Find your database
A single home for the curated databases powering Net4Brain, from brain cancer imaging catalogs to multi-modal neuroscience collections.
Database Categories
Our databases are organized by data type and research focus
Imaging Data
MRI, fMRI, PET scans
Non-Imaging Data
MRI, fMRI, PET scans
Clinical Data
Patient records, outcomes
Genomics
Gene expression, sequences
Browse databases
70+ Imaging datasets • 70 Active databases
Brain Cancer Imaging Resources
Imaging Dataset Catalog with cards, table, and analytics views. Filterable by base collection, access level, tumor type, and categories.
Curated Databases Catalog
Image-Based and Non-Image-Based data collections with detailed entries including DOI, modalities, and linked publications.
Featured datasets
Yale-Brain-Mets-Longitudinal
Information from the public dataset page confirms Yale-Brain-Mets-Longitudinal as a public TCIA MRI+demographic release (DOI 10.7937/3YAT-E768; status complete; updated 2025/05/30). Data-access section shows downloadable NIfTI radiology package and demographic/scanner spreadsheet (CC BY 4.0), with version note stating June-2025 clinical-sheet deduplication update. Source counts differ by release cut: TCIA page reports 1,430 subjects and 11,892 studies, while paper text reports 11,884 studies from 1,430 patients; both are retained as versioned provenance. Supplementary metadata tables (Clinical_data; Acquisition_data; image_acquisition_parameters) support acquisition ranges including slice_thickness 0.6-6.0 mm and spacing_between_slices 0.5-7.5 mm.
brain metastasis
FOMO60K
Information from the public Hugging Face dataset page (FOMO-MRI/FOMO60K) confirms a public multi-source MRI aggregation (subset of FOMO300K), with download gated by contact-sharing acceptance and collection-level CC BY-NC-SA 4.0 framing. Page + paper agree on FOMO60K counts: 60,529 scans, 13,900 MRI sessions, 11,187 subjects (16 source datasets). Paper-specific preprocessing details for FOMO60K include RAS reorientation, within-session affine co-registration to highest-resolution image (FreeSurfer mri_coreg), skull stripping/defacing safeguards, and DWI handling rules. Metadata/provenance files explicitly listed are participants table (age/sex/handedness/group), scan-to-source mapping table (scan-to-source mapping), and MRI acquisition table (acquisition information). Paper also reports demographic/acquisition summary statistics (female 53%, mean age 35.2 years, handedness/right-handed predominance, vendor/field-strength/2D-vs-3D/sequence distributions) at corpus session level; these are retained as evidence context and may not be strictly FOMO60K-only aggregates. Dependency chain: FOMO60K is a curated/processed subset of FOMO300K.
other/unspecified
FOMO300K
Information from the public dataset page and primary publication confirms FOMO300K as a large-scale heterogeneous brain MRI aggregation for self-supervised learning. The paper reports 318,877 scans from 82,678 MRI sessions and 59,969 subjects aggregated from 920 public sources; Hugging Face card presents dataset-level download and citation context for FOMO300K. Minimal preprocessing is described (RAS orientation, 4D-to-3D conversion, scan filtering, and modality-specific handling such as DWI shell/b0 treatment), with broad multi-sequence and multi-pathology heterogeneity retained. Dependency chain: FOMO300K is a larger multi-source aggregation that includes almost all OpenMind (anon-res) scans, and FOMO60K is a subset of FOMO300K.
other/unspecified
OpenMind
Information from the public dataset page and published metadata tables confirms OpenMind as a large-scale anonymized 3D head-and-neck MRI dataset release for self-supervised learning, hosted under AnonRes/OpenMind. The dataset card reports 114k MRI images from 30 scanner centers and 23 modality/technique families. Metadata (published metadata export) lists 113,921 image entries with per-image demographic/acquisition fields (partial coverage for age/sex/handedness/health status) and anonymization artifacts (anat_mask, anon_mask). Dependency chain: OpenMind (anon-res) is an upstream source collection from which almost all scans are included in FOMO300K.
other/unspecified
LUMIERE
Information from the public dataset page and primary publication confirms LUMIERE as a single-center longitudinal GBM MRI dataset with expert RANO ratings and rationale, 91 patients, 638 study dates, and 2487 MRI images. Data include CT1/T1/T2/FLAIR, automated segmentations (DeepBraTumIA and HD-GLIO-AUTO), radiomics/CoLlAGe features, MRI acquisition metadata, and pathology subset fields (MGMT/IDH1) with survival context.
glioma
Yale-Lung-Mets-Path
Information from the public dataset page confirms TCIA Brain-Mets-Lung-MRI-Path-Segs snapshot: subjects=103, data types include Histopathology/WSI + MR + Segmentation + Demographic/Diagnosis/Follow-Up/Molecular Test, status complete, updated 2025/11/13, DOI 10.7937/k0sm-y874. Data-access section indicates Version 2 update added MRI intensity-normalization preprocessing for radiomics while imaging files were unchanged. Review of supplementary tables (Clinical_Data; Image_Data; Radiomics) confirms 111 case rows with 103 unique patients (8 recurrence pairs), acquisition heterogeneity (0.7T/1.5T/3T; Siemens/GE/Philips/Hitachi), and cohort-level demographic table consistency with paper.
brain metastasis
Cite Net4Brain
Using Net4Brain in your research? View our citation guidelines and help us track the impact of this resource.
Contribute a database
Help expand our curated collection. Submit your dataset for review and inclusion in the Net4Brain catalog.