SENORA-MRI: A Multimodal Brain MRI Dataset with Metadata and Lesion Masks for Stroke and Epilepsy in Northern Nigeria

SENORA-MRI (Stroke and Epilepsy Neuroimaging Open Repository for Africa — Magnetic Resonance Imaging) is a multimodal brain MRI dataset with comprehensive clinical metadata and manual lesion segmentation from Medserve Kano Diagnostic Center in collaboration with Aminu Kano Teaching Hospital (AKTH), Kano, Northern Nigeria. This dataset addresses a critical gap in global neuroimaging representation by providing structured, open-access data from a population of over 100 million people previously absent from public repositories. It was developed in response to the MICCAI Open Data Initiative and supported by the MICCAI 2026 Open Dataset Micro-Grant. COHORT------103 participants (2024–2026)- 99 stroke cases (89 ischemic, 3 hemorrhagic)- 3 verified epilepsy cases with EEG correlation- Age: 52.7 ± 22.5 years (range 3–95)- Sex: 66 male / 37 female- Geographic coverage: Kano State (n=71), Jigawa (n=8), Katsina (n=6), Kaduna (n=4), Bauchi (n=2), and other Northern Nigerian states IMAGING-------Scanner: Siemens Magnetom Essenza 1.5T at Medserve Kano Diagnostic CenterEight sequences per participant in NIfTI format, organised in BIDS v1.8.0:- T1-weighted (T1w): TR=489 ms, TE=8.9 ms, slice=5 mm- T2-weighted (T2w): TR=3540 ms, TE=70 ms, slice=5 mm- FLAIR: TR=3980 ms, TE=78 ms, slice=5 mm- DWI: TR=4100 ms, TE=120 ms, slice=5.5 mm- T2*-weighted GRE: TR=7 ms, TE=2.9 ms, slice=8 mm- T1-weighted post-contrast (T1wCE): TR=829 ms, TE=8.9 ms, slice=5 mm- T2-weighted coronal (T2wCOR): TR=4470 ms, TE=82 ms, slice=5 mm- T2-weighted sagittal (T2wSAG): TR=4570 ms, TE=80 ms, slice=5 mm METADATA--------15 structured fields per participant (participants.tsv):- Demographics (age, sex, BMI)- Socioeconomic status (High / Middle / Low)- Education level (Primary / Secondary / Tertiary / Not recorded)- Stroke subtype (Ischemic / Hemorrhagic / Seizure)- Presentation status (Acute / Chronic / Unknown)- Handedness (Right / Left)- Quality control flags (PASS / PARTIAL / FAIL)- State of origin SEGMENTATION------------- 23 manual lesion masks produced by two consultant radiologists- Segmentation software: ITK-SNAP v3.8.0- Inter-rater reliability: Dice = 0.777 ± 0.144- Consensus masks generated by averaging and thresholding at 0.5 QUALITY CONTROL---------------Three-stage protocol:1. DICOM header validation (pydicom v2.3.0)2. Visual inspection for motion artefacts and sequence completeness3. BIDS Validator v1.9.0 compliance checkResults: 94 PASS, 3 PARTIAL, 6 FAIL (all retained with QC annotations) ETHICS------Approved by AKTH Health Research Ethics Committee (NHREC/28/01/2020/AKTH/EC/3867; AKTH/MAC/SUB/12A/P-3/VI/3967, dated 30th September 2024).HIPAA Safe Harbor de-identification applied (name removal, date shifting, master key destruction).Written informed consent obtained for research participation and open data release. DOWNLOAD--------The full image dataset is split into two ZIP archives for bandwidth efficiency:- SENORA-MRI_v1.0_part1.zip (~2.5 GB)- SENORA-MRI_v1.0_part2.zip (~2.5 GB) Download BOTH parts and extract them to the same folder to reconstruct the complete BIDS dataset. A metadata preview ZIP (SENORA-MRI_metadata_preview.zip) containing README.md, participants.tsv, and dataset_description.json is also provided for inspection without downloading the full imaging data. CITATION--------Zubair AA, Ibrahim U, Abbas RM, Mukhtar Dafi S, Muhammad Musa N, Anas A, Ismail II, Muhammad IH, Yakubu M, Wogu E. SENORA-MRI: A Multimodal Brain MRI Dataset with Metadata and Lesion Masks for Stroke and Epilepsy in Northern Nigeria. Zenodo, 2026. DOI: 10.5281/zenodo.20757772 LICENSE-------CC BY 4.0 FUNDING & ACKNOWLEDGEMENT-------MICCAI 2026 Open Dataset Micro-GrantAfrican Institute for Research Advancement and Innovation (AIRA Africa)CAMERA AfricaABDN DOI---10.5281/zenodo.20757772