{"success":true,"data":{"author":"Liu Mengxing","bucketAccessType":"public","bucketListable":true,"bucketName":"openneuro.org","bucketPrefix":"ds008384","category":"Neuroscience","cloudProvider":"aws","createdAt":"2026-11-16T14:11:23.695Z","description":"7 Tesla fMRI from 14 healthy adults (10 female, ages 20 to 33) who performed a hierarchical decision-making task over four sessions on different days. On each trial a cloud of coloured moving dots (the cue) indicated which of two categorisation subtasks to apply to a composite face/scene image: face (male vs female) or scene (city vs landscape). Mappings between cue attributes (red, yellow, left, right motion) and subtasks defined four latent task contexts, two colour-based and two motion-based, which switched covertly every 10 to 20 trials and had to be inferred from feedback; subjects reported both their answer and the cue attribute used, so their believed context is known on every trial. Colour and motion coherence varied independently. Per session: five task runs (task-Quantum4S, 40 trials, 430 volumes) and a T1w. BIDS layout, 679 files, about 121 GB. Useful for latent state inference, prefrontal task-set representations and hierarchical reinforcement learning. Released under CC0 on OpenNeuro; cite Liu Mengxing and the dataset DOI.","displayTitle":"QUANTUM_4S: 7T fMRI of Context Inference in a Hierarchical Decision Task","doi":"10.18112/openneuro.ds008384.v1.0.1","fuseMountBrowsable":true,"id":"bde64c28-7236-52fe-a2d4-3e2f2643286d","license":"CC0-1.0","licenseEvidenceUrl":"https://openneuro.org/datasets/ds008384","objectCountEstimate":15,"probeStatus":"ok","referenceUrl":"https://doi.org/10.18112/openneuro.ds008384.v1.0.1","region":"us-east-1","schemaCard":{"accessHints":[{"language":"python-anonymous","snippet":"import boto3\nfrom botocore import UNSIGNED\nfrom botocore.client import Config\ns3 = boto3.client('s3', region_name='us-east-1', config=Config(signature_version=UNSIGNED))\ns3.list_objects_v2(Bucket='openneuro.org', Prefix='ds008384/', Delimiter='/')"},{"language":"python-local","snippet":"open('/mnt/archive/.datalad/config', 'rb')"}],"caveats":["Files are binary — use a domain-specific parser (e.g. metpy / pynexrad / pydicom) rather than text I/O."],"format":"binary","generatedAt":"2026-11-16T14:11:26.861Z","layout":{"example":"ds008384/.datalad/config","pattern":"{prefix}/{seg1}/{object}"},"mountPath":"/mnt/archive","sampleHeaders":[{"bytes":"[datalad \"dataset\"]\n\tid = 45c3d6f7-0fed-4da1-bd09-94e2eef5dab9\n","encoding":"utf8","key":"ds008384/.datalad/config","truncatedAt":512}],"sampleObjects":["ds008384/.datalad/config"],"sidecars":[{"bytes":"# QUANTUM_4S: 7T fMRI of context inference in a hierarchical decision task\n\n## Overview\nFourteen healthy adults (10 female, aged 20-33 years) performed a hierarchical decision-making task across four fMRI sessions on different days. On each trial, a cloud of colored moving dots (cue) indicated which of two categorization subtasks to perform on a composite face/scene image (probe): face (male vs. female) or scene (city vs. landscape). The mapping between cue attributes (red, yellow, left, right) and subtasks defined four latent task contexts, two based on color and two based on motion. The context changed covertly every 10-20 trials and had to be inferred from feedback. Subjects reported both the subtask answer and the cue attribute they used, so their believed context could be identified on every trial. Color and motion coherence of the cue were varied independently across trials.\n\n## Dataset structure\n- 14 subjects, 4 sessions each (`ses-1` to `ses-4`, numbered in chronological order)\n- per session: 5 runs of the task (`task-Quantum4S`, 40 trials and 430 volumes per run) and one T1-weighted anatomical image\n\n## Acquisition\nGE SIGNA 7T, Magnetic Resonance Research Facility, University of Iowa.\n- Functional: multiband EPI (factor 2), TR = 1712 ms, TE = 20.5 ms, flip angle 60 deg, FoV 220 mm, 2 x 2 x 2 mm voxels, 68 transverse slices\n- Anatomical: MPRAGE, TR = 3600 ms, TE = 3 ms, flip angle 8 deg, FoV 220 mm, 0.70 x 0.43 x 0.43 mm voxels, 220 sagittal slices\n\nAnatomical images are defaced.\n\n## Ethics\nThe study was approved by the University of Iowa Institutional Review Board (IRB No. 202207126). All participants gave written informed consent, including consent for data sharing.\n","encoding":"utf8","key":"ds008384/README","name":"README","size":1705,"truncated":false}],"sourceObjectKey":"ds008384/.datalad/config","starterCell":"with open('/mnt/archive/.datalad/config', 'rb') as fh:\n    head = fh.read(1024)\nprint(len(head), head[:64])","totalObjects":"15 top-level prefixes"},"summary":"7 Tesla fMRI from 14 healthy adults (10 female, ages 20 to 33) who performed a hierarchical decision-making task over four sessions on different days.","title":"QUANTUM_4S: 7T fMRI of Context Inference in a Hierarchical Decision Task","updatedAt":"2026-11-16T15:35:46.979Z","visibility":"public","websiteUrl":"https://openneuro.org/datasets/ds008384"}}