Longitudinal data gathering

[plbenveniste]

In this issue, I am detailing my investigations in gathering data for longitudinal segmentation of MS lesions.

Datasets of interests:

• bavaria-quebec-spine-ms-unstitched
• canproco
• ms-basel-2018
• ms-dresden-mp2rage-2025
• ms-karolinska-2020
• umass-ms-siemens-espree1.5
• umass-ms-ge-pioneer3
• umass-ms-ge-excite1.5
• umass-ms-ge-hdxt1.5
• ms-nmo-beijing

Also tagging this comment about potential data from UCSF: neuropoly/idea-projects#17 (comment)

[plbenveniste]

First aggregation of data:

2025-10-20 07:16:17.592 | INFO     | __main__:main:70 - Command line: python longitudinal_segmentation/dataset_aggregation/build_msd.py -pd /home/plbenveniste/net/longitudinal_ms/data/ -po /home/plbenveniste/net/longitudinal_ms/msd_data/ --canproco-exclude /home/plbenveniste/net/longitudinal_ms/exclude_file/canproco/exclude.yml --exclude dataset_aggregation/exclude.yml
2025-10-20 07:16:20.003 | INFO     | __main__:main:118 - Number of subjects in Bavaria: 331
2025-10-20 07:16:24.664 | INFO     | __main__:main:143 - Number of subjects in CanProCo: 346
2025-10-20 07:16:24.709 | INFO     | __main__:main:168 - Number of subjects in Basel: 1
2025-10-20 07:16:24.964 | INFO     | __main__:main:196 - Number of subjects in Karo: 38
2025-10-20 07:16:25.096 | INFO     | __main__:main:225 - Number of subjects in Umass1: 2
2025-10-20 07:16:28.577 | INFO     | __main__:main:255 - Number of subjects in Umass2: 144
2025-10-20 07:16:28.775 | INFO     | __main__:main:285 - Number of subjects in Umass3: 10
2025-10-20 07:16:29.012 | INFO     | __main__:main:314 - Number of subjects in Umass4: 7
2025-10-20 07:16:30.372 | INFO     | __main__:main:342 - Number of subjects in Beijing: 13
2025-10-20 07:16:30.372 | INFO     | __main__:main:346 - Total number of subjects before exclusion: 887
2025-10-20 07:16:30.372 | INFO     | __main__:main:351 - Total number of subjects after exclusion: 887
2025-10-20 07:16:30.389 | INFO     | __main__:main:367 - JSON file saved at /home/plbenveniste/net/longitudinal_ms/msd_data/dataset_2025-10-20.json

More data should be aggregated with the addition of UCSF data from Roland Henry and new timepoints from canproco