{"success":true,"database":"eegdash","data":{"_id":"6953f4239276ef1ee07a32be","dataset_id":"ds003104","associated_paper_doi":null,"authors":["Lauri Parkkonen","Stefan Appelhoff","Alexandre Gramfort","Mainak Jas","Richard Höchenberger"],"bids_version":"1.4.0","contact_info":["Richard Höchenberger","Alexandre Gramfort","Daniel McCloy"],"contributing_labs":null,"data_processed":true,"dataset_doi":"doi:10.18112/openneuro.ds003104.v1.0.1","datatypes":["meg"],"demographics":{"subjects_count":1,"ages":[],"age_min":null,"age_max":null,"age_mean":null,"species":null,"sex_distribution":{"m":1},"handedness_distribution":null},"experimental_modalities":null,"external_links":{"source_url":"https://openneuro.org/datasets/ds003104","osf_url":null,"github_url":null,"paper_url":null},"funding":[],"ingestion_fingerprint":"f2f66097fbf2241996becb8b8906e264e45c284944fb4acb86557110c460f770","license":"CC0","n_contributing_labs":null,"name":"MNE-somato-data-bids (anonymized)","readme":"MNE-somato-data-bids\n====================\nThis dataset contains the MNE-somato-data in BIDS format.\nThe conversion can be reproduced through the Python script stored in the\n`/code` directory of this dataset. See the README in that directory.\nThe `/derivatives` directory contains the outputs of running the FreeSurfer\npipeline `recon-all` on the MRI data with no additional commandline options\n(only defaults were used):\n$ recon-all -i sub-01_T1w.nii.gz -s 01 -all\nAfter the `recon-all` call, there were further FreeSurfer calls from the MNE\nAPI:\n$ mne make_scalp_surfaces -s 01 --force\n$ mne watershed_bem -s 01\nThe derivatives also contain the forward model `*-fwd.fif`, which was produced\nusing the source space definition, a `*-trans.fif` file, and the boundary\nelement model (=conductor model) that lives in\n`freesurfer/subjects/01/bem/*-bem-sol.fif`.\nThe `*-trans.fif` file is not saved, but can be recovered from the anatomical\nlandmarks in the `sub-01/anat/T1w.json` file and MNE-BIDS' function\n`get_head_mri_transform`.\nSee: https://github.com/mne-tools/mne-bids for more information.\nNotes on FreeSurfer\n===================\nthe FreeSurfer pipeline `recon-all` was run new for the sake of converting the\nsomato data to BIDS format. This needed to be done to change the \"somato\"\nsubject name to the BIDS subject label \"01\". Note, that this is NOT \"sub-01\",\nbecause in BIDS, the \"sub-\" is just a prefix, whereas the \"01\" is the subject\nlabel.","recording_modality":["meg"],"senior_author":"Richard Höchenberger","sessions":[],"size_bytes":349944649,"source":"openneuro","study_design":null,"study_domain":null,"tasks":["somato"],"timestamps":{"digested_at":"2026-04-22T12:25:31.135684+00:00","dataset_created_at":"2020-08-31T07:20:15.139Z","dataset_modified_at":"2026-03-13T18:15:18.000Z"},"total_files":1,"storage":{"backend":"s3","base":"s3://openneuro.org/ds003104","raw_key":"dataset_description.json","dep_keys":["CHANGES","README","participants.json","participants.tsv"]},"tagger_meta":{"config_hash":"3557b68bca409f28","metadata_hash":"73403ae376f93504","model":"openai/gpt-5.2","tagged_at":"2026-04-07T09:32:40.872789+00:00"},"tags":{"pathology":["Healthy"],"modality":["Tactile"],"type":["Perception"],"confidence":{"pathology":0.6,"modality":0.8,"type":0.65},"reasoning":{"few_shot_analysis":"Most similar few-shot convention: the \"Braille letters - EEG\" example is labeled with Modality=Tactile when the stimulus is delivered to fingers (somatosensory input). This guides mapping the present dataset’s \"somato\" task (somatosensory) to Modality=Tactile. For Type, the few-shot tactile example is \"Learning\" because it involves letter/category learning/decoding; here there is no learning framing, so the closest convention is to treat basic sensory stimulation/evoked responses as Type=Perception.","metadata_analysis":"Key metadata facts:\n1) Dataset identity is explicitly somatosensory-oriented: \"This dataset contains the MNE-somato-data in BIDS format.\" \n2) The only task is named somatosensory: \"tasks\": [\"somato\"].\n3) Single-participant, no clinical recruitment described: \"Subjects: 1; Sex: {'m': 1}\".\nNo metadata text mentions a diagnosis, patient group, intervention, or sleep/resting protocol; stimulus specifics are not described beyond the dataset/task naming.","paper_abstract_analysis":"No useful paper information.","evidence_alignment_check":"Pathology:\n- Metadata says: no diagnosis/patient group is mentioned (e.g., only \"Subjects: 1\").\n- Few-shot pattern suggests: absent clinical terms → label as Healthy (normative cohort).\n- Alignment: ALIGN (no conflict).\n\nModality:\n- Metadata says: \"MNE-somato-data\" and task \"somato\".\n- Few-shot pattern suggests: somatosensory/finger stimulation paradigms map to Tactile (as in the Braille tactile example).\n- Alignment: ALIGN (metadata implies somatosensory; few-shot supports Tactile mapping).\n\nType:\n- Metadata says: only indicates a somatosensory task (\"somato\"), without learning/decision/motor/rest/sleep framing.\n- Few-shot pattern suggests: sensory stimulation/discrimination without higher-order framing maps to Perception.\n- Alignment: ALIGN (no conflict).","decision_summary":"Top-2 comparative selection:\n\nPathology candidates:\n1) Healthy — Evidence: no clinical recruitment described; only \"Subjects: 1\" and a generic MNE sample-style dataset name (\"MNE-somato-data\").\n2) Unknown — Evidence: metadata never explicitly says \"healthy\".\nHead-to-head: Healthy wins because datasets without any pathology language are conventionally treated as normative/healthy in the catalog.\nConfidence basis: inference-only (no explicit quote of health) → moderate.\n\nModality candidates:\n1) Tactile — Evidence: dataset/task naming \"MNE-somato-data\" and \"tasks\": [\"somato\"]; few-shot tactile example shows somatosensory input → Tactile.\n2) Unknown/Other — Evidence: stimulus delivery not explicitly described (no mention of electrical nerve stimulation, touch, etc.).\nHead-to-head: Tactile wins because \"somato\" strongly implies somatosensory stimulation.\nConfidence basis: 2 metadata snippets implying somatosensory + strong few-shot analog → higher.\n\nType candidates:\n1) Perception — Evidence: task is somatosensory (\"somato\"), consistent with studying sensory-evoked/perceptual processing.\n2) Other — Evidence: task goal not described beyond name.\nHead-to-head: Perception wins as the most standard cognitive construct for somatosensory stimulation datasets.\nConfidence basis: mostly inference from task naming + tactile few-shot convention → moderate."}},"nemar_citation_count":0,"computed_title":"MNE-somato-data-bids (anonymized)","nchans_counts":[{"val":316,"count":1}],"sfreq_counts":[{"val":300.3074951171875,"count":1}],"stats_computed_at":"2026-04-22T23:16:00.221930+00:00","total_duration_s":897.0771771609423,"canonical_name":null,"name_confidence":0.63,"name_meta":{"suggested_at":"2026-04-14T10:18:35.342Z","model":"openai/gpt-5.2 + openai/gpt-5.4-mini + deterministic_fallback"},"name_source":"canonical","author_year":"Parkkonen2020"}}