SALD_t1w / README.md
Gaurav08's picture
Update README.md
0b44cb8 verified
|
Raw
History Blame Contribute Delete
2.97 kB
metadata
dataset_info:
  features:
    - name: volume_id
      dtype: int32
    - name: slice_id
      dtype: int32
    - name: age
      dtype: int32
    - name: sex
      dtype: string
    - name: image
      dtype: image
  splits:
    - name: train
      num_bytes: 5053069526.112
      num_examples: 126464
  download_size: 4796863468
  dataset_size: 5053069526.112
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-nc-sa-4.0
language:
  - en
tags:
  - medical
  - mri
  - pretraining
  - SALD
size_categories:
  - 100K<n<1M

This dataset is derived from the Southwest University Adult Lifespan Dataset (SALD), a large-scale cross-sectional neuroimaging dataset designed to characterize brain structure and function across the adult lifespan. The study recruited community-dwelling adults between ages 19 and 80 from Southwest University in Chongqing, China. At each session, participants contributed T1-weighted structural MRI images alongside behavioral and cognitive assessments. Here, we provide a 2D slice-based version of T1-weighted (T1w) MRI volumes, extracted from cognitively normal subjects across the full age range.

📦 Dataset Structure

Each entry corresponds to a single 2D slice from a 3D MRI volume:

  • volume_id → Unique identifier for the subject/volume
  • slice_id → Index of the slice within the volume
  • age → Age of the participant at time of scan
  • sex → Biological sex of the participant ('M' / 'F')
  • image → 2D MRI slice

⚙️ Preprocessing

  • Only T1-weighted (T1w) volumes were used
  • Volumes were converted into 2D axial slices
  • Slices were normalized and resized to 256×256

⚠️ Caveat

Late slices in each volume have noise in them.

🚀 Usage

from datasets import load_dataset
import matplotlib.pyplot as plt

ds = load_dataset("chehablab/SALD_t1w", split="train")
sample = ds[314]
img = sample["image"]
plt.imshow(img, cmap="gray")
plt.title(f"Volume {sample['volume_id']} | Slice {sample['slice_id']} | Age {sample['age']} | Sex {sample['sex']}")
plt.axis("off")
plt.show()

📚 Citation

If you use this dataset, please acknowledge our lab Chehab Lab and cite the original SALD dataset:

@article {Wei177279,
    author = {Wei, Dongtao and Zhuang, Kaixiang and Chen, Qunlin and Yang, Wenjing and Liu, Wei and Wang, Kangcheng and Sun, Jiangzhou and Qiu, Jiang},
    title = {Structural and functional MRI from a cross-sectional Southwest University Adult lifespan Dataset (SALD)},
    year = {2018},
    doi = {10.1101/177279},
    publisher = {Cold Spring Harbor Laboratory},
    journal = {bioRxiv}
}

License

This work is licensed under a Creative Commons CC BY-NC-SA 4.0 License. CC BY-NC-SA 4.0

Chehab Lab @ 2026