NEBULA101 (ds005613)

Brain MRI and deep behavioural phenotyping from a Geneva study on why people differ in their ability to learn and use languages. Healthy young adults, most of them speaking several languages, completed questionnaires on language background, motivation, reading and music, plus a long battery of linguistic and general cognitive tasks spread over two online and one in-person session, and then a single imaging session. The data are on OpenNeuro in BIDS format under CC0, and a data descriptor in Scientific Data documents them.

Composition

101 adults, 68 women and 33 men, aged 18 to 38 with a mean of 23.35 years. The study recruited 104 people; one withdrew consent to data sharing and two could not be scanned because of claustrophobia. People with neurological or psychiatric diagnoses, serious chronic illness, professional music or interpreting training, diagnosed dyslexia, or more than 10 well-known languages were not included. Questionnaire scores, task scores and language history data come as phenotype and derivative tables. There is no patient group.

Acquisition

All scans come from a Siemens 3 T Magnetom Prisma with a 64-channel head coil at Campus Biotech in Geneva. Each participant has a defaced 1 mm MPRAGE T1-weighted scan, one multi-shell diffusion run at 1.5 mm with b-values up to 2850 s/mm², one resting-state BOLD run and three runs of an auditory language localiser in which participants listened to intact and degraded passages of Alice in Wonderland in their first and second language. BOLD runs use a repetition time of 2 s and multiband factor 3. Phase-difference field maps accompany the functional and diffusion scans. Data were collected between July 2022 and June 2023.

Annotations

There are no image labels. Event files for the localiser runs, the localiser stimuli and per-participant behavioural and questionnaire scores are the non-imaging data.

Known limitations

The sample is young, mostly female and highly educated, and recruited from one French-speaking region. Three further fMRI tasks and one anatomical scan acquired in the same session belong to other projects and are not part of this release. Some behavioural measures are missing for individual participants, as logged in the dataset's validation folder.