{"success":true,"database":"eegdash","data":{"_id":"69a33a3b897a7725c66f3edf","dataset_id":"ds006136","associated_paper_doi":null,"authors":["Vladimir Omelyusik","Tyler S. Davis","Satish S. Nair","Behrad Noudoost","Patrick Hackett","Elliot H. Smith","Shervin Rahimpour","John D. Rolston","Bornali Kundu"],"bids_version":"1.10.0","contact_info":["Vladimir Omelyusik","Bornali Kundu"],"contributing_labs":null,"data_processed":false,"dataset_doi":"doi:10.18112/openneuro.ds006136.v1.0.1","datatypes":["ieeg"],"demographics":{"subjects_count":13,"ages":[],"age_min":null,"age_max":null,"age_mean":null,"species":null,"sex_distribution":null,"handedness_distribution":null},"experimental_modalities":null,"external_links":{"source_url":"https://openneuro.org/datasets/ds006136","osf_url":null,"github_url":null,"paper_url":null},"funding":["K12 grant from the Neurosurgery Career Development Award, parent award from the NINDS as well as a Neurosurgery Research Education Fund award","NIH/NINDS career development award to JDR (K23 NS114178)","NIMH MH122023 and NSF-OAC 2417875 (SSN)"],"ingestion_fingerprint":"7ef58303e0652dae780864e62769057377903ab6910e89152243ef549a338f8a","license":"CC0","n_contributing_labs":null,"name":"OWM-Dataset","readme":"# OWM-Dataset\n## Description\nThe dataset contains processed intracranial EEG recordings from frontal (LMFG, RMFG) and temporal (LMTG, RMTG) areas of 13 subjects (epilepsy patients) while they performed a load-3 object working memory task. Please see the associated publication (Paper): https://doi.org/10.1016/j.neuroimage.2026.121718\n## Data structure\n### Included trials\nThe dataset includes trials which were used for the final analyses (i.e., after artifact rejection; see the Methods section of the Paper for a full description of preprocessing procedures). Note that since some artifact rejection procedures were performed at the single-trial level, trial indexes are not matched across channels even for the same subject (i.e., trial 1 of channel 1 may not correspond to trial 1 of channel 2) and have to be read separately. The sourcedata/ folder contains per-subject trial indexes for trial matching.\n### Trial structure\nEach trial is 6498 ms long (1000 ms of fixation, 1500 ms of encoding and 3998 ms of delay).\n### Storage format\nTo comply with the .edf format, trials for every channel were concatenated into a single one-dimensional array. Due to a different number of trials across channels, each array was padded with 0s on the right, ensuring the same data length for all channels within a subject. The total number of concatenated trials per channel and the padding length are recorded in the \"_channels.tsv\" file for each subject.\n### Performance\nThe sourcedata/ folder contains performance results for every subject. Rows of each table correspond to trials (the order matches LFP recordings). Columns represent whether the subject selected the presented stimuli during the search period (0 = no, 1 = yes).\n## Reading the data and replicating the results\nThe Paper repository (https://github.com/V-Marco/FT-bursting-WM) includes a Python function for reading the data, performing trial matching, appending performance information, and representing the recordings as a 2D table. The repository also includes examples on replicating the main figures.","recording_modality":["ieeg"],"senior_author":"Bornali Kundu","sessions":["01","02"],"size_bytes":299778229,"source":"openneuro","storage":{"backend":"s3","base":"s3://openneuro.org/ds006136","raw_key":"dataset_description.json","dep_keys":["CHANGES","README"]},"study_design":null,"study_domain":null,"tasks":["OWM"],"timestamps":{"digested_at":"2026-04-22T12:29:09.935326+00:00","dataset_created_at":"2025-04-16T16:28:47.452Z","dataset_modified_at":"2026-01-29T22:08:24.000Z"},"total_files":14,"computed_title":"OWM-Dataset","nchans_counts":[{"val":9,"count":2},{"val":8,"count":2},{"val":7,"count":2},{"val":12,"count":2},{"val":14,"count":1},{"val":6,"count":1},{"val":17,"count":1},{"val":5,"count":1},{"val":18,"count":1},{"val":11,"count":1}],"sfreq_counts":[{"val":1000.0,"count":14}],"stats_computed_at":"2026-04-22T23:16:00.312453+00:00","total_duration_s":90986.0,"tagger_meta":{"config_hash":"3557b68bca409f28","metadata_hash":"e258ece040fd8444","model":"openai/gpt-5.2","tagged_at":"2026-04-07T09:32:40.872789+00:00"},"tags":{"pathology":["Epilepsy"],"modality":["Visual"],"type":["Memory"],"confidence":{"pathology":0.8,"modality":0.6,"type":0.8},"reasoning":{"few_shot_analysis":"Most similar few-shot example by population is the pediatric epilepsy dataset (Pathology=Epilepsy), which shows the convention that when metadata explicitly recruits/uses epilepsy patients, Pathology should be labeled Epilepsy regardless of task. By cognitive construct, the digit span dataset (Type=Memory) illustrates that explicit mention of a working-memory paradigm maps to Type=Memory (even when other conditions like rest are included). For Modality, few-shot conventions emphasize labeling the stimulus channel (e.g., visual discrimination → Visual; auditory digit presentation → Auditory). Here the task is 'object working memory', which most plausibly implies visually presented objects, guiding Modality toward Visual over Unknown.","metadata_analysis":"Key population/task facts from README:\n1) Population: \"13 subjects (epilepsy patients)\".\n2) Task/construct: \"performed a load-3 object working memory task\".\nAdditional task-structure context: \"1000 ms of fixation, 1500 ms of encoding and 3998 ms of delay\" and performance described as \"whether the subject selected the presented stimuli during the search period\".\nThese support a working-memory paradigm with presented stimuli, but the sensory modality (e.g., 'visual', 'pictures') is not explicitly stated in the provided metadata.","paper_abstract_analysis":"No useful paper information. (Only a DOI link is provided; no abstract text is included in the metadata payload.)","evidence_alignment_check":"Pathology:\n- Metadata says: \"13 subjects (epilepsy patients)\".\n- Few-shot suggests: Epilepsy should be used when epilepsy patients are the recruited/recorded population (see epilepsy HFO example).\n- Alignment: ALIGN.\n\nModality:\n- Metadata says: \"object working memory task\" and mentions \"fixation\", \"encoding\", and \"presented stimuli\" during a \"search period\" but does not explicitly say 'visual' or 'auditory'.\n- Few-shot suggests: Modality should follow stimulus channel; object-based WM tasks are typically visual, but this is an inference.\n- Alignment: PARTIAL (few-shot convention clear, but stimulus modality not explicitly stated in metadata).\n\nType:\n- Metadata says: \"object working memory task\" and specifies a memory-trial structure (fixation/encoding/delay).\n- Few-shot suggests: Working memory paradigms map to Type=Memory (see digit span example).\n- Alignment: ALIGN.","decision_summary":"Top-2 comparative selection:\n\n1) Pathology\n- Candidate A: Epilepsy\n  Evidence: \"13 subjects (epilepsy patients)\".\n- Candidate B: Healthy\n  Evidence: none; only implied by absence of other diagnoses, but contradicted by explicit epilepsy statement.\n- Decision: Epilepsy (explicit recruitment/clinical context).\n- Confidence justification: 1 strong explicit quote + strong few-shot alignment.\n\n2) Modality\n- Candidate A: Visual\n  Evidence: \"object working memory task\" + \"fixation\" + \"encoding\" + \"presented stimuli\" and a \"search period\" (strongly typical of visually presented objects, but not explicitly stated).\n- Candidate B: Unknown\n  Evidence: no explicit modality keywords (e.g., 'visual', 'auditory') in provided metadata.\n- Decision: Visual (best-supported by paradigm semantics, but inferential).\n- Confidence justification: contextual inference only; no direct modality quote.\n\n3) Type\n- Candidate A: Memory\n  Evidence: explicit \"object working memory task\"; trial includes \"encoding\" and \"delay\".\n- Candidate B: Attention\n  Evidence: could involve attention demands, but working memory is explicitly named as the purpose.\n- Decision: Memory.\n- Confidence justification: 2+ explicit task/structure quotes + clear few-shot analog for WM→Memory."}},"canonical_name":null,"name_confidence":0.72,"name_meta":{"suggested_at":"2026-04-14T10:18:35.343Z","model":"openai/gpt-5.2 + openai/gpt-5.4-mini + deterministic_fallback"},"name_source":"author_year","author_year":"Omelyusik2025"}}