OSCR

Pathology-Anchored Transcranial Sonography: A Cascaded Super-Resolution Deep Learning System for Early-Stage Parkinson's Disease Grading.

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 › Comparison studies ↔ Final_test/Process_fold_Efficient.py, lines 182–233 · score 0.51 · EfficientNet b7, b0, Transformer, models
  2. [2] § Methods › Comparison studies ↔ Final_test/Process_fold_efficient_v2.py, lines 182–220 · score 0.51 · EfficientNet b0, b7, Transformer, models
  3. [3] § Methods › Cascaded super-resolution deep learning system ↔ Final_test/Process_fold_transformer.py, lines 64–95 · score 0.51 · random rotation, augmentation, mirror, flipping, resized, classification

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 · 353 lines · 14 KB · no license · 1 match

  1. import sys
  2. from tqdm import tqdm
  3. import torch
  4. from torch import nn
  5. from torchvision import transforms
  6. from torch.utils.data import DataLoader, Dataset
  7. from sklearn.metrics import roc_curve, auc, confusion_matrix, classification_report
  8. from sklearn.model_selection import StratifiedKFold
  9. from PIL import Image
  10. import numpy as np
  11. import matplotlib.pyplot as plt
  12. import os
  13. from torchvision import transforms, models
  14. # ====================
  15. # 配置设备
  16. # ====================
  17. device = "cuda" if torch.cuda.is_available() else "mps" if torch.backends.mps.is_available() else "cpu"
  18. print(f"Using {device} device")
  20. # ====================
  21. # 数据集类
  22. # ====================
  23. class VideoFrameDataset(Dataset):
  24. def __init__(self, file_path, transform=None):
  25. self.file_path = file_path
  26. self.transform = transform
  27. self.imgs = []
  28. self.labels = []
  29. with open(self.file_path) as f:
  30. samples = [x.strip().rsplit(' ', 1) for x in f.readlines()]
  31. for img_path, label in samples:
  32. self.imgs.append(img_path)
  33. self.labels.append(int(label))
  35. def __len__(self):
  36. return len(self.imgs)
  38. def __getitem__(self, idx):
  39. image = Image.open(self.imgs[idx]).convert("RGB")
  40. if self.transform:
  41. image = self.transform(image)
  42. label = torch.tensor(self.labels[idx], dtype=torch.int64)
  43. return image, label
  45. # ====================
  46. # 数据增强
  47. # ====================
  48. data_transforms = transforms.Compose([
  49. transforms.Resize([300, 300]),
  50. transforms.RandomRotation(45),
  51. transforms.CenterCrop(256),
  52. transforms.RandomHorizontalFlip(p=0.5),
  53. transforms.RandomVerticalFlip(p=0.5),
  54. transforms.ColorJitter(brightness=0.4, contrast=0.4, saturation=0.4, hue=0.2),
  55. transforms.RandomGrayscale(p=0.2),
  56. transforms.ToTensor(),
  57. transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225]),
  58. ])
  60. # ====================
  61. # 模型保存和加载
  62. # ====================
  63. def save_model(model, path):
  64. torch.save(model.state_dict(), path)
  65. print(f"Model saved to {path}")
  67. def load_model(model, path):
  68. if os.path.exists(path):
  69. model.load_state_dict(torch.load(path))
  70. print(f"Model loaded from {path}")
  71. else:
  72. print(f"Model path {path} does not exist.")
  74. # ====================
  75. # 定义训练和测试函数
  76. # ====================
  77. def train(dataloader, model, loss_fn, optimizer):
  78. model.train()
  79. total_loss = 0
  80. correct = 0
  81. for X, y in tqdm(dataloader, desc="Training"):
  82. X, y = X.to(device), y.to(device)
  83. pred = model(X)
  84. loss = loss_fn(pred, y)
  85. optimizer.zero_grad()
  86. loss.backward()
  87. optimizer.step()
  88. total_loss += loss.item() * X.size(0)
  89. correct += (pred.argmax(1) == y).sum().item()
  90. avg_loss = total_loss / len(dataloader.dataset)
  91. accuracy = correct / len(dataloader.dataset)
  92. return avg_loss, accuracy
  94. def test(dataloader, model, threshold=0.1):
  95. model.eval()
  96. total_loss = 0
  97. correct = 0
  98. all_labels = []
  99. all_preds = []
  100. with torch.no_grad():
  101. for X, y in tqdm(dataloader, desc="Testing"):
  102. X, y = X.to(device), y.to(device)
  103. pred = model(X)
  104. loss = loss_fn(pred, y)
  105. total_loss += loss.item() * X.size(0)
  106. all_labels.extend(y.cpu().numpy())
  107. pred_probs = torch.softmax(pred, dim=1)
  108. pred_classes = (pred_probs[:, 1] > threshold).long() # Class 1 is predicted as positive
  109. all_preds.extend(torch.softmax(pred, dim=1).cpu().numpy())
  110. correct += (pred.argmax(1) == y).sum().item()
  111. avg_loss = total_loss / len(dataloader.dataset)
  112. accuracy = correct / len(dataloader.dataset)
  113. return np.array(all_labels), np.array(all_preds), avg_loss, accuracy
  115. # ====================
  116. # 计算与显示评估指标
  117. # ====================
  118. def calculate_metrics(labels, preds, num_classes):
  119. pred_classes = preds.argmax(axis=1)
  120. cm = confusion_matrix(labels, pred_classes)
  121. class_report = classification_report(labels, pred_classes, target_names=[f'Class {i}' for i in range(num_classes)], output_dict=True)
  122. print("Classification Report:")
  123. print(classification_report(labels, pred_classes, target_names=[f'Class {i}' for i in range(num_classes)]))
  124. plt.figure(figsize=(8, 8))
  125. plt.imshow(cm, interpolation='nearest', cmap=plt.cm.Blues)
  126. plt.title("Confusion Matrix")
  127. plt.colorbar()
  128. tick_marks = np.arange(num_classes)
  129. plt.xticks(tick_marks, [f'Class {i}' for i in range(num_classes)], rotation=45)
  130. plt.yticks(tick_marks, [f'Class {i}' for i in range(num_classes)])
  131. plt.ylabel('True Label')
  132. plt.xlabel('Predicted Label')
  133. thresh = cm.max() / 2.0
  134. for i, j in np.ndindex(cm.shape):
  135. plt.text(j, i, f"{cm[i, j]}", horizontalalignment="center", color="white" if cm[i, j] > thresh else "black")
  136. plt.tight_layout()
  137. plt.show()
  138. return class_report
  140. def plot_roc_curve(labels, preds, fold):
  141. plt.figure(figsize=(10, 8))
  142. fpr = dict()
  143. tpr = dict()
  144. roc_auc = dict()
  145. for i in range(4): # 几分类任务
  146. fpr[i], tpr[i], _ = roc_curve(labels == i, preds[:, i])
  147. roc_auc[i] = auc(fpr[i], tpr[i])
  148. plt.plot(fpr[i], tpr[i], label=f'Class {i} (AUC = {roc_auc[i]:.2f})')
  149. plt.plot([0, 1], [0, 1], 'k--', lw=2)
  150. plt.xlim([0.0, 1.0])
  151. plt.ylim([0.0, 1.05])
  152. plt.xlabel('False Positive Rate')
  153. plt.ylabel('True Positive Rate')
  154. plt.title(f'ROC Curve (Fold {fold})')
  155. plt.legend(loc="lower right")
  156. plt.grid()
  157. plt.show()
  158. return roc_auc
  160. def plot_training_curves(train_losses, valid_losses, train_accuracies, valid_accuracies, fold):
  161. epochs = range(1, len(train_losses) + 1)
  162. plt.figure(figsize=(12, 5))
  163. plt.subplot(1, 2, 1)
  164. plt.plot(epochs, train_losses, label='Train Loss')
  165. plt.plot(epochs, valid_losses, label='Valid Loss')
  166. plt.xlabel('Epochs')
  167. plt.ylabel('Loss')
  168. plt.title(f'Loss Curve (Fold {fold})')
  169. plt.legend()
  170. plt.grid()
  171. plt.subplot(1, 2, 2)
  172. plt.plot(epochs, train_accuracies, label='Train Accuracy')
  173. plt.plot(epochs, valid_accuracies, label='Valid Accuracy')
  174. plt.xlabel('Epochs')
  175. plt.ylabel('Accuracy')
  176. plt.title(f'Accuracy Curve (Fold {fold})')
  177. plt.legend()
  178. plt.grid()
  179. plt.tight_layout()
  180. plt.show()
  182. # ====================
  183. # 主程序
  184. # ====================
  185. if __name__ == "__main__":
  186. mode = "test" # 训练模式："train"，测试模式："test"
  187. model_path = "./saved_models/fold_1_model.pth"
  188. new_data_path = "./new_data/test_data.txt" # 新数据集路径（测试模式）
  190. dataset = VideoFrameDataset(file_path='./data/train_final.txt', transform=data_transforms)
  191. kfold = StratifiedKFold(n_splits=5, shuffle=True, random_state=42)
  192. labels = np.array(dataset.labels)
  194. if mode == "train":
  195. os.makedirs('./saved_models', exist_ok=True)
  197. all_fold_metrics = []
  199. for fold, (train_idx, valid_idx) in enumerate(kfold.split(np.zeros(len(labels)), labels), 1):
  200. print(f'Fold {fold}')
  201. train_subset = torch.utils.data.Subset(dataset, train_idx)
  202. valid_subset = torch.utils.data.Subset(dataset, valid_idx)
  204. train_dataloader = DataLoader(train_subset, batch_size=32, shuffle=True)
  205. valid_dataloader = DataLoader(valid_subset, batch_size=32, shuffle=False)
  207. # 修改为 EfficientNet-B0
  208. model = models.efficientnet_b7(pretrained=False)
  209. model.classifier[1] = nn.Linear(model.classifier[1].in_features, 4)
  210. model = model.to(device)
  212. optimizer = torch.optim.Adam(model.parameters(), lr=0.001, weight_decay=1e-5)
  213. scheduler = torch.optim.lr_scheduler.ReduceLROnPlateau(optimizer, mode='min', patience=3, factor=0.5, verbose=True)
  214. loss_fn = nn.CrossEntropyLoss()
  216. train_losses = []
  217. valid_losses = []
  218. train_accuracies = []
  219. valid_accuracies = []
  221. # 创建日志文件
  222. log_file = open(f'./logs/final_efficient07_fold_{fold}_log.txt', 'w')
  223. log_file.write("Epoch\tTrain Loss\tTrain Acc\tValid Loss\tValid Acc\n")
  225. for epoch in range(50): # 限制到50个epoch
  226. train_loss, train_accuracy = train(train_dataloader, model, loss_fn, optimizer)
  227. _, _, valid_loss, valid_accuracy = test(valid_dataloader, model)
  228. scheduler.step(valid_loss)
  230. train_losses.append(train_loss)
  231. valid_losses.append(valid_loss)
  232. train_accuracies.append(train_accuracy)
  233. valid_accuracies.append(valid_accuracy)
  235. # 打印每个 epoch 的指标
  236. print(f"Epoch {epoch + 1}:")
  237. print(f" Train Loss: {train_loss:.4f}, Train Accuracy: {train_accuracy:.4f}")
  238. print(f" Valid Loss: {valid_loss:.4f}, Valid Accuracy: {valid_accuracy:.4f}")
  240. # 将指标写入日志文件
  241. log_file.write(
  242. f"{epoch + 1}\t{train_loss:.4f}\t{train_accuracy:.4f}\t{valid_loss:.4f}\t{valid_accuracy:.4f}\n")
  244. # 关闭日志文件
  245. log_file.close()
  247. # 绘制训练曲线
  248. # plot_training_curves(train_losses, valid_losses, train_accuracies, valid_accuracies, fold)
  250. # 验证集测试与结果计算
  251. fold_labels, fold_preds, _, fold_accuracy = test(valid_dataloader, model)
  252. # fold_auc = plot_roc_curve(fold_labels, fold_preds, fold)
  254. # 计算分类报告和混淆矩阵
  255. # fold_metrics = calculate_metrics(fold_labels, fold_preds, num_classes=4)
  257. # 存储折叠结果
  258. # all_fold_metrics.append({
  259. # 'accuracy': fold_accuracy,
  260. # 'auc': fold_auc,
  261. # 'loss': valid_losses[-1],
  262. # 'metrics': fold_metrics
  263. # })
  265. # 保存模型
  266. model_save_path = f'./saved_models/efficientnet_b7_{fold}_final.pth'
  267. torch.save(model.state_dict(), model_save_path)
  268. print(f"Model for Fold {fold} saved at {model_save_path}")
  270. # ====================
  271. # 综合评估
  272. # ====================
  273. overall_accuracy = np.mean([m['accuracy'] for m in all_fold_metrics])
  274. overall_auc = np.mean([np.mean(list(m['auc'].values())) for m in all_fold_metrics])
  275. overall_loss = np.mean([m['loss'] for m in all_fold_metrics])
  277. print(f"\nFinal Overall Results:")
  278. print(f"Accuracy: {overall_accuracy:.4f}")
  279. print(f"AUC: {overall_auc:.4f}")
  280. print(f"Loss: {overall_loss:.4f}")
  282. # 平均 F1-Score, Precision, Recall
  283. precision_scores = [np.mean([m['metrics'][f'Class {i}']['precision'] for i in range(3)]) for m in all_fold_metrics]
  284. recall_scores = [np.mean([m['metrics'][f'Class {i}']['recall'] for i in range(3)]) for m in all_fold_metrics]
  285. f1_scores = [np.mean([m['metrics'][f'Class {i}']['f1-score'] for i in range(3)]) for m in all_fold_metrics]
  287. print(f"Average Precision: {np.mean(precision_scores):.4f}")
  288. print(f"Average Recall: {np.mean(recall_scores):.4f}")
  289. print(f"Average F1-Score: {np.mean(f1_scores):.4f}")
  290. elif mode == "test":
  291. test_dataset = VideoFrameDataset(file_path='./data/test_final.txt', transform=data_transforms)
  292. test_dataloader = DataLoader(test_dataset, batch_size=32, shuffle=False)
  294. # 修改为 EfficientNet-B0
  295. model = models.efficientnet_b7(pretrained=False)
  296. model.classifier[1] = nn.Linear(model.classifier[1].in_features, 4)
  297. model = model.to(device)
  298. loss_fn = nn.CrossEntropyLoss()
  300. # 遍历保存的模型并测试每个模型
  301. saved_models_dir = './saved_models/'
  302. if not os.path.exists(saved_models_dir):
  303. print("No saved models found.")
  304. sys.exit(1) # 非零状态码表示异常退出
  306. all_test_results = [] # 存储每个模型的测试结果
  308. for fold in range(1, 6): # 假设你有5折模型
  309. model_path = os.path.join(saved_models_dir, f'efficientnet_b7_{fold}_final.pth')
  310. if not os.path.exists(model_path):
  311. print(f"Model for Fold {fold} not found at {model_path}. Skipping...")
  312. continue
  314. print(f"Loading model from {model_path}...")
  315. model.load_state_dict(torch.load(model_path))
  316. model.eval()
  318. # 测试模型
  319. test_labels, test_preds, test_loss, test_accuracy = test(test_dataloader, model)
  320. test_auc = plot_roc_curve(test_labels, test_preds, fold)
  321. test_metrics = calculate_metrics(test_labels, test_preds, num_classes=4)
  323. # 保存测试结果
  324. all_test_results.append({
  325. 'fold': fold,
  326. 'accuracy': test_accuracy,
  327. 'auc': test_auc,
  328. 'loss': test_loss,
  329. 'metrics': test_metrics
  330. })
  332. print(f"Fold {fold} Test Results:")
  333. print(f" Accuracy: {test_accuracy:.4f}")
  334. print(f" Loss: {test_loss:.4f}")
  335. # 打印 AUC 信息
  336. print(f" AUCs:")
  337. for class_name, auc_value in test_auc.items():
  338. print(f" {class_name}: {auc_value:.4f}")
  339. avg_auc = np.mean(list(test_auc.values()))
  340. print(f" Average AUC: {avg_auc:.4f}")
  342. # 汇总所有折的测试结果
  343. if all_test_results:
  344. overall_test_accuracy = np.mean([r['accuracy'] for r in all_test_results])
  345. overall_test_auc = np.mean([np.mean(list(r['auc'].values())) for r in all_test_results])
  346. overall_test_loss = np.mean([r['loss'] for r in all_test_results])
  348. print("\nFinal Test Results:")
  349. print(f" Overall Accuracy: {overall_test_accuracy:.4f}")
  350. print(f" Overall AUC: {overall_test_auc:.4f}")
  351. print(f" Overall Loss: {overall_test_loss:.4f}")
  352. else:
  353. print("No valid test results were obtained.")

Process_fold_Efficient.py at commit ae26dfc, no license · at the source

Overview

Authors: Yishen Zhao1,2,3, Weiguo Cui1,2,3, Shuang Liang1,2,3, Yu Gu1,2,3, Xue Bai1,2,3, Lu Zhang1,2,3
  1. School of Biomedical Engineering, Capital Medical University, Beijing, 100069 China
  2. Beijing Key Laboratory of Fundamicationental Research on Biomechanics in Clinical Application, Capital Medical University, Beijing, 100069 China
  3. Laboratory for Clinical Medicine, Capital Medical University, Beijing, 100069 China
Institutions: Capital Medical University (China)
Journal: NPJ Parkinson's disease, volume 12, issue 1, article 145
Dates: received 23 September 2025; accepted 29 March 2026; published online 11 April 2026
Type: Research article · Language: English
License: CC BY-NC-ND
Identifiers: DOI 10.1038/s41531-026-01348-1 · PMID 41965365 · PMCID PMC13269911 · OpenAlex W7153630582
Open access: gold, a free copy (OpenAlex)
Status: code verified
Categories: Parkinson's (population), clinical / translational (subfield)
Methods: Connectivity, Statistics, Machine learning, fMRI & imaging
Keywords: Computational biology and bioinformatics, Diseases, Medical research, Neurology, Neuroscience
Topic: Neurological disorders and treatments (Neurology, Medicine), according to OpenAlex
Funding: Beijing Municipal Natural Science Foundation (Z220015)
Citations: cited by 1 paper (Europe PMC); 34 references in the paper

Abstract

The abstract is not reproduced here: the paper's license (CC BY-NC-ND) does not allow it. Read it in the paper, at the publisher or on Europe PMC.

Repository

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

DrZhaoys/Early-Stage-Parkinson-s-Disease-Grading

License: none: the authors keep all their rights
State: the link answers, verified on 29 September 2026
Evidence: files inventoried
Commit: ae26dfc225d58fd3c8e67d27e8edefd085daa0d6, 17 March 2026
Languages: Python (45)
Size: 46 files, 45 scripts
Software Heritage: not archived
Found in: “Code availability”
Not found: README, license file, CITATION.cff, environment file, tests, continuous integration, documentation
Tools: PyTorch (36 files), Pillow (35 files), Matplotlib (33 files), NumPy (30 files), scikit-learn (22 files), OpenCV (16 files), pandas (16 files), scikit-image (16 files), Hugging Face Transformers (1 file)
Availability: 1 check, the latest on 29 September 2026: the link answers
  • 29 September 2026: the link answers
45 files
At the source: github.com/DrZhaoys/Early-Stage-Parkinson-s-Disease-Grading

Code availability statement

The paper has a code availability statement. Its license (CC BY-NC-ND) does not allow reproducing it here; in short, from what the harvester recognized in it:

Read it in the paper: doi.org/10.1038/s41531-026-01348-1.

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;
  • 45 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

Datasets cited

Data availability statement

The paper has a data availability statement. Its license (CC BY-NC-ND) does not allow reproducing it here; in short, from what the harvester recognized in it:

Read it in the paper: doi.org/10.1038/s41531-026-01348-1.

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, 29 September 2026: the first record

Recorded: type, language, journal, volume, issue, pages, dates, 6 authors, 5 keywords, 1 funder, 26 references.

Cite

This paper

Zhao, Y., Cui, W., Liang, S., Gu, Y., Bai, X., & Zhang, L. (2026). Pathology-Anchored Transcranial Sonography: A Cascaded Super-Resolution Deep Learning System for Early-Stage Parkinson's Disease Grading. NPJ Parkinson's disease, 12(1), 145. https://doi.org/10.1038/s41531-026-01348-1

BibTeX

@article{zhao2026pathology,
author = {Zhao, Yishen and Cui, Weiguo and Liang, Shuang and Gu, Yu and Bai, Xue and Zhang, Lu},
title = {{Pathology-Anchored Transcranial Sonography: A Cascaded Super-Resolution Deep Learning System for Early-Stage Parkinson's Disease Grading}},
journal = {NPJ Parkinson's disease},
year = {2026},
month = apr,
volume = {12},
number = {1},
pages = {145},
publisher = {Nature Publishing Group},
issn = {2373-8057},
doi = {10.1038/s41531-026-01348-1},
url = {https://doi.org/10.1038/s41531-026-01348-1},
pmid = {41965365},
pmcid = {PMC13269911}
}

RIS

TY - JOUR
AU - Zhao, Yishen
AU - Cui, Weiguo
AU - Liang, Shuang
AU - Gu, Yu
AU - Bai, Xue
AU - Zhang, Lu
TI - Pathology-Anchored Transcranial Sonography: A Cascaded Super-Resolution Deep Learning System for Early-Stage Parkinson's Disease Grading
T2 - NPJ Parkinson's disease
J2 - NPJ Parkinsons Dis
PY - 2026
DA - 2026/04/11
VL - 12
IS - 1
SP - 145
SN - 2373-8057
PB - Nature Publishing Group
DO - 10.1038/s41531-026-01348-1
UR - https://doi.org/10.1038/s41531-026-01348-1
LA - en
ER -

CSL-JSON

{
"id": "10.1038/s41531-026-01348-1",
"type": "article-journal",
"title": "Pathology-Anchored Transcranial Sonography: A Cascaded Super-Resolution Deep Learning System for Early-Stage Parkinson's Disease Grading",
"container-title": "NPJ Parkinson's disease",
"author": [
{
"family": "Zhao",
"given": "Yishen"
},
{
"family": "Cui",
"given": "Weiguo"
},
{
"family": "Liang",
"given": "Shuang"
},
{
"family": "Gu",
"given": "Yu"
},
{
"family": "Bai",
"given": "Xue"
},
{
"family": "Zhang",
"given": "Lu"
}
],
"container-title-short": "NPJ Parkinsons Dis",
"volume": "12",
"issue": "1",
"page": "145",
"DOI": "10.1038/s41531-026-01348-1",
"PMID": "41965365",
"PMCID": "PMC13269911",
"ISSN": "2373-8057",
"publisher": "Nature Publishing Group",
"URL": "https://doi.org/10.1038/s41531-026-01348-1",
"language": "en",
"issued": {
"date-parts": [
[
2026,
4,
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.patter.2026.101538 [code]
A multi-modal foundation model for brain disease diagnosis and medical imaging.
Journal: Patterns (New York, N.Y.)
In common: Hugging Face Transformers, OpenCV, scikit-image, 6 other tools, clinical / translational
[2] doi:10.1038/s42003-026-10957-8 [code]
Brain defence by the extracellular matrix protein Cochlin.
Journal: Communications biology
In common: Hugging Face Transformers, OpenCV, scikit-image, 6 other tools
[3] doi:10.1038/s41467-026-76837-1 [code]
Drug screen and machine learning predict neuroprotective agents in a preclinical human model of childhood dementia.
Journal: Nature communications
In common: Hugging Face Transformers, OpenCV, scikit-image, 6 other tools
[4] 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: Hugging Face Transformers, OpenCV, scikit-image, 6 other tools
[5] doi:10.7554/elife.107933 [code]
Modality-agnostic decoding of vision and language from fMRI.
Journal: eLife
In common: Hugging Face Transformers, OpenCV, scikit-image, 6 other tools
[6] doi:10.1038/s43856-026-01606-6 [code]
Validation of remote multimodal AI screening for Parkinson disease across diverse settings.
Journal: Communications medicine
In common: Hugging Face Transformers, OpenCV, PyTorch, 4 other tools, Parkinson's, clinical / translational, 1 reference
[7] doi:10.1162/imag.a.1326 [code]
RAVEN: Robust, generalizable, multi-resolution structural MRI upsampling using autoencoders.
Journal: Imaging neuroscience (Cambridge, Mass.)
In common: Hugging Face Transformers, OpenCV, scikit-image, 5 other tools
[8] doi:10.1111/joa.70203 [code]
Two-step workflow integrating automatic registration and manual refinement for the accurate alignment of serial histological sections in 3D reconstruction.
Journal: Journal of anatomy
In common: Hugging Face Transformers, OpenCV, scikit-image, 5 other tools
[9] doi:10.1038/s43856-026-01817-x [code]
Visual prompt engineering for multimodal and irregularly sampled medical data.
Journal: Communications medicine
In common: Hugging Face Transformers, OpenCV, Pillow, 5 other tools, clinical / translational
[10] doi:10.1093/braincomms/fcag253 [code]
Disease detection and classification in temporal lobe epilepsy: step-wise versus simultaneous AI decision models in a multisite neuroimaging study.
Journal: Brain communications
In common: Hugging Face Transformers, OpenCV, Pillow, 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.