{"dataset":{"id":"415","dataset_id":"nm000232","name":"THINGS-EEG2: A large and rich EEG dataset for modeling human visual object recognition","description":"THINGS-EEG2 is a large-scale EEG dataset comprising recordings from 10 subjects viewing 16,540 distinct training images and 200 test images presented via rapid serial visual presentation at 5 Hz. The dataset includes 63-channel EEG data sampled at 1000 Hz across 4 sessions per subject, with approximately 32,540 training trials and 16,000 test trials, designed to support computational modeling of human visual object recognition. Stimuli are drawn from the THINGS database, and the dataset includes resting-state recordings and behavioral annotations for each trial.","owner_user_id":19,"status":"active","github_repo":"nemarDatasets/nm000232","concept_doi":"10.82901/nemar.nm000232","latest_version_doi":"10.82901/nemar.nm000232.v1.1.0","created_at":"2026-04-11 14:54:14","updated_at":"2026-07-10 21:58:56","zenodo_concept_id":"21210570","is_sandbox":0,"visibility":"public","ezid_status":"public","enrichment_json":"{\n  \"version\": \"2.0\",\n  \"pipeline_stage\": \"validated\",\n  \"authors\": {\n    \"Alessandro T. Gifford\": {\n      \"orcid\": \"0000-0002-8923-9477\"\n    },\n    \"Kshitij Dwivedi\": {},\n    \"Gemma Roig\": {},\n    \"Radoslaw M. Cichy\": {}\n  },\n  \"related_identifiers\": [\n    {\n      \"identifier\": \"10.17605/OSF.IO/3JK45\",\n      \"identifier_type\": \"DOI\",\n      \"relation_type\": \"IsDerivedFrom\"\n    },\n    {\n      \"identifier\": \"10.1016/j.neuroimage.2022.119754\",\n      \"identifier_type\": \"DOI\",\n      \"relation_type\": \"References\"\n    },\n    {\n      \"identifier\": \"https://github.com/nemarDatasets/nm000232\",\n      \"identifier_type\": \"URL\",\n      \"relation_type\": \"IsDescribedBy\"\n    },\n    {\n      \"identifier\": \"https://nemar.org/dataset/nm000232\",\n      \"identifier_type\": \"URL\",\n      \"relation_type\": \"IsDescribedBy\"\n    }\n  ],\n  \"funding_references\": [\n    {\n      \"funder_name\": \"Hessian Ministry of Higher Education, Research, Science and the Arts (HMWK)\"\n    },\n    {\n      \"funder_name\": \"DFG\",\n      \"award_number\": \"CI 241/1-1\"\n    },\n    {\n      \"funder_name\": \"DFG\",\n      \"award_number\": \"CI 241/3-1\"\n    },\n    {\n      \"funder_name\": \"DFG\",\n      \"award_number\": \"CI 241/1-7\"\n    },\n    {\n      \"funder_name\": \"ERC\",\n      \"award_number\": \"ERC-2018-StG 803370\"\n    },\n    {\n      \"funder_name\": \"HMWK\"\n    }\n  ],\n  \"title\": \"THINGS-EEG2: A large and rich EEG dataset for modeling human visual object recognition\",\n  \"license\": \"CC-BY 4.0\",\n  \"dataset_type\": \"raw\",\n  \"resource_type_general\": \"Dataset\",\n  \"modalities\": [\n    \"eeg\"\n  ],\n  \"resource_type_specific\": \"EEG Dataset\",\n  \"sizes\": [\n    \"212.2 GB (1504 files)\"\n  ],\n  \"formats\": [\n    \".bdf\",\n    \".eeg\",\n    \".extracted\",\n    \".jpg\",\n    \".json\",\n    \".mat\",\n    \".md\",\n    \".npy\",\n    \".py\",\n    \".tsv\",\n    \".txt\",\n    \".vhdr\",\n    \".vmrk\",\n    \".yml\",\n    \".zip\"\n  ],\n  \"description\": \"THINGS-EEG2 is a large-scale EEG dataset comprising recordings from 10 subjects viewing 16,540 distinct training images and 200 test images presented via rapid serial visual presentation at 5 Hz. The dataset includes 63-channel EEG data sampled at 1000 Hz across 4 sessions per subject, with approximately 32,540 training trials and 16,000 test trials, designed to support computational modeling of human visual object recognition. Stimuli are drawn from the THINGS database, and the dataset includes resting-state recordings and behavioral annotations for each trial.\",\n  \"methods_description\": \"EEG data were recorded using a Brain Products actiCHamp system with 63 channels in a 10-10 cap layout at 1000 Hz sampling rate. Stimuli were presented via rapid serial visual presentation (RSVP) at 5 Hz. Each of 4 sessions contained 5 training runs (~3,360 trials each) of unique images, 1 test run (~4,080 trials) of 200 repeated images, and 2 resting-state runs (before and after the main task). Online band-pass filtering was applied at 0.01-100 Hz.\",\n  \"keywords\": [\n    {\n      \"term\": \"EEG\"\n    },\n    {\n      \"term\": \"visual object recognition\"\n    },\n    {\n      \"term\": \"rapid serial visual presentation\"\n    },\n    {\n      \"term\": \"neural encoding\"\n    },\n    {\n      \"term\": \"image classification\"\n    },\n    {\n      \"term\": \"human vision\"\n    }\n  ],\n  \"source_hash\": \"562a1589334324046776308484a5c086521536e9babd277e77c2b8f818ed3912\"\n}","last_activity_at":"2026-07-05 21:21:05","source":null,"source_id":null,"subject_count":10,"modalities":"eeg","age_min":24,"age_max":34,"file_size":212160209984,"total_files":1504,"tasks":"rest,rest1,rest2,test,train","metadata_columns_error":null,"staleness_warn_stage":null,"staleness_admin_notified_at":null,"authors":"Alessandro T. Gifford, Kshitij Dwivedi, Gemma Roig, Radoslaw M. Cichy","license":"CC-BY 4.0","readme":"[![DOI](https://img.shields.io/badge/DOI-10.82901%2Fnemar.nm000232-blue)](https://doi.org/10.82901/nemar.nm000232)\n\nTHINGS-EEG2: A large and rich EEG dataset for modeling human visual object recognition\n========================================================================================\n\nOverview\n--------\nEEG dataset of 10 subjects who viewed 16,540 distinct training images and 200\ntest images (each repeated ~80 times) using rapid serial visual presentation\n(RSVP) at 5 Hz, recorded on a BrainVision actiCHamp system at 1000 Hz.\nThe source files store 63 EEG channels (the online reference electrode is\nnot stored). Stimuli are drawn from the THINGS database (Hebart et al. 2019).\n\nEach subject completed 4 separate sessions; each session contained:\n  - 5 training runs (~3,360 trials each) covering ~16,540 unique images\n  - 1 test run (~4,080 trials) of 200 images repeated 20× per session\n  - 2 resting-state runs (one before, one after the main experiment)\n\nTotal: ~32,540 training trials + ~16,000 test trials per subject across 4 sessions.\n\nRecording setup\n---------------\n- Manufacturer: Brain Products (actiCHamp)\n- 63 EEG channels (one electrode served as online reference and is not\n  stored in the source files)\n- 10-10 cap layout\n- Sampling rate: 1000 Hz\n- Online band-pass: 0.01-100 Hz\n- Triggers recorded as BrainVision stimulus annotations (not as a\n  dedicated stim channel)\n\nTasks (BIDS labels)\n-------------------\n- task-train: training run (RSVP of unique images)\n- task-test:  test run (RSVP of repeated test images)\n- task-rest:  resting state (eyes open, fixation cross)\n\nRun numbering\n-------------\n- task-train: run-01..run-05 per session (5 training parts)\n- task-test:  single run per session\n- task-rest:  run-01 (before main task) and run-02 (after main task)\n\nEvents\n------\nevents.tsv columns:\n  onset, duration, sample, value, trial_type\n  tot_img_number     - global image ID (1-16540 for train; 1-200 for test;\n                       'n/a' for target catch trials)\n  img_category       - integer category index\n  category_name      - human-readable category, e.g. \"01175_roller_coaster\"\n  block, sequence    - hierarchical position within the run\n  img_in_sequence    - image position within its 20-image sequence\n  soa                - actual stimulus onset asynchrony (~200 ms)\n\ntrial_type values:\n  image  - normal training/test image presentation\n  target - random catch trial (subject must press a button)\n  rest_marker - resting-state start/end marker\n\nSubject information\n-------------------\nparticipants.tsv contains age and sex (both extracted from the\nbehavioural .mat files in the source data).\n\nFolder layout\n-------------\n/sub-XX/ses-YY/eeg/        - main BIDS data (BDF + sidecars)\n/sourcedata/               - original BrainVision .eeg/.vhdr/.vmrk and\n                             behavioural .mat files\n/derivatives/preprocessed_eeg/   - authors' preprocessed train/test epochs\n/derivatives/resting_state/      - authors' preprocessed resting state\n/stimuli/                  - image set (training_images.zip, test_images.zip)\n                             plus image_metadata.npy\n/code/                     - this conversion script\n\nReference\n---------\nGifford, A.T., Dwivedi, K., Roig, G., & Cichy, R.M. (2022). A large and rich\nEEG dataset for modeling human visual object recognition. NeuroImage, 264,\n119754. https://doi.org/10.1016/j.neuroimage.2022.119754\n\nCode: https://github.com/gifale95/eeg_encoding\nOSF:  https://osf.io/3jk45/\n","bids_version":"1.9.0","sessions_count":4,"publish_date":"2026-05-11 18:18:55","embedding_dirty":0,"license_tier":"attribution","zarr_status":"ready","zarr_converted_at":"2026-09-03 13:07:58","zarr_store_count":319,"zarr_index_etag":"b7c35f80885532055cc4af156ee79eae","zarr_source_commit":"5732fd1a9b9c873c94b25c7e7e70e550dc64faa7","archive_status":null,"archive_size":null,"archive_retry_count":0,"records_status":"ready","archive_skip_reason":"dataset 241.5 GB exceeds 100.0 GB archive limit; use direct download","zarr_errors":0,"zarr_failure_count":0,"zarr_deterministic":0,"zarr_failed_at":null,"num_dataset_citations":0,"num_datapaper_citations":82,"n_channels":63,"electrode_system":"10-10","has_hed":0,"hed_version":"8.2.0","is_exemplar":0,"bytes_present":null,"data_complete":null,"withdrawn_at":null,"withdrawn_reason":null,"archive_complete":null,"archive_absent_files":null,"archive_declared_files":null,"zarr_pool_breaks":0,"total_recording_duration":314348,"recording_duration_min":316,"recording_duration_max":2294,"recording_count":319,"recordings_unavailable":0,"recordings_measured":319,"channel_count_min":63,"channel_count_max":63,"sampling_frequency":1000,"power_line_frequency":50,"eeg_reference":"FCz (online); offline re-referenced as needed","placement_scheme":"International 10-10","sweep_stamps":"{\"enrichment_updated_at\":\"2026-07-05 21:32:21\",\"metadata_updated_at\":\"2026-07-05 21:32:54\",\"archive_checked_at\":\"2026-07-05 21:33:14\",\"zarr_checked_at\":\"2026-06-07 17:58:34\",\"records_checked_at\":\"2026-07-05 21:33:59\",\"citations_updated_at\":\"2026-09-10 03:00:14\",\"channel_montage_checked_at\":\"2026-06-28 22:59:21\",\"hed_checked_at\":\"2026-06-30 04:29:29\",\"data_checked_at\":null,\"availability_report_at\":\"2026-07-23 01:08:52\",\"signal_defaults_at\":\"2026-09-02 11:46:50\",\"recording_stats_at\":\"2026-09-04 03:00:20\"}","participants":10,"num_citations":82,"latest_version":"v1.1.0","zarr_verify_status":null,"zarr_verified_at":null,"owner_username":"bruaristimunha","owner_github":"bruAristimunha","file_size_formatted":"198 GB","zarr_data_failures":null,"zarr_index_url":"https://zarr.nemar.org/nm000232/zarr/index.json","attestation_deposit_type":null,"attestation_key_status":null,"attestation_deidentified":null,"attestation_no_duplicate":null,"attestation_upstream_source":null,"attestation_accepted_at":null}}