{"success":true,"database":"eegdash","data":{"_id":"6953f4249276ef1ee07a33ae","dataset_id":"ds005028","associated_paper_doi":null,"authors":["Nand Chandravadia","Shrita Pendekanti","Dustin Roberts","Robert Tran","Saarang Panchavati","Corey Arnold","Nader Pouratian","William Speier"],"bids_version":"1.8.0","contact_info":["Bill Speier"],"contributing_labs":null,"data_processed":false,"dataset_doi":"doi:10.18112/openneuro.ds005028.v1.0.0","datatypes":["eeg"],"demographics":{"subjects_count":11,"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/ds005028","osf_url":null,"github_url":null,"paper_url":null},"funding":["National Science Foundation Graduate Research Fellowship Program under grant no. DGE-2034835"],"ingestion_fingerprint":"1b08a2419f8d2a65e2e8a2d354efe876c436f781d5386474779d671f38dd39dd","license":"CC0","n_contributing_labs":null,"name":"Comparing P300 Flashing paradigms in online typing with language models","readme":"This dataset was created using BCI2000. The goal of this study was to explore the online typing performance of the P300 speller using language models and various flashing paradigms. For more information see Chandravadia et al. (https://www.medrxiv.org/content/10.1101/2022.06.24.22276882v1).\nIf you reference this dataset in your publications, please acknowledge its authors.\nThis dataset is made available under CC0.\nNote: subject 5 was not included in the analysis because the testing stage did not include all three flashing paradigms.","recording_modality":["eeg"],"senior_author":"William Speier","sessions":["test","train"],"size_bytes":442568172,"source":"openneuro","study_design":null,"study_domain":null,"tasks":["cb","comb","rc"],"timestamps":{"digested_at":"2026-04-22T12:27:13.248573+00:00","dataset_created_at":"2024-03-11T23:10:26.788Z","dataset_modified_at":"2024-03-11T23:12:16.000Z"},"total_files":105,"storage":{"backend":"s3","base":"s3://openneuro.org/ds005028","raw_key":"dataset_description.json","dep_keys":["CHANGES","README"]},"nemar_citation_count":0,"computed_title":"Comparing P300 Flashing paradigms in online typing with language models","nchans_counts":[],"sfreq_counts":[],"stats_computed_at":"2026-04-22T23:16:00.308850+00:00","tags":{"pathology":["Unknown"],"modality":["Visual"],"type":["Attention"],"confidence":{"pathology":0.4,"modality":0.7,"type":0.7},"reasoning":{"few_shot_analysis":"Most similar few-shot conventions are the oddball-style datasets, where target detection elicits a P300 component. For example, the few-shot 'Cross-modal Oddball Task' is labeled with Type='Clinical/Intervention' primarily because it explicitly recruits Parkinson’s disease patients; however, it demonstrates the convention that oddball/P300 paradigms map to an attention/cognitive-control construct rather than motor output. Also, the BCI2000-based 'EEG Motor Movement/Imagery Dataset' shows that the platform (BCI2000) does not determine Type—task goal does. For the present dataset, the paradigm is a P300 speller with flashing stimuli, which aligns more with an attention/oddball ERP framing than with motor or learning.","metadata_analysis":"Key metadata indicating a P300 speller with flashing stimuli and online typing: (1) README: \"The goal of this study was to explore the online typing performance of the P300 speller using language models and various flashing paradigms.\" (2) README: \"This dataset was created using BCI2000.\" (3) Title: \"Comparing P300 Flashing paradigms in online typing with language models\". Participant health/diagnosis is not described; only \"participants_overview\": \"Subjects: 11\".","paper_abstract_analysis":"No paper abstract text was provided in the metadata (only a link to medRxiv). No useful paper information.","evidence_alignment_check":"Pathology: Metadata says nothing about diagnosis/clinical recruitment (e.g., only \"Subjects: 11\"); few-shot patterns do not justify inferring a clinical or healthy cohort. ALIGNMENT: no conflict—insufficient evidence, so Pathology must be Unknown.\nModality: Metadata says \"P300 speller\" and \"flashing paradigms\" (implying visual flashes); few-shot conventions for oddball/speller paradigms typically treat flashing as Visual input. ALIGNMENT: aligns (inference supported by paradigm name).\nType: Metadata says \"P300 speller\" with performance in online typing; few-shot oddball examples suggest P300/oddball paradigms index attention/target detection processes. ALIGNMENT: aligns—P300 speller is essentially an attention-to-target ERP paradigm, even though the application is BCI typing.","decision_summary":"Pathology top-2: (a) Unknown—supported by lack of any diagnostic/health statement (\"Subjects: 11\"); (b) Healthy—plausible for a BCI methods paper, but not stated. Winner: Unknown (metadata has no recruitment/diagnosis facts). Confidence reflects absence of supporting quotes.\nModality top-2: (a) Visual—supported by \"P300 Flashing paradigms\" and \"P300 speller\"; (b) Other—if flashes were not visual (unlikely). Winner: Visual.\nType top-2: (a) Attention—P300 speller/oddball-style target detection; supported by \"P300 speller\" wording and flashing paradigm; (b) Decision-making/Other—could be framed as BCI communication/typing optimization with language models. Winner: Attention because the core neural construct is P300 target attention rather than value-based choice."}},"total_duration_s":null,"tagger_meta":{"config_hash":"3557b68bca409f28","metadata_hash":"79c0c430b1bf564f","model":"openai/gpt-5.2","tagged_at":"2026-04-07T09:32:40.872789+00:00"},"canonical_name":null,"name_confidence":0.74,"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":"Chandravadia2024","size_human":"422.1 MB"}}