Mouse HD-EEG under chronic sleep restriction

Description

This is the chronic-sleep-restriction mouse EEG resource of Han et al. (2022), “Dataset of high-density and screw EEG recordings under chronic sleep restriction in mice,” accompanying the Scientific Data descriptor “Nine-day continuous recording of EEG and 2-hour of high-density EEG under chronic sleep restriction in mice” (Han et al., 2022). It was collected by the Choi laboratory (Center for Neuroscience, Korea Institute of Science and Technology / KAIST, with collaborators at the University of Michigan) to study how chronic partial sleep loss reshapes both global and local cortical oscillations in freely-moving mice. The recording design pairs a 9-day continuous low-channel “screw” EEG (frontal + parietal) with accelerometry against intermittent 2-hour high-density EEG (HD-EEG) sessions captured from a flexible polyimide microarray, so that homeostatic sleep regulation can be tracked at the whole-night scale and the topography of slow waves and spindles can be mapped at high spatial resolution. The cohort is nine male C57BL/6 × 129S4/SvJae hybrid mice, ~12 weeks old at surgery.

Citation

• Paper DOI: 10.1038/s41597-022-01354-x
• Dataset DOI: 10.12751/g-node.mvi0dp
• Reference: Han, H.-B., Kim, B., Kim, Y., Jeong, Y., & Choi, J. H. (2022). Nine-day continuous recording of EEG and 2-hour of high-density EEG under chronic sleep restriction in mice. Scientific Data 9, 225. Dataset: G-Node, DOI 10.12751/g-node.mvi0dp.

Catalog citation block

Han, H.-B., Kim, B., Kim, Y., Jeong, Y., & Choi, J. H. (2022). Nine-day continuous recording of EEG and 2-hour of high-density EEG under chronic sleep restriction in mice. Scientific Data, 9, 225. https://doi.org/10.1038/s41597-022-01354-x

Dataset: Han, H.-B., Kim, B., Kim, Y., & Choi, J. H. (2022). Dataset of high-density and screw EEG recordings under chronic sleep restriction in mice [Data set]. G-Node. https://doi.org/10.12751/g-node.mvi0dp (CC BY 4.0).

Download and access

Field	Value
Official source	https://gin.g-node.org/hiobeen/Mouse_EEG_ChronicSleepRestriction_Kim_et_al
Access class	open
Size	~90–100 GB (full upstream, reported)
Format	bids
BIDS	true
Mirrors	none recorded

Source & access

• Primary repository: G-Node GIN — https://gin.g-node.org/hiobeen/Mouse_EEG_ChronicSleepRestriction_Kim_et_al
• GIN DOI: 10.12751/g-node.mvi0dp (registered 2022-02-12)
• Data descriptor: Han et al. (2022), Scientific Data 9:225, 10.1038/s41597-022-01354-x (PMC9126869)
• License: Creative Commons Attribution 4.0 International (CC BY 4.0)
• Access: GIN is git-annex backed. Recommended retrieval is via gin-cli (gin get hiobeen/Mouse_EEG_ChronicSleepRestriction_Kim_et_al then gin download --content); the web ZIP/TAR.GZ contains git-annex pointer files unless content is pulled. The full upstream dataset is large (~90-100 GB reported).

Note: a separate, smaller (~11 GiB) Choi-lab dataset — “high-density EEG recordings with auditory and optogenetic stimulation in mice” (Hwang et al., GIN DOI 10.12751/g-node.e5tyek) — exists on the same platform but is not this dataset (different size, paradigm, and authorship order).

Data structure

The dataset is laid out in EEGLAB-compatible BIDS. Per subject and session, EEG is stored as paired EEGLAB .set (header) + .fdt (binary) files named like sub-##-..._screw.set and sub-##-..._hdeeg.set. Sidecar .tsv tables carry channel information (*_channels.tsv) and sleep-stage event annotations (*_events.tsv: wake / NREM / REM scored in 10-second epochs on selected days). A montage.csv / electrode-coordinate file gives the HD-array layout. Underlying storage is git-annex; large binaries must be content-pulled (see Source & access). The descriptor also ships Python analysis scripts (slow-wave/spindle detection, topographic mapping, spectral analysis).

License

• Name: CC-BY-4.0
• 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	36-channel flexible polyimide microarray (CompuMedics Synamps2) plus 2-channel skull-screw EEG (Grass QP511 / Digidata 1440A)
Device class	microarray
Channels	36 (36 HD-array channels (task-hdeeg) + 2 screw channels and an accelerometer (task-screw); registry value: 36+2)
Sampling rate (Hz)	500
Online filters	0.3–100 Hz analog band-pass and 60 Hz notch
Reference	interparietal-bone electrode
Mains frequency (Hz)	60
Paradigms	sleep-restriction, sleep-wake
Participants	9
Population	9 male C57BL/6 × 129S4/SvJae hybrid mice, ~12 weeks at surgery
Clinical groups	none
Sessions	9
Durations	9-day continuous 2-ch screw EEG (~24 h/day) and 2-h HD-EEG sessions (ZT1–ZT3) on 8 days; baseline, 5 sleep-restriction days, 3 recovery days; wake/NREM/REM scored in 10-s epochs

Acquisition

Two recording subsystems were used per animal:

• High-density EEG (task-hdeeg): 36 active channels read from a flexible polyimide microarray (40 leads total, 8 μm film thickness; 500 μm electrode radius; 1.2-2.0 mm center-to-center spacing; ~100 kΩ at 1 kHz). Amplified with a CompuMedics Synamps2 system, digitized at 1 kHz and downsampled to 500 Hz.
• Conventional “screw” EEG (task-screw): 2 channels from stainless-steel skull screws over frontal and parietal cortex (0.8 mm; 10-50 kΩ at 1 kHz), plus a co-recorded accelerometer channel for movement/state. Amplified with a Grass Technologies QP511 amplifier into a Molecular Devices Digidata 1440A at 500 Hz.

Both streams used a 0.3-100 Hz analog bandpass and a 60 Hz notch (line frequency in South Korea). The reference/ground was an interparietal-bone electrode. Recording lengths: the screw EEG ran ~24 h/day continuously across 9 days; HD-EEG was collected in 2-hour windows (Zeitgeber time ZT1-ZT3) on 8 days.

Participants

Nine subjects (sub-01 through sub-09), all male C57BL/6 × 129S4/SvJae hybrid mice, ~12 weeks old at the time of electrode-implant surgery (Han et al., 2022). No clinical/disease groups — the manipulation is an experimental sleep-restriction protocol applied to all animals, not a patient-vs-control design. One animal lacks a complete HD-EEG series: the 2-hour HD-EEG sessions are available for 8 of the 9 mice/days, while the continuous screw EEG covers all nine.

Tasks / conditions

The protocol spans nine sessions per mouse (ses-01..ses-09) labelled by sleep-restriction phase: BL (baseline), SR1-SR5 (five consecutive sleep-restriction days), and R1-R3 (three recovery days). Sleep deprivation was 18 hours/day for 5 consecutive days, enforced by a slowly rotating motorized wheel (~2.3 RPM, ~4-second on-intervals) that interrupts sleep onset without forced locomotion. Each session contributes the continuous screw/accelerometer recording (task-screw) and, on HD-EEG days, a 2-hour high-density block (task-hdeeg). There is no cognitive task — the conditions are sleep/wake states under homeostatic pressure.

Used on this site

L0.1, L7.2

Loader

Fetches only the stated subset. Recorded loader: gin-cli — see snippet.

# GIN is git-annex backed: the web ZIP holds pointer files unless content is pulled.
gin get hiobeen/Mouse_EEG_ChronicSleepRestriction_Kim_et_al
cd Mouse_EEG_ChronicSleepRestriction_Kim_et_al
gin get-content sub-01   # one animal only

Caveats worth knowing

• HD-EEG sessions on 8 of the 9 protocol days (one animal lacks a complete series).
• 60 Hz notch already applied (Korea).

Notable caveats

• Line noise: 60 Hz mains (Korea) with a 60 Hz analog notch already applied — do not assume 50 Hz, and expect a notch gap at 60 Hz and harmonics.
• Sampling/bandwidth: Final sampling is 500 Hz with a 0.3-100 Hz passband; the HD-EEG was acquired at 1 kHz then downsampled, so usable bandwidth is capped near 100 Hz (high-gamma analyses are limited).
• Two heterogeneous montages: 36-ch HD array vs. 2-ch screw EEG are different spatial samplings amplified on different hardware (Synamps2 vs. Grass/Digidata); they should not be pooled channel-wise.
• Reference: interparietal-bone reference, not a standard scalp reference — re-referencing choices matter for topography.
• Coverage gap: HD-EEG present for 8 of 9 mice/days; screw EEG for all 9. Session count and HD-EEG availability vary, so per-subject file counts are not uniform.

Related datasets

• 1024-channel macaque V1/V4 resting state

# Loader recorded in the catalog: gin-cli — see snippet
# TODO(confirm) the exact call and the subset to fetch. Official source: https://gin.g-node.org/hiobeen/Mouse_EEG_ChronicSleepRestriction_Kim_et_al