Datasets:

EEGDash
/

ds004844

dataset_id stringclasses 1 value	title stringclasses 1 value	source stringclasses 1 value	source_url stringclasses 1 value	doi stringclasses 1 value	license stringclasses 1 value	loader dict	catalog stringclasses 1 value	generated_by stringclasses 1 value
ds004844	T22	openneuro	https://openneuro.org/datasets/ds004844	10.18112/openneuro.ds004844.v1.0.0	CC0	{ "library": "eegdash", "class": "EEGDashDataset", "kwargs": { "dataset": "ds004844" } }	https://huggingface.co/spaces/EEGDash/catalog	huggingface-space/scripts/push_metadata_stubs.py

T22

Dataset ID: ds004844

Metcalfe2023_T22

At a glance: EEG · Visual decision-making · healthy · 17 subjects · 68 recordings · CC0

Load this dataset

This repo is a pointer. The raw EEG data lives at its canonical source (OpenNeuro / NEMAR); EEGDash streams it on demand and returns a PyTorch / braindecode dataset.

# pip install eegdash
from eegdash import EEGDashDataset

ds = EEGDashDataset(dataset="ds004844", cache_dir="./cache")
print(len(ds), "recordings")

If the dataset has been mirrored to the HF Hub in braindecode's Zarr layout, you can also pull it directly:

from braindecode.datasets import BaseConcatDataset
ds = BaseConcatDataset.pull_from_hub("EEGDash/ds004844")

Dataset metadata


Subjects	17
Recordings	68
Tasks (count)	1
Sessions	4
Channels	72 (×68)
Sampling rate (Hz)	1024 (×68)
Total duration (h)	21.3
Size on disk	22.3 GB
Recording type	EEG
Experimental modality	Visual
Paradigm type	Decision-making
Population	Healthy
BIDS version	1.8.0
Source	openneuro
License	CC0
NEMAR citations	0

Tasks

Drive

Upstream README

Verbatim from the dataset's authors — the canonical description.

TX22 dataset: Predicting and influencing trust-based decisions about control authority hand-off and take-over during simulated, semi-automated driving in a leader-follower paradigm.Vehicle survivability is critically important in todays military. Significant DoD investments have focused on developing and integrating autonomous vehicle technologies to mitigate the effects of human error and thus enhance surviability and mission effectiveness. In a previous experiment (SANDR designation: ARL_TX20), we explored how a human operators acceptance and use of advanced technology is influenced by their trust and related factors, like subjective workload and automation reliability. Nevertheless, more critical than measuring and achieving a certain level of trust is the need for a capability to resolve observed (or predicted) discrepancies between trust and trustworthiness that will undermine effective joint system performance. Using the same paradigm as we developed for our previous experiment (ARL_TX20), here we explore our ability to (a) make accurate real-time predictions of instances where intervention is necessary and (b) use those predictions to provide feedback to the driver that is intended to support active "trust management" by influencing the trust-based decisions of the driver.

People

Authors

Jason S. Metcalfe
Victor Paul
Benamin Haynes
Corey Atwater
Amar Marathe
Gregory Gremillion
Kim Drnec
William Nothwang
Justin R. Estepp
Margaret Bowers
Jamie Lukos
Tony Johnson
Mike Dunkel
Stephen Gordon
Jon Touryan
Kevin King (senior)

Contact

Kevin King

Provenance

Backend: s3 — s3://openneuro.org/ds004844
Exact size: 23,976,121,518 bytes (22.3 GB)
Ingested: 2026-04-06
Stats computed: 2026-04-04

Auto-generated from dataset_summary.csv and the EEGDash API. Do not edit this file by hand — update the upstream source and re-run scripts/push_metadata_stubs.py.

Downloads last month: 34

Number of rows:

Total file size:

7.33 kB

T22