OSCR

Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET.

Code ↔ Paper

3 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 3 matches
  1. [1] § Methods › Experiment › Training Details ↔ Code/utils/dataset.py, lines 17–148 · score 0.79 · Gaussian noise, elastic, magnitude, trilinear, augmentation, flips
  2. [2] § Methods › Experiment › Training Details ↔ Code/train.py, lines 100–156 · score 0.69 · cosine annealing learning, Adam, dropout, channel, cycle, scheduler
  3. [3] § Methods › Experiment › Benchmark ↔ Code/train.py, lines 100–156 · score 0.66 · cosine annealing, combined loss function, min, max, cycle, scheduler

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 · 294 lines · 9.6 KB · no license · 2 matches

  1. # Run this code like python train.py --config example.yaml
  2. # If you have a checkpoint, run this code like python train.py --config example.yaml --checkpoint /path/to/the/chechpoint.pth
  4. import os
  5. import time
  6. import torch
  7. from torch import optim
  8. from torch.optim.lr_scheduler import CosineAnnealingLR
  9. from monai.networks.nets import UNet
  10. from utils.dataset import load_data
  11. from utils.cyclic_perceptual_loss import (
  12. CombinedLoss,
  13. update_plane_selection,
  14. calculate_cycle_epochs,
  15. )
  16. from utils.utils import (
  17. update_checkpoint_and_log,
  18. load_checkpoint,
  19. parse_arguments,
  20. load_config,
  21. seed_everything,
  22. )
  25. def train_one_epoch(
  26. model, train_loader, optimizer, loss_function, device, selection, grad_accum_steps
  27. ):
  28. """
  29. Train the model for one epoch on the given DataLoader and return the average loss.
  31. Args:
  32. model (nn.Module): The neural network model.
  33. train_loader (DataLoader): The DataLoader for the training set.
  34. optimizer (Optimizer): The optimizer.
  35. loss_function (callable): The loss function.
  36. device (torch.device): The device to use (CPU or GPU).
  37. selection (str): Which plane is currently selected ("axial", "coronal", or "sagittal").
  38. grad_accum_steps (int): Number of gradient accumulation steps.
  40. Returns:
  41. float: The average training loss for this epoch.
  42. """
  43. model.train()
  44. epoch_loss = 0.0
  46. for idx, batch_data in enumerate(train_loader):
  47. inputs = batch_data["image"].to(device)
  48. targets = batch_data["label"].to(device)
  50. outputs = model(inputs)
  51. loss = loss_function(outputs, targets, selection)
  53. # Gradient Accumulation
  54. (loss / grad_accum_steps).backward()
  55. if (idx + 1) % grad_accum_steps == 0 or idx == len(train_loader) - 1:
  56. optimizer.step()
  57. optimizer.zero_grad()
  59. epoch_loss += loss.item()
  61. # Clear memory for unused tensors
  62. del inputs, targets, outputs
  64. average_loss = epoch_loss / len(train_loader)
  65. return average_loss
  68. def validate_one_epoch(model, val_loader, loss_function, device, selection):
  69. """
  70. Run validation on the given DataLoader for one epoch and return the average validation loss.
  72. Args:
  73. model (nn.Module): The neural network model.
  74. val_loader (DataLoader): The DataLoader for the validation set.
  75. loss_function (callable): The loss function.
  76. device (torch.device): The device to use (CPU or GPU).
  77. selection (str): Which plane is currently selected ("axial", "coronal", or "sagittal").
  79. Returns:
  80. float: The average validation loss for this epoch.
  81. """
  82. model.eval()
  83. val_loss = 0.0
  85. with torch.no_grad():
  86. for val_data in val_loader:
  87. val_inputs = val_data["image"].to(device)
  88. val_targets = val_data["label"].to(device)
  90. val_outputs = model(val_inputs)
  91. loss = loss_function(val_outputs, val_targets, selection)
  92. val_loss += loss.item()
  94. del val_inputs, val_targets, val_outputs
  96. average_loss = val_loss / len(val_loader)
  97. return average_loss
  100. def train_model(train_loader, val_loader, config, device, checkpoint_path=None):
  101. """
  102. Train and validate the model on the provided data loaders. Manages checkpoints and logging.
  104. Args:
  105. train_loader (DataLoader): The DataLoader for the training set.
  106. val_loader (DataLoader): The DataLoader for the validation set.
  107. config (dict): Configuration dictionary loaded from YAML.
  108. device (torch.device): The device to use for training.
  109. checkpoint_path (str, optional): Path to an existing checkpoint to resume from.
  110. """
  112. # Prepare directories
  113. log_dir = config["LOG_DIR"]
  114. model_dir = config["MODEL_DIR"]
  115. os.makedirs(log_dir, exist_ok=True)
  116. os.makedirs(model_dir, exist_ok=True)
  118. # File for logging
  119. log_file = os.path.join(log_dir, "log.txt")
  121. # Load settings
  122. max_epochs = config["MAX_EPOCHS"]
  123. learning_rate = float(config["LEARNING_RATE"])
  124. use_dropout = config["USE_DROPOUT"]
  125. dropout = config["DROPOUT"]
  126. early_stopping_patience = config["EARLY_STOPPING_PATIENCE"]
  127. channels_setting = tuple(config["CHANNELS_SETTING"])
  128. cycle_duration = config["CYCLE_DURATION"]
  129. cycle_factor = config["CYCLE_FACTOR"]
  130. grad_accum_steps = config["GRAD_ACCUM_STEPS"]
  132. # Model creation
  133. if use_dropout:
  134. model = UNet(
  135. spatial_dims=3,
  136. in_channels=1,
  137. out_channels=1,
  138. channels=channels_setting,
  139. strides=(2, 2, 2, 2),
  140. dropout=dropout,
  141. ).to(device)
  142. else:
  143. model = UNet(
  144. spatial_dims=3,
  145. in_channels=1,
  146. out_channels=1,
  147. channels=channels_setting,
  148. strides=(2, 2, 2, 2),
  149. ).to(device)
  151. loss_function = CombinedLoss(config=config)
  153. optimizer = optim.Adam(model.parameters(), learning_rate)
  154. scheduler = CosineAnnealingLR(
  155. optimizer, T_max=int((cycle_duration + 0.5) / 2), eta_min=0
  156. )
  158. # Calculate the plane-switching epochs (Algorithm 2 in the paper)
  159. ax_epochs, co_epochs, sa_epochs = calculate_cycle_epochs(
  160. max_epochs, cycle_duration, cycle_factor
  161. )
  162. print("ax: ", ax_epochs)
  163. print("co: ", co_epochs)
  164. print("sa: ", sa_epochs)
  165. print("Max Epoch: ", max_epochs)
  167. # Initialize checkpoint-related variables
  168. best_val_loss = float("inf")
  169. no_improvement_epochs = 0
  170. selection = "axial" # Default plane
  171. start_epoch = 0
  172. early_stopping_trigger_epoch = (
  173. ax_epochs[2] if len(ax_epochs) > 2 else 0
  174. ) # Early stopping is activated in the late stages to prevent it from activating when the model is overfitted into a specific plane
  175. print(
  176. f"Early stopping will start to check after epoch: {early_stopping_trigger_epoch}"
  177. )
  178. best_checkpoint_path = None
  180. # Load checkpoint if provided
  181. if checkpoint_path:
  182. start_epoch, best_val_loss, selection, no_improvement_epochs = load_checkpoint(
  183. checkpoint_path, model, optimizer, scheduler
  184. )
  185. else:
  186. print("No checkpoint provided.")
  188. log_message = ""
  189. epoch_time_start = time.time()
  191. # Main training loop
  192. for epoch in range(start_epoch, max_epochs):
  193. # Print elapsed time every 50 epochs
  194. if epoch > 0 and epoch % 50 == 0:
  195. epoch_time_end = time.time()
  196. elapsed = epoch_time_end - epoch_time_start
  197. hours = int(elapsed // 3600)
  198. minutes = int((elapsed % 3600) // 60)
  199. seconds = int(elapsed % 60)
  200. print(f"{epoch} epochs done: {hours}h {minutes}m {seconds}s elapsed.")
  201. epoch_time_start = time.time()
  203. # Check if we need to switch planes
  204. new_selection, reset_best = update_plane_selection(
  205. epoch, ax_epochs, co_epochs, sa_epochs, selection
  206. )
  208. if new_selection != selection:
  209. selection = new_selection
  210. if selection == "axial":
  211. print("Perceptual Loss is now based on axial slices.")
  212. elif selection == "coronal":
  213. print("Perceptual Loss is now based on coronal slices.")
  214. elif selection == "sagittal":
  215. print("Perceptual Loss is now based on sagittal slices.")
  217. if reset_best:
  218. best_val_loss = float("inf")
  220. # Train for one epoch
  221. train_loss = train_one_epoch(
  222. model=model,
  223. train_loader=train_loader,
  224. optimizer=optimizer,
  225. loss_function=loss_function,
  226. device=device,
  227. selection=selection,
  228. grad_accum_steps=grad_accum_steps,
  229. )
  231. # Validate
  232. val_loss = validate_one_epoch(
  233. model=model,
  234. val_loader=val_loader,
  235. loss_function=loss_function,
  236. device=device,
  237. selection=selection,
  238. )
  240. print(
  241. f"Epoch {epoch+1}/{max_epochs} | Train Loss: {train_loss:.4f}, Validation Loss: {val_loss:.4f}"
  242. )
  243. log_message += (
  244. f"Epoch {epoch+1}, Train Loss: {train_loss}, Validation Loss: {val_loss}\n"
  245. )
  247. scheduler.step()
  249. # Update checkpoint and log
  250. best_val_loss, no_improvement_epochs, stop_training, best_checkpoint_path = (
  251. update_checkpoint_and_log(
  252. average_val_loss=val_loss,
  253. selection=selection,
  254. epoch=epoch,
  255. no_improvement_epochs=no_improvement_epochs,
  256. log_message=log_message,
  257. model=model,
  258. optimizer=optimizer,
  259. scheduler=scheduler,
  260. best_val_loss=best_val_loss,
  261. model_dir=model_dir,
  262. early_stopping_patience=early_stopping_patience,
  263. log_file=log_file,
  264. early_stopping_trigger_epoch=early_stopping_trigger_epoch,
  265. best_checkpoint_path=best_checkpoint_path,
  266. )
  267. )
  269. if stop_training:
  270. print(f"Early stopping at epoch {epoch+1}")
  271. break
  274. def main():
  275. seed_everything(deterministic=True) # Fix seed
  276. args = parse_arguments()
  277. config = load_config(args.config)
  278. if isinstance(config["CUDA_SETTING"], int):
  279. torch.cuda.set_device(config["CUDA_SETTING"])
  280. device = torch.device(
  281. config["CUDA_SETTING"] if torch.cuda.is_available() else "cpu"
  282. )
  283. train_loader, val_loader = load_data(config)
  284. train_model(
  285. train_loader=train_loader,
  286. val_loader=val_loader,
  287. config=config,
  288. device=device,
  289. checkpoint_path=args.checkpoint,
  290. )
  293. if __name__ == "__main__":
  294. main()

train.py at commit 1ff09cd, no license · at the source

Overview

Authors: Junho Moon1, Symac Kim2, Haejun Chung1,2,3, Ikbeom Jang4,5,6, Alzheimer's Disease Neuroimaging Initiative
ORCID iDs: Junho Moon, Haejun Chung, Ikbeom Jang
  1. Department of Artificial Intelligence Semiconductor Engineering, Hanyang University, Seoul, South Korea
  2. Department of Artificial Intelligence, Hanyang University, Seoul, South Korea
  3. Department of Electronic Engineering, Hanyang University, Seoul, South Korea
  4. Division of Computer Engineering, Hankuk University of Foreign Studies, Yongin, South Korea
  5. Division of AI Data Convergence, Hankuk University of Foreign Studies, Yongin, South Korea
  6. Division of Language & AI, Hankuk University of Foreign Studies, Seoul, South Korea
Institutions: Hanyang University (South Korea); Hankuk University of Foreign Studies (South Korea)
Journal: Human brain mapping, volume 47, issue 5, article e70508
Dates: received 20 September 2025; accepted 7 March 2026; published online 12 April 2026; in print April 2026
Type: Research article · Language: English
License: CC BY-NC
Identifiers: DOI 10.1002/hbm.70508 · PMID 41968275 · PMCID PMC13070738 · OpenAlex W4399837632
Open access: gold, a free copy (OpenAlex)
Preprint: arxiv:2406.12632
Status: code verified
Categories: structural MRI / diffusion (modality), PET / SPECT (modality), human (organism), Alzheimer's / dementia (population), cognitive (subfield)
Methods: Connectivity, Statistics, Machine learning, fMRI & imaging
Keywords: 3D image translation, Alzheimer's disease, cross‐modal image synthesis, cyclic 2.5D perceptual loss, dementia, MRI, PET
MeSH: Alzheimer Disease*, Brain*, Imaging, Three-Dimensional*, Magnetic Resonance Imaging*, Neuroimaging*, Positron-Emission Tomography*, tau Proteins*, Generative Artificial Intelligence, Humans (* major topic)
Topic: Computer Graphics and Visualization Techniques (Computer Graphics and Computer-Aided Design, Computer Science), according to OpenAlex
Funding: Artificial Intelligence Graduate School Program (Hanyang University), Ministry of Science and ICT (MSIT) (RS-2023-00261368, RS-2020-II201373); Institute for Information & Communication Technology Planning & Evaluation (IITP), Ministry of Science and ICT (MSIT) (IITP-(2025)-RS-2023-00253914, RS-2024-00412644); National Research Foundation of Korea (NRF), Ministry of Science and ICT (MSIT) (RS-2024-00455720, RS-2024-00414119, RS-2024-00338048); Ministry of Trade, Industry, and Energy (MOTIE) (RS-2025-13002970, RS‐2025‐13002970); Ministry of Science and ICT, South Korea (RS‐2024‐00414119, RS‐2020‐II201373, RS‐2024‐00412644, RS‐2024‐00455720, RS‐2023‐00261368, IITP‐(2025)‐RS‐2023‐00253914, RS‐2024‐00338048); Hankuk University of Foreign Studies Research Fund of 2025; Hankuk University of Foreign Studies; Korea National Institute of Health (2025ER040300, 2024ER040700); National Supercomputing Center, Korea Institute of Science and Technology Information (KSC-2024-CRE-0021, KSC-2025-CRE-0065, KSC‐2025‐CRE‐0065, KSC‐2024‐CRE‐0021); Ministry of Health and Welfare (RS‐2025‐02220534); Korea Creative Content Agency (KOCCA), Ministry of Culture, Sports and Tourism (MCST) (RS-2024-00332210); Korea Health Industry Development Institute (KHIDI), Ministry of Health & Welfare (MOHW) (RS-2025-02220534)
Citations: not cited yet (Europe PMC); 138 references in the paper

Abstract

Positron emission tomography (PET) provides an in vivo molecular marker for various diseases, including Alzheimer's disease and related dementias (ADRD). PET has become increasingly integrated into diagnostic decision‐making, disease staging, and clinical trial enrichment. However, its widespread use remains constrained by high costs, government regulations, and the invasiveness of radiotracer injection. Modern diagnostic frameworks emphasize the importance of multimodal biomarker assessment, such as the “amyloid/tau/neurodegeneration” (A/T/N) framework for Alzheimer's disease; however, they are constrained by these barriers. Medical image synthesis or translation offers a potential solution by enabling the reconstruction of unavailable modalities. The clinical utility of PET depends on accurately capturing regional uptake patterns rather than exact voxel‐wise intensities, motivating the use of perceptual loss functions to assess higher‐level semantic features in generative models. While 2D, 3D, and 2.5D perceptual losses are utilized in 3D synthesis, each encounters challenges, including limited volumetric context, the scarcity of pretrained 3D models, and difficulty balancing optimization across anatomical planes. In this work, we address cross‐modal synthesis of tau PET from structural magnetic resonance imaging (MRI), generating 3D pseudo‐[18F]flortaucipir standardized uptake value ratio (SUVR) maps from 3D T1‐weighted MR images. We propose a cyclic 2.5D perceptual loss that cyclically optimizes the axial, coronal, and sagittal planes over training phases, thereby enhancing volumetric consistency. Furthermore, we standardize PET SUVRs by scanner manufacturer, reducing inter‐manufacturer variability and better preserving high‐uptake regions. We evaluate the proposed approach on cohorts spanning the ADRD spectrum using data from the Alzheimer's Disease Neuroimaging Initiative and the Standardized Centralized Alzheimer's Disease and Related Dementias Neuroimaging cohort. Our approach is broadly applicable across various generative frameworks and achieves high quantitative and qualitative performance on diverse architectures, including U‐Net, UNETR, SwinUNETR, CycleGAN, and Pix2Pix. Notably, it achieves better agreement between synthesized SUVRs and measured PET scans in key brain regions relevant to Alzheimer‐type tau pathology. The code is publicly available at https://github.com/labhai/Cyclic‐2.5D‐Perceptual‐Loss.

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

Repository

Its files are read in the Code ↔ Paper reader above, with 3 matches between paragraphs and lines of code.

labhai/Cyclic-2.5D-Perceptual-Loss

License: none: the authors keep all their rights
State: the link answers, verified on 28 September 2026
Evidence: files inventoried
Commit: 1ff09cd443a70018b6ad45a4656495b6814caef2, 17 April 2026
Languages: Python (5)
Size: 9 files, 5 scripts
Software Heritage: not archived
Found in: “Data Availability Statement”
Holds: README, environment (Code/requirements.txt)
Not found: license file, CITATION.cff, tests, continuous integration, documentation
Tools: MONAI (4 files), PyTorch (4 files), NumPy (2 files), NiBabel (1 file)
Availability: 1 check, the latest on 28 September 2026: the link answers
  • 28 September 2026: the link answers
6 files
At the source: github.com/labhai/Cyclic-2.5D-Perceptual-Loss

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;
  • 5 scripts, each with its path and the digest of its content;
  • 3 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

The code for the proposed method is available at https://github.com/labhai/Cyclic‐2.5D‐Perceptual‐Loss (https://github.com/labhai/Cyclic-2.5D-Perceptual-Loss). Data used in preparation of this article were obtained from the Alzheimer's Disease Neuroimaging Initiative (ADNI) database (adni.loni.usc.edu). As such, the investigators within the ADNI contributed to the design and implementation of ADNI and/or provided data but did not participate in analysis or writing of this report. A complete listing of ADNI investigators can be found at: http://adni.loni.usc.edu/wp‐content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf (http://adni.loni.usc.edu/wp-content/uploads/how_to_apply/ADNI_Acknowledgement_List.pdf). Data used in the preparation of this article were also obtained from the National Alzheimer's Coordinating Center (NACC) database, including MRI and/or PET data from the Standardized Centralized Alzheimer's and Related Dementias Neuroimaging (SCAN) Initiative. NACC/SCAN data are available to qualified researchers via the NACC Data Request Process and are subject to the NACC Data Use Agreement; the authors do not have permission to share NACC/SCAN data directly.

Reproduced under the paper's license (CC BY-NC), 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, 5 authors, 7 keywords, 9 MeSH terms, 12 funders, 124 references.

Cite

This paper

Moon, J., Kim, S., Chung, H., Jang, I., & Alzheimer's Disease Neuroimaging Initiative. (2026). Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET. Human brain mapping, 47(5), e70508. https://doi.org/10.1002/hbm.70508

BibTeX

@article{moon2026cyclic,
author = {Moon, Junho and Kim, Symac and Chung, Haejun and Jang, Ikbeom and {Alzheimer's Disease Neuroimaging Initiative}},
title = {{Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET}},
journal = {Human brain mapping},
year = {2026},
month = apr,
volume = {47},
number = {5},
pages = {e70508},
publisher = {Wiley},
issn = {1065-9471},
doi = {10.1002/hbm.70508},
url = {https://doi.org/10.1002/hbm.70508},
pmid = {41968275},
pmcid = {PMC13070738}
}

RIS

TY - JOUR
AU - Moon, Junho
AU - Kim, Symac
AU - Chung, Haejun
AU - Jang, Ikbeom
AU - Alzheimer's Disease Neuroimaging Initiative
TI - Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET
T2 - Human brain mapping
J2 - Hum Brain Mapp
PY - 2026
DA - 2026/04/01
VL - 47
IS - 5
SP - e70508
SN - 1065-9471
PB - Wiley
DO - 10.1002/hbm.70508
UR - https://doi.org/10.1002/hbm.70508
LA - en
ER -

CSL-JSON

{
"id": "10.1002/hbm.70508",
"type": "article-journal",
"title": "Cyclic 2.5D Perceptual Loss for Cross-Modal 3D Medical Image Synthesis: T1w MRI to Tau PET",
"container-title": "Human brain mapping",
"author": [
{
"family": "Moon",
"given": "Junho"
},
{
"family": "Kim",
"given": "Symac"
},
{
"family": "Chung",
"given": "Haejun"
},
{
"family": "Jang",
"given": "Ikbeom"
},
{
"literal": "Alzheimer's Disease Neuroimaging Initiative"
}
],
"container-title-short": "Hum Brain Mapp",
"volume": "47",
"issue": "5",
"page": "e70508",
"DOI": "10.1002/hbm.70508",
"PMID": "41968275",
"PMCID": "PMC13070738",
"ISSN": "1065-9471",
"publisher": "Wiley",
"URL": "https://doi.org/10.1002/hbm.70508",
"language": "en",
"issued": {
"date-parts": [
[
2026,
4,
1
]
]
}
}

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:10.64898/2026.05.06.26352540 [code]
Generating synthetic tau-PET scans in Alzheimer’s disease from MRI, blood biomarkers and demographics with deep learning
Journal: medRxiv (preprint)
In common: NiBabel, NumPy, PET / SPECT, Alzheimer's / dementia, structural MRI / diffusion, 11 references
[2] doi:10.3389/frai.2026.1771088 [code]
Few-shot deployment of pretrained MRI transformers in brain imaging tasks.
Journal: Frontiers in artificial intelligence
In common: MONAI, NiBabel, PyTorch, 1 other tool, structural MRI / diffusion, 5 references
[3] doi:10.1002/alz.71571 [code]
Vulnerability of anterior medial temporal lobe subregions to early tau-related neurodegeneration in Alzheimer's disease: Converging evidence from tau-PET and plasma p-tau217.
Journal: Alzheimer's & dementia : the journal of the Alzheimer's Association
In common: PyTorch, NumPy, PET / SPECT, Alzheimer's / dementia, structural MRI / diffusion, 6 references
[4] doi:10.1002/alz.71822 [code]
SynthPET: A 3D generative AI approach for FDG-PET image synthesis from T1-weighted MRI and ASL CBF in Alzheimer's disease.
Journal: Alzheimer's & dementia : the journal of the Alzheimer's Association
In common: PET / SPECT, Alzheimer's / dementia, structural MRI / diffusion, 7 references
[5] doi:10.1126/sciadv.aee2305 [code]
Prediction of mild cognitive impairment progression using time-sensitive multimodal biomarkers.
Journal: Science advances
In common: NiBabel, NumPy, PET / SPECT, Alzheimer's / dementia, structural MRI / diffusion, 6 references
[6] doi:10.1093/brain/awaf375 [code]
Reference proteins to improve Core 1 and Core 2 Alzheimer's disease CSF and plasma biomarkers.
Journal: Brain : a journal of neurology
In common: NumPy, PET / SPECT, Alzheimer's / dementia, 6 references
[7] doi:10.3389/fnagi.2026.1847611 [code]
APOE ε4-associated hippocampal atrophy trajectories across the Alzheimer's disease continuum: a systematic review, meta-analysis, and longitudinal validation.
Journal: Frontiers in aging neuroscience
In common: NumPy, Alzheimer's / dementia, structural MRI / diffusion, 7 references
[8] doi:10.34133/research.1392 [code]
Neuroimaging Epicenters as Vulnerable Nodes in Plasma p-tau217/Aβ42-Positive Alzheimer's Disease.
Journal: Research (Washington, D.C.)
In common: NiBabel, NumPy, Alzheimer's / dementia, 6 references
[9] doi:10.3389/fnut.2026.1837406 [code]
MIND diet moderates the associations between cerebrovascular and neurodegenerative disease burden and cognition.
Journal: Frontiers in nutrition
In common: Alzheimer's / dementia, structural MRI / diffusion, cognitive, 7 references
[10] doi:10.1038/s41598-026-54109-8
A computational model to describe multi-regional brain architecture during neurodegeneration in Alzheimer's disease.
Journal: Scientific reports
In common: Alzheimer's / dementia, structural MRI / diffusion, 6 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.