bilalahmad176176/BrainAge-Golden-Preprocessed

BrainAge Golden预处理缓存数据集包含6,050个预处理后的脑部MRI张量，专为脑龄预测模型训练设计。每个.pt文件对应一个受试者，包含：归一化的T1加权脑部体积数据（形状128×144×112）、表格数据（86维，包括70个脑区体积、3维性别独热编码和13维站点独热编码）、年龄标量以及元数据字典。数据来源于12个公开神经影像数据集（如BCP、IXI、ABIDE等），年龄范围0-86岁。数据集已应用完整预处理流程：包括HD-BET颅骨剥离、N4偏置校正、MNI空间配准、Z-score强度归一化、哈佛-牛津图谱分割和体积测量。总大小约24GB，每个文件约4MB。用户可直接下载用于模型训练，跳过耗时的预处理步骤。

The BrainAge Golden Preprocessed Cache dataset contains 6,050 preprocessed brain MRI tensors ready for training a brain-age prediction model. Each .pt file (one per subject) includes: a Z-normalized T1w brain volume in MNI space (shape 128×144×112), tabular data (86 dimensions: 70 regional volumes with log1p/12 scaling, 3 one-hot sex encodings, and 13 one-hot site encodings), chronological age in years, and a meta dictionary with subject_id, site, sex, age, and split information. Data is sourced from 12 public neuroimaging datasets (e.g., BCP, IXI, ABIDE) with an age range of 0–86 years. The preprocessing pipeline includes HD-BET skull-stripping, N4 bias correction, affine registration to MNI152, Z-score intensity normalization, Harvard-Oxford atlas segmentation, volume measurement, and tensor packaging. Total size is ~24 GB (~4 MB per file). This cache allows users to skip the 40+ hour preprocessing step and jump directly to model training.