1))
  154. ground_truth_stages = np.vstack((ground_truth_stages_control, ground_truth_stages_other)).astype(int)
  156. data, data_denoised, stage_value = generate_data_Zscore_sustain( ground_truth_subtypes,
  157. ground_truth_stages,
  158. ground_truth_sequences,
  159. Z_vals,
  160. Z_max)
  162. # choose which subjects will be cases and which will be controls
  163. MIN_CASE_STAGE = int(np.round((N_stages + 1) * 0.8))
  164. index_case = np.where(ground_truth_stages >= MIN_CASE_STAGE)[0]
  165. index_control = np.where(ground_truth_stages == 0)[0]
  168. labels = 2 * np.ones(data.shape[0], dtype=int) # 2 - MCI, default assignment here
  169. labels[index_case] = 0
  170. labels[index_control] = 1
  172. sustain = ZscoreSustain(data, Z_vals, Z_max, SuStaInLabels, N_startpoints, N_S_max, N_iterations_MCMC, output_folder, dataset_name, use_parallel_startpoints)
  174. elif sustainType == 'mixed':
  176. data = simfuncs_mixed.generate_mixed_data(seed=42)
  178. zscore_data = data["zscore_data"]
  179. z_vals = data["z_vals"]
  180. z_max = data["z_max"]
  181. score_vals = data["score_vals"]
  183. n_biomarkers_ordinal = data["n_biomarkers_ordinal"]
  184. n_biomarkers_event = data["n_biomarkers_event"]
  186. ordinal_prob_nl = data["ordinal_prob_nl"]
  187. ordinal_prob_score = data["ordinal_prob_score"]
  188. event_prob_yes = data["event_prob_yes"]
  189. event_prob_no = data["event_prob_no"]
  191. if zscore_data is not None and zscore_data.shape[1] > 0:
  192. zscore_labels = [f"zscore_{i+1}" for i in range(zscore_data.shape[1])]
  193. else:
  194. zscore_labels = []
  196. ordinal_labels = [f"ordinal_{i+1}" for i in range(n_biomarkers_ordinal)]
  197. event_labels = [f"event_{i+1}" for i in range(n_biomarkers_event)]
  199. if n_biomarkers_ordinal > 0:
  200. ordinal_score_vals = score_vals[:n_biomarkers_ordinal, :]
  201. else:
  202. ordinal_prob_nl = None
  203. ordinal_prob_score = None
  204. ordinal_score_vals = np.zeros((0, 0), dtype=int)
  206. if n_biomarkers_event == 0:
  207. event_prob_no = None
  208. event_prob_yes = None
  210. if zscore_data is not None:
  211. num_subjects = zscore_data.shape[0]
  212. elif ordinal_prob_nl is not None:
  213. num_subjects = ordinal_prob_nl.shape[0]
  214. elif event_prob_no is not None:
  215. num_subjects = event_prob_no.shape[0]
  216. else:
  217. raise ValueError("Mixed simulation returned no modality data.")
  219. ground_truth_subj_ids = list(np.arange(1, num_subjects + 1).astype('str'))
  220. ground_truth_sequences = data["gt_sequence"]
  221. ground_truth_subtypes = data["gt_subtypes"]
  222. ground_truth_stages = data["gt_stages"]
  223. N_S_ground_truth = ground_truth_sequences.shape[0]
  225. sustain = MixedTypeSustain(
  226. zscore_data=zscore_data,
  227. z_vals=z_vals,
  228. z_max=z_max,
  229. zscore_biomarker_labels=zscore_labels,
  230. ordinal_prob_nl=ordinal_prob_nl,
  231. ordinal_prob_score=ordinal_prob_score,
  232. ordinal_score_vals=ordinal_score_vals,
  233. ordinal_biomarker_labels=ordinal_labels,
  234. event_prob_yes=event_prob_yes,
  235. event_prob_no=event_prob_no,
  236. event_biomarker_labels=event_labels,
  237. N_startpoints=N_startpoints,
  238. N_S_max=N_S_max,
  239. N_iterations_MCMC=N_iterations_MCMC,
  240. output_folder=output_folder,
  241. dataset_name=dataset_name,
  242. use_parallel_startpoints=use_parallel_startpoints
  243. )
  244. validate = False
  246. #****** plot the ground truth sequences
  247. ground_truth_sequences = np.expand_dims(ground_truth_sequences, axis=2)
  248. ground_truth_fractions_actual, _ = np.histogram(ground_truth_subtypes, bins=np.arange(N_S_ground_truth + 1) - 0.5)
  249. ground_truth_fractions_actual = ground_truth_fractions_actual/len(ground_truth_subtypes)
  250. ground_truth_fractions_actual = np.expand_dims(ground_truth_fractions_actual, axis=1)
  251. ground_truth_nsamples = np.inf
  253. #ordering of positional variance diagrams (PVDs)
  254. plot_subtype_order = np.arange(N_S_ground_truth)
  255. #ordering of biomarkers in each PVD
  256. plot_biomarker_order = ground_truth_sequences[plot_subtype_order[0], :].astype(int).ravel()
  257. #plot PVDs given subtype and biomarker ordering
  258. figs, ax = sustain._plot_sustain_model(ground_truth_sequences, ground_truth_fractions_actual, ground_truth_nsamples, \
  259. subtype_order=plot_subtype_order, biomarker_order=plot_biomarker_order, title_font_size=12)
  260. figs[0].suptitle('Ground truth sequences')
  261. figs[0].savefig(os.path.join(output_folder, 'PVD_true.png'))
  262. figs[0].show()
  264. #************* run SuStaIn to infer subtype sequences and subjects' subtypes/stages estimates
  265. samples_sequence, \
  266. samples_f, \
  267. ml_subtype, \
  268. prob_ml_subtype, \
  269. ml_stage, \
  270. prob_ml_stage, \
  271. prob_subtype_stage = sustain.run_sustain_algorithm(plot=True)
  273. #save the most likely subtype, the associated subtype probability,
  274. # the most likely stage and the associated stage probability for each subject
  275. df = pd.DataFrame()
  276. df['subj_id'] = ground_truth_subj_ids
  277. df['ml_subtype'] = ml_subtype
  278. df['prob_ml_subtype'] = prob_ml_subtype
  279. df['ml_stage'] = ml_stage
  280. df['prob_ml_stage'] = prob_ml_stage
  281. df.to_csv(os.path.join(output_folder, 'Subject_subtype_stage_estimates.csv'), index=False)
  283. FONT_SIZE = 15
  285. #plot the inferred subtypes as histograms binned by true subtype
  286. plt.style.use('seaborn-deep')
  287. bins = np.arange(0, N_S_ground_truth+1)
  288. X_hist = list()
  289. labels_hist = list()
  290. for i in range(N_S_max):
  291. X_hist.append(ml_subtype[ground_truth_subtypes==i].ravel())
  292. labels_hist.append('Subtype ' + str(i+1))
  293. fig, ax = plt.subplots()
  294. ax.hist(X_hist, bins, label=labels_hist)
  295. ax.set_xticks(np.arange(0, N_S_ground_truth)+0.5)
  296. ax.set_xticklabels(np.arange(1, N_S_ground_truth+1))
  297. ax.set_xlabel('Estimated subtype', fontsize=FONT_SIZE)
  298. ax.set_title('')
  299. ax.legend(loc='upper right', fontsize=FONT_SIZE)
  300. fig.savefig(os.path.join(output_folder, 'Subtype_estimate_histograms.png'))
  301. fig.show()
  303. #plot the inferred stages as boxplots binned by true stage
  304. df_boxplot = pd.DataFrame()
  305. df_boxplot['subtypes_true'] = ground_truth_subtypes
  306. df_boxplot['subtypes_est'] = ml_subtype
  307. df_boxplot['stages_true'] = ground_truth_stages
  308. df_boxplot['stages_est'] = ml_stage
  309. fig, ax = plt.subplots()
  310. df_boxplot.boxplot(column='stages_est', by='stages_true', grid=False, fontsize=FONT_SIZE, ax=ax)
  311. ax.set_xlabel('True stages', fontsize=FONT_SIZE)
  312. ax.set_ylabel('Estimated stages', fontsize=FONT_SIZE)
  313. ax.set_title('')
  314. fig.savefig(os.path.join(output_folder, 'Stage_estimate_boxplots.png'))
  315. fig.show()
  317. print('Maximum likelihood model finished. Saved figures and output files in ' + output_folder + ' folder.')
  319. # ************* cross-validation: splits the whole dataset into a number of folds (N_folds),
  320. # runs SuStaIn on the training data of each fold and evaluates out-of-sample log likelihood of fold's test data,
  321. # also evaluates cross-validation information criterion (CVIC) and displays cross-validated positional variance diagram (PVD)
  322. if validate:
  324. print('Running cross-validation. This may take an hour or longer. Set validate = False to disable.')
  326. test_idxs = []
  328. cv = sklearn.model_selection.StratifiedKFold(n_splits=N_folds, shuffle=True)
  329. cv_it = cv.split(data, labels)
  331. for train, test in cv_it:
  332. test_idxs.append(test)
  333. test_idxs = np.array(test_idxs)
  335. #For parallelization, you can call this several different ways
  336. #passing in one or a specific set ofcross-validation folds:
  337. #CVIC, loglike_matrix = sustain.cross_validate_sustain_model(test_idxs, 0) #just the first fold
  338. #CVIC, loglike_matrix = sustain.cross_validate_sustain_model(test_idxs, [0,5]) #first and sixth
  340. #You can also just run all folds at once
  341. CVIC, loglike_matrix = sustain.cross_validate_sustain_model(test_idxs)
  343. if CVIC == [] and loglike_matrix == []:
  344. return
  346. #output CV folds' out-of-sample log likelihoods
  347. df_loglike = pd.DataFrame(data = loglike_matrix, columns = ["Subtype " + str(i+1) for i in range(N_S_max)])
  348. df_loglike.to_csv(os.path.join(output_folder, 'Log_likelihoods_cv_folds.csv'), index=False)
  350. #this part estimates cross-validated positional variance diagrams
  351. for i in range(N_S_max):
  352. sustain.combine_cross_validated_sequences(i+1, N_folds)
  354. print('Cross-validation finished. Saved figures and output files in ' + output_folder + ' folder.')
  356. if __name__ == '__main__':
  357. np.random.seed(42)
  358. main()

simrun.py at commit 708fa22, under MIT · at the source

Overview

Authors: Isabelle Lajoie1,2, Sanjay Kalra3,4, Mahsa Dadar1,2
ORCID iDs: Isabelle Lajoie
  1. Douglas Mental Health University Health Centre, Montreal, Quebec, Canada
  2. Department of Psychiatry, McGill University, Montreal, Quebec, Canada
  3. Neuroscience and Mental Health Institute, University of Alberta, Edmonton, Alberta, Canada
  4. Division of Neurology, Department of Medicine, University of Alberta, Edmonton, Alberta, Canada
Institutions: Douglas Mental Health University Institute (Canada); McGill University (Canada); University of Alberta (Canada); Neuroscience and Mental Health Institute (Canada)
Journal: Imaging neuroscience (Cambridge, Mass.), volume 4, article IMAG.a.1264
Dates: received 11 December 2025; accepted 12 May 2026; published online 4 June 2026
Type: Research article · Language: English
License: CC BY
Identifiers: DOI 10.1162/imag.a.1264 · PMID 42253609 · PMCID PMC13238002 · OpenAlex W4417129219
Open access: diamond, a free copy (OpenAlex)
Preprint: doi:10.64898/2025.12.02.25341482
Status: code verified
Categories: structural MRI / diffusion (modality), human (organism), other condition (population), clinical / translational (subfield)
Methods: Statistics, Machine learning, fMRI & imaging
Keywords: subtype and stage inference, progression modeling, machine learning, longitudinal
Topic: Amyotrophic Lateral Sclerosis Research (Neurology, Medicine), according to OpenAlex
Funding: ALS Society of Canada; Brain Canada; Canadian Institutes of Health Research (CIHR); Shelly Mrkonjic ALS Research Fund
Citations: not cited yet (Europe PMC); 50 references in the paper

Abstract

Amyotrophic lateral sclerosis (ALS) is clinically and biologically heterogeneous, yet data-driven imaging subtyping approaches have rarely been validated longitudinally or linked to clinical and survival outcomes. We aimed to identify and validate distinct ALS subtypes and disease stages using deformation-based morphometry (DBM) and the Subtype and Stage Inference (SuStaIn) model, and to characterize their cross-sectional and longitudinal imaging, clinical, cognitive, and survival profiles. Data from 198 ALS patients and 144 healthy controls in the Canadian ALS Neuroimaging Consortium (CALSNIC) multicenter cohort were analyzed. Baseline regional DBM w-scores from 14 ALS-relevant regions served as input to SuStaIn to infer subtypes and stages. Longitudinal consistency of subtype and stage assignments (e.g. adherence to the expected disease evolution) was assessed using follow-up visits. Imaging and clinical trajectories were compared across subtypes using linear mixed-effects models incorporating stage and elapsed time. Associations between longitudinal variables and SuStaIn stage were estimated using mixed models, while baseline clinical and cognitive differences were assessed with ordinary least squares regression. Survival differences were evaluated using Kaplan–Meier curves and log-rank tests. SuStaIn identified one normal-appearing group (S0) and three ALS atrophy subtypes. S0 showed no baseline atrophy but exhibited longitudinal motor decline and the most favorable survival (log-rank p < 0.05 to p < 0.01). S1 exhibited classical motor/corticospinal tract-dominant degeneration, greater lower motor neuron burden, and intermediate survival. S2 showed limbic-onset atrophy progressing toward motor pathways, with preserved cognition and a milder course. S3 demonstrated extensive fronto-parietal and striatal atrophy, longitudinal motor–thalamic degeneration, and the shortest survival. Subtype and stage assignments demonstrated high longitudinal consistency (>90%). SuStaIn stage was strongly associated with widespread brain atrophy (and ventricular expansion), with the strongest effects in limbic–subcortical regions. Stage also correlated with ALS Functional Rating Scale-Revised (ALSFRS-R) decline and forced vital capacity (FVC) reduction, indicating that stage reflects disease-linked progression. This study establishes a robust, longitudinally validated model of ALS heterogeneity, showing that SuStaIn-derived subtypes define distinct disease trajectories, whereas the normal-appearing group reflects an early, structurally preserved state with a more favorable survival profile. By integrating probabilistic staging with longitudinal modeling, these findings clarify dynamic subtype-specific progression patterns and support the use of SuStaIn for biologically informed patient stratification, prognostication, and clinical trial enrichment in ALS.

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

Repositories

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

BIC-MNI/minc-tools

License: other
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Commit: e3825986359ecd75d82aa88ff2015d36e234e55d, 13 September 2015
Languages: C (183), C/C++ (83), Perl (16), Shell (8), MATLAB (3)
Size: 443 files, 293 scripts
Software Heritage: not archived
Found in: the text, “Image preprocessing”
Holds: README, license file, documentation
Not found: CITATION.cff, environment file, tests, continuous integration
Tools: SPM (3 files)
Availability: 1 check, the latest on 27 September 2026: the link answers
  • 27 September 2026: the link answers
295 files
At the source: github.com/BIC-MNI/minc-tools

VANDAlab/Preprocessing_Pipeline

License: none: the authors keep all their rights
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Commit: 1e2f4abd90baa404fedf0d91f914fa3be3051820, 15 April 2026
Languages: Shell (10)
Size: 12 files, 10 scripts
Software Heritage: not archived
Found in: “Data and Code Availability”
Holds: README
Not found: license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: ANTs (10 files)
Availability: 1 check, the latest on 27 September 2026: the link answers
  • 27 September 2026: the link answers
11 files
At the source: github.com/VANDAlab/Preprocessing_Pipeline

ucl-pond/pySuStaIn

License: MIT
State: the link answers, verified on 27 September 2026
Evidence: files inventoried
Commit: 708fa22d89c9692dbeafcc2c31a1c8460ced5640, 30 March 2026
Languages: Python (14), Jupyter (3)
Size: 30 files, 17 scripts
Software Heritage: not archived
Found in: “Data and Code Availability”
Holds: README, license file, environment (requirements.txt, setup.py), tests, 3 notebooks
Not found: CITATION.cff, continuous integration, documentation
Tools: NumPy (14 files), Matplotlib (10 files), SciPy (8 files), pandas (7 files), scikit-learn (3 files), statsmodels (2 files), Pillow (1 file), seaborn (1 file)
Availability: 1 check, the latest on 27 September 2026: the link answers
  • 27 September 2026: the link answers
19 files
At the source: github.com/ucl-pond/pySuStaIn

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:

  • 3 repositories of the authors' code, each at its verified commit, with its license and how the link was found in the paper;
  • 320 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 and Code Availability

Imaging and clinical data can be requested from the CALSNIC consortium (https://calsnic.org/). The scripts for MRI data preprocessing and DBM calculations are available at https://github.com/VANDAlab/Preprocessing_Pipeline. The python SuStaIn code is available at https://github.com/ucl-pond/pySuStaIn. The following open access tools were employed: MINC (https://github.com/BIC-MNI/minc-tools) and ANTs (http://stnava.github.io/ANTs/).

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, 27 September 2026: the first record

Recorded: type, language, journal, volume, pages, dates, 3 authors, 4 keywords, 4 funders, 50 references.

Cite

This paper

Lajoie, I., Kalra, S., & Dadar, M. (2026). Data-driven subtyping and staging of ALS: A multicenter, longitudinal, deformation-based morphometry study. Imaging neuroscience (Cambridge, Mass.), 4, IMAG.a.1264. https://doi.org/10.1162/imag.a.1264

BibTeX

@article{lajoie2026data,
author = {Lajoie, Isabelle and Kalra, Sanjay and Dadar, Mahsa},
title = {{Data-driven subtyping and staging of ALS: A multicenter, longitudinal, deformation-based morphometry study}},
journal = {Imaging neuroscience (Cambridge, Mass.)},
year = {2026},
month = jun,
volume = {4},
pages = {IMAG.a.1264},
publisher = {MIT Press},
issn = {2837-6056},
doi = {10.1162/imag.a.1264},
url = {https://doi.org/10.1162/imag.a.1264},
pmid = {42253609},
pmcid = {PMC13238002}
}

RIS

TY - JOUR
AU - Lajoie, Isabelle
AU - Kalra, Sanjay
AU - Dadar, Mahsa
TI - Data-driven subtyping and staging of ALS: A multicenter, longitudinal, deformation-based morphometry study
T2 - Imaging neuroscience (Cambridge, Mass.)
J2 - Imaging Neurosci (Camb)
PY - 2026
DA - 2026/06/04
VL - 4
SP - IMAG.a.1264
SN - 2837-6056
PB - MIT Press
DO - 10.1162/imag.a.1264
UR - https://doi.org/10.1162/imag.a.1264
LA - en
ER -

CSL-JSON

{
"id": "10.1162/imag.a.1264",
"type": "article-journal",
"title": "Data-driven subtyping and staging of ALS: A multicenter, longitudinal, deformation-based morphometry study",
"container-title": "Imaging neuroscience (Cambridge, Mass.)",
"author": [
{
"family": "Lajoie",
"given": "Isabelle"
},
{
"family": "Kalra",
"given": "Sanjay"
},
{
"family": "Dadar",
"given": "Mahsa"
}
],
"container-title-short": "Imaging Neurosci (Camb)",
"volume": "4",
"page": "IMAG.a.1264",
"DOI": "10.1162/imag.a.1264",
"PMID": "42253609",
"PMCID": "PMC13238002",
"ISSN": "2837-6056",
"publisher": "MIT Press",
"URL": "https://doi.org/10.1162/imag.a.1264",
"language": "en",
"issued": {
"date-parts": [
[
2026,
6,
4
]
]
}
}

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:10.1162/imag.a.1235 [code]
Intracranial volume: To adjust or not to adjust? It is not a matter of if, but how.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: ANTs, pandas, SciPy, 1 other tool, structural MRI / diffusion, 9 references
[2] doi:10.1093/braincomms/fcag176 [code]
Tau topography subtypes account for clinical heterogeneity and longitudinal trajectories in early-onset Alzheimer's disease.
Journal: Brain communications
In common: SPM, Pillow, statsmodels, 6 other tools, clinical / translational, 3 references
[3] doi:10.1371/journal.pbio.3003856 [code]
Aging and metabolism contribute separately to brain-body health.
Journal: PLoS biology
In common: ANTs, SPM, statsmodels, 6 other tools, structural MRI / diffusion, clinical / translational, 1 reference
[4] doi:10.1038/s41467-026-71151-2 [code]
Common and distinct neural correlates of social interaction processing and theory of mind in narratives.
Journal: Nature communications
In common: ANTs, SPM, Pillow, 7 other tools, 1 reference
[5] doi:10.1162/imag.a.1276 [code]
High-resolution whole-brain magnetic resonance spectroscopic imaging in youth at risk for psychosis.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: ANTs, Pillow, seaborn, 5 other tools, clinical / translational, 2 references
[6] 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: ANTs, Pillow, statsmodels, 6 other tools, structural MRI / diffusion, other condition
[7] doi:10.1038/s41398-026-03965-z [code]
Disentangling individual heterogeneity reveals robust network and molecular signatures of major depressive disorder with suicidal ideation.
Journal: Translational psychiatry
In common: SPM, Pillow, statsmodels, 6 other tools, 1 reference
[8] doi:10.1093/braincomms/fcag127 [code]
Individualized phenotyping of functional amyotrophic lateral sclerosis pathology in sensorimotor cortex.
Journal: Brain communications
In common: ANTs, SPM, seaborn, 5 other tools, other condition, 1 reference
[9] doi:10.1038/s41467-026-71568-9 [code]
Convergent and selective representations of pain, appetitive processes, aversive processes, and cognitive control in the insula.
Journal: Nature communications
In common: ANTs, SPM, Pillow, 6 other tools
[10] 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: ANTs, seaborn, scikit-learn, 4 other tools, structural MRI / diffusion, clinical / translational, 2 references

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.