Iowa Parkinson's disease resting EEG ("Rest eyes open")

Description

Scalp EEG from an Iowa Parkinson’s disease cohort, collected in the Narayanan and Singh labs at the University of Iowa to test whether resting-state EEG can index cognitive impairment in Parkinson’s disease (Anjum et al., 2024). The dataset comprises eyes-open resting recordings from 149 participants — 100 patients with Parkinson’s disease (PD) and 49 demographically matched healthy controls. PD patients were recruited from the Movement Disorders Clinic at the University of Iowa (recruitment spanning roughly 2017—2022) and recorded while ON their usual dopaminergic medication; controls were drawn from the general Iowa City community. The motivating analysis used these data to optimize a Linear Predictive Coding / “LEAPD” index and traditional spectral measures to estimate general cognitive function (MoCA, NIH Toolbox), reporting strong correlations with cognition (MoCA: ρ ≈ 0.68, p < 0.001) (Anjum et al., 2024). On OpenNeuro the dataset is titled simply “Rest eyes open” (accession ds004584).

Citation

• Paper DOI: 10.1038/s41531-023-00602-0
• Dataset DOI: 10.18112/openneuro.ds004584.v1.0.0
• Reference: Anjum MF, Espinoza AI, Cole RC, Singh A, May P, Uc EY, Dasgupta S, Narayanan NS (2024). Resting-state EEG measures cognitive impairment in Parkinson’s disease. npj Parkinson’s Disease 10, 6. Dataset: Singh A, Cole R, Espinoza A, Cavanagh J, Narayanan N. Rest eyes open. OpenNeuro ds004584 v1.0.0.

Catalog citation block

Anjum MF, Espinoza AI, Cole RC, Singh A, May P, Uc EY, Dasgupta S, Narayanan NS. “Resting-state EEG measures cognitive impairment in Parkinson’s disease.” npj Parkinson’s Disease. 2024 Jan 3;10:6. doi:10.1038/s41531-023-00602-0. (Preprint: medRxiv 2022.07.26.22278079, doi:10.1101/2022.07.26.22278079.)

Dataset: Singh A, Cole R, Espinoza A, Cavanagh J, Narayanan N. “Rest eyes open.” OpenNeuro ds004584, v1.0.0. doi:10.18112/openneuro.ds004584.v1.0.0.

Download and access

Field	Value
Official source	https://openneuro.org/datasets/ds004584
Access class	open
Size	TODO(confirm)
Format	eeglab_set
BIDS	true
Mirrors	Narayanan Lab datasets page (bit.ly/CogPDEEG)

Source & access

• Repository: OpenNeuro — ds004584 (“Rest eyes open”).
• DOI: 10.18112/openneuro.ds004584.v1.0.0 (snapshot v1.0.0).
• License: CC0 (public domain dedication), per dataset_description.json.
• Lab mirror: linked from the Narayanan Lab datasets page via the short link bit.ly/CogPDEEG.
• Authors of record (OpenNeuro): Arun Singh, Rachel Cole, Arturo Espinoza, Jim Cavanagh, Nandakumar Narayanan.

Data structure

The dataset is organized in BIDS (BIDSVersion v1.2.1) with EEG stored in EEGLAB format (.set header + .fdt binary). No preprocessing derivatives are included in the OpenNeuro release; the cleaned/analyzed data underlying the paper are distributed separately by the lab. OpenNeuro snapshots are served via git-annex/DataLad, so the dataset can also be retrieved with datalad install / datalad get.

License

• Name: CC0
• Snippets on this site: allowed — the site may ship short derived snippets and precomputed products from this dataset, labelled as derivatives (re-referenced, filtered, cropped) with citation, license and DOI.

Acquisition

Field	Value
Device	Brain Vision system with 64-channel actiCAP (Brain Products)
Device class	research-cap
Channels	64 (64 recorded; 60 analyzable (Pz reference; Iz, I1, I2 excluded))
Sampling rate (Hz)	500
Online filters	0.1 Hz online high-pass
Reference	Pz (online)
Mains frequency (Hz)	60
Paradigms	rest-eo
Participants	149
Population	100 PD (on dopaminergic medication; 68 M / 32 F; ~68.5 y) + 49 controls (~70.9 y)
Clinical groups	PD, HC
Sessions	1
Durations	eyes-open rest, ~3 min on average, once per participant

Acquisition

EEG was recorded with a Brain Vision system and 64-channel actiCAP (Brain Products GmbH) at a 500 Hz sampling rate, with a 0.1 Hz online high-pass filter and Pz as the online reference (Anjum et al., 2024). The analysis pipeline used 60 of the 64 electrodes per participant (electrodes Iz, I1, and I2 excluded in addition to the Pz reference). Resting recordings were eyes open — chosen deliberately to avoid drowsiness and the posterior dominant alpha rhythm — and were ~3 minutes long on average. Offline, line-noise components at 60, 180, and 200 Hz were removed via FFT-based subtraction and the data were band-pass filtered with a 6th-order Butterworth filter (Anjum et al., 2024); note these are the authors’ analysis steps, not necessarily applied to the raw files on OpenNeuro.

Participants

The released cohort is N = 149: 100 PD patients and 49 controls. Age (mean ± SD): PD 68.53 ± 8.06 years, controls 70.91 ± 7.62 years. Sex: PD 68 male / 32 female; controls 26 male / 23 female (Anjum et al., 2024). Cognition by MoCA: PD 24.31 ± 4.02, controls 26.67 ± 1.86; the paper dichotomized cognitive impairment as MoCA < 26 and cognitively normal as 26—30. PD diagnosis followed the United Kingdom Brain Bank criteria; controls were community members without known neurological disease. The published study additionally used a separate out-of-sample validation set of 32 PD patients, but that validation set is not part of the ds004584 release (Anjum et al., 2024).

Tasks / conditions

A single resting-state condition: eyes-open rest while seated, recorded once per participant (PD patients ON medication). There are no task runs or multiple sessions in this dataset. (Related task datasets from the same labs — Simon/conflict ds004580, oddball ds004574, interval-timing ds004579 — are separate OpenNeuro accessions and are not part of iowapd.)

Used on this site

L1.1, L1.6, L6.3, L7.6, C6, C7, w-filter-sandbox

Loader

Fetches only the stated subset. Recorded loader: DataLad (OpenNeuro ds004584).

# OpenNeuro serves ds004584 as a DataLad/git-annex repository; fetch one subject only.
datalad install https://github.com/OpenNeuroDatasets/ds004584.git
cd ds004584
datalad get sub-001   # replace with the subject id you need

Caveats worth knowing

• PD recorded on medication (attenuates beta signatures).
• 100 vs 49 group imbalance.
• Pz reference channel is flat.

Notable caveats

• Reference and channel count: Pz is the online reference and is therefore flat/absent as a usable signal; the paper additionally drops Iz, I1, I2, yielding 60 analyzable channels rather than the nominal 64.
• Line noise: US 60 Hz mains — expect a 60 Hz line and harmonics (the authors notched 60/180/200 Hz).
• Medication state: PD recordings are ON dopaminergic medication, which attenuates the OFF-state beta signatures of PD; this is a confound for any disease-vs-control oscillatory contrast.
• Short recordings: ~3 min eyes-open segments limit low-frequency resolution and the number of independent epochs for spectral/connectivity estimates.
• Group imbalance: the cohort is imbalanced (100 PD vs 49 controls) and PD is male-skewed (68/32), relevant for any classifier or group-level statistic.

Related datasets

• ERP CORE (Compendium of Open Resources and Experiments)
• EEG Motor Movement/Imagery Dataset (EEGMMIDB)
• LEMON (MPI-Leipzig Mind-Brain-Body)
• Dortmund Vital Study resting EEG
• SRM Resting-state EEG

# Loader recorded in the catalog: DataLad (OpenNeuro ds004584)
# TODO(confirm) the exact call and the subset to fetch. Official source: https://openneuro.org/datasets/ds004584