BANDA

BANDA is the adolescent anxiety and depression study of the Human Connectome Project, funded under the Connectomes Related to Human Disease program. Researchers at MIT, Massachusetts General Hospital, McLean Hospital and Boston University scanned adolescents with and without current anxious or depressive disorders using an HCP-style protocol, and followed their symptoms for a year afterwards. The aims were to give researchers a shared sample for adolescent-specific atlases and tools and to test brain markers that predict how symptoms develop, persist or remit. The data are held in NIMH Data Archive collection 3037. Release 1.0 appeared in February 2023 (NDA study 1920) and Release 1.1 in March 2024; the second release only corrected and extended the clinical and behavioral tables.

Composition

The release covers 215 adolescents, 152 of whom had a current DSM-5 anxious and/or depressive disorder. The others include typical adolescents recruited as controls, who had no current or lifetime anxiety or depressive disorder. Age at imaging was 14 to 17. Unprocessed imaging is shared for 207 subjects and minimally preprocessed imaging for 203. Per the release completeness file, 202 subjects have T1w, T2w, functional and diffusion data. Clinical self-report measures were repeated about 6 and 12 months after the scan.

Acquisition

All imaging took place at the Martinos Center on one Siemens Prisma 3T scanner with a 64-channel head coil, in a session of about 1 h 36 min. The protocol has 0.8 mm T1w MPRAGE and T2w images with volumetric navigators for motion correction, multiband diffusion MRI at 1.5 mm with two shells (b=1500 and 3000), four resting-state fMRI runs at 2 mm and TR 800 ms, and three fMRI tasks on emotion interference, emotion processing and incentive processing. Diffusion and fMRI are harmonized with HCP-Development. Three protocol versions were used during the study.

Annotations

There are no image labels. Subjects carry K-SADS diagnoses, suicide risk ratings, family history, self- and parent-report scales, NIH Toolbox and Penn cognitive tests, IQ, puberty stage and demographics.

Known limitations

All subjects come from one scanner and one metropolitan area. The sad-face condition of the emotion processing task was added after the first 17 subjects. Sex and age distributions of the released sample are not given in the release documentation. Eye-tracking data were collected but are not shared.