QNLD Negative BOLD

The Quantitative Neuroimaging Laboratory at Weill Cornell Medicine in New York built this cohort to study the negative BOLD response: the drop in fMRI signal that many brain regions show while a person performs a task. Cognitively normal young and older adults each had structural MRI, twelve event-related task fMRI runs, resting-state fMRI and three PET scans, plus a neuropsychological battery. The aim is to describe how task-evoked deactivation behaves, how it relates to functional networks, and how it changes with age and with amyloid and tau pathology. Data are shared in BIDS on OpenNeuro (ds006148).

Composition

356 people consented (97 aged 20 to 40 and 259 aged 60 to 80), and 259 completed at least one scan: 62 young and 197 older participants. The release holds 4688 MRI and fMRI scans and 719 PET scans (251 FDG, 232 florbetaben, 236 MK-6240). 189 participants have every imaging modality. By visual reading, 44 of the older participants are amyloid positive and 63 tau positive, all cognitively unimpaired. Each participant has up to three sessions. The task fMRI covers four cognitive domains (fluid reasoning, processing speed, episodic memory, crystallized knowledge) with three tasks each. The snapshot also includes cognitive scores and derived data such as preprocessed resting-state fMRI and tau PET in native space.

Acquisition

All MRI was acquired on one 3 T Siemens Magnetom Prisma with a 64-channel head coil between October 2020 and June 2024: 0.5 mm MPRAGE, FLAIR, T2-weighted, multiband EPI fMRI (2 mm, TR 1008 ms, ten-minute runs), two diffusion runs with opposite phase encoding, pseudo-continuous ASL and multi-echo gradient echo reconstructed into QSM and SWI. PET was acquired on a Siemens Biograph mCT. MK-6240 was acquired with two protocols, one of them a split 0 to 60 and 90 to 120 minute scan used for 120 participants.

Annotations

Two neuroradiologists read the florbetaben and MK-6240 scans as positive or negative, globally and per lobe.

Known limitations

Age, sex, education and race at the individual level are, per the data descriptor, controlled access under a data use agreement with Weill Cornell Medicine. Some participants miss modalities or tasks, and the counts above refer to snapshot 1.0.5 described in the paper. Participants were recruited near one New York imaging center, and most are White.