A paired dataset of multi-modal MRI at 3 Tesla and 7 Tesla with manual hippocampal subfield segmentations

Citation Author(s):
Shuyu
Li
Lei
Chu
Baoqiang
Ma
Xiaoxi
Dong
Submitted by:
Shuyu Li
Last updated:
Sat, 04/20/2024 - 00:52
DOI:
10.21227/wcge-v904
Data Format:
.IMA
Research Article Link:
Syn_SegNet: A joint deep neural network for ultrahigh-field 7T MRI synthesis and hippocampal subfield segmentation in routine 3T
License:
Creative Commons Attribution
779 Views
Categories:
Neuroscience
Keywords:
hippocampus; hippocampal subfield; 7T MRI;
0
0 ratings - Please login to submit your rating.
ACCESS DATASET CITE SHARE/EMBED

Abstract 

The paired 3T-7T MRI dataset includes T1- and T2-weighted whole-brain MRI of each participant performed on an investigational 7T MAGNETOM MR system with a 32-channel receiver and a 1-channel transmit birdcage head coil (Nova Medical, Wilmington, MA, USA) and a 3T MAGNETOM Prisma MR system equipped with a 64-channel eachphased-array head coil (Siemens Healthineers, Erlangen, Germany) on the same day at Beijing MRI Center for Brain Research (BMCBR). Twenty volunteers aged 18-25 years (10 males and 10 females) recruited from a university were enrolled.
The subiculum (SUB), cornu ammonis (CA)2, CA1, CA4-dentage gyrus (DG), entorhinal cortex (ERC), CA3 and Tail of hippocampus were manually delineated using ITK-SNAP (www.itksnap.org) based on a previously published protocol (Wisse et al., 2012)[1] for the left and right hippocampus separately on every slice of the individual subjects’ ultrahigh resolution 7T T2-weighted images on the coronal plane.

[1]. Wisse, L.E.M., et al., Subfields of the hippocampal formation at 7T MRI: In vivo volumetric assessment. NeuroImage, 2012. 61(4): p. 1043-1049.(https://www.sciencedirect.com/science/article/abs/pii/S1053811912002960)

Instructions: 

The data format is *.IMA. This dataset provides 3T T1, 3T T2, 7T T1, 7T T2 MRI, and hippocampal subfield labels for each subject.
Two experienced medical experts outlined the hippocampal subfields of the same 4 subjects twice (for one month interval) to ensure both had mastered the segmentation protocol by calculating inter-rater and intra-rater reliability which were evaluated by Dice Similarity Indices (DSI) for the left, right and both hemispheres combined. The two experts jointly outlined the final manual segmentation of HS for 20 subjects as the gold standard for automatic segmentation. Any disputed delineation was discussed to modification and agreed upon.

The source code of Syn_SegNet has been made publicly available at (https://github.com/lixw777/Syn_SegNet).

If you have any questions, please contact: shuyuli@bnu.edu.cn

Dataset Files

LOGIN TO ACCESS DATASET FILES

 

QUESTIONS?