Wearanize plus (Wearanize+)

—A multimodal dataset with wearable-based overnight sleep recordings

Table of Contents: Overview | Access Instructions | Devices and Modalities | Dataset Versions | Script Descriptions | PlugNPlay Channel Descriptions | Ethical Statements | Funding | Reference Paper | Citation | People

Overview

The Wearanize+ dataset comprises overnight sleep recordings from 130 healthy participants (one night each) aged between 18 and 39 years (mean = 23.16 years; SD = 4.34; 89F & 41M). Each participant's sleep was recorded simultaneously using three wearable devices—a Zmax EEG headband, an Empatica E4 wristband, and an ActivPAL leg patch—alongside full polysomnography (PSG) using SOMNOscreen plus or Mentalab Explore Pro (for a few participants). It also includes the responses to three widely used questionnaires—the Pittsburgh Sleep Quality Index (PSQI), the Mannheim Dream Questionnaire (MADRE), and the Patient Health Questionnaire (PHQ-9)—providing information on the participants' sleep, dreams, and overall health. The PSG data has been manually sleep-scored by an expert sleep scorer and automatically sleep-scored by USleep v2.0. Both sets of scores are included in the dataset. For more details, see the reference paper.

See Access Instructions for a step-by-step guide to obtaining access to the dataset. This repository also contains the scripts used to preprocess, synchronize, and prepare the dataset. See Script Descriptions for more details.

The dataset is the outcome of a research project bearing the same name, carried out at the Trigon Building of the Donders Centre for Cognitive Neuroimaging, Radboud University (Nijmegen, The Netherlands). The study was conducted by members of Donders Sleep & Memory Lab, in collaboration with Radboud University Medical Center (Radboudumc, Nijmegen, The Netherlands) and Hochschule Rhein-Waal (Kleve, Germany), between October 2023 and August 2024.

The dataset can facilitate a range of applications, including device-specific validations of the three wearables, development of (device-specific) automatic sleep-stage scorers (autoscorers) based on the provided PSG-based sleep scores, methods for handling missing or corrupted data, and evaluation of alternative (as well as compound) sensor modalities for sleep scoring. It has already been used for developing Zmax-based autoscorers, such as ezscore and u-sleep-w, and validating an automatic Zmax–Somnoscreen synchronization method. One of our key objectives behind creating this dataset is to leverage these multimodal recordings to build robust, multi-wearable sleep-scoring models that approach PSG-grade performance while minimizing the impact of EEG artifacts.

Access Instructions

The Wearanize+ dataset is hosted on the Radboud Data Repository (RDR) in this collection and is available for research upon signing a Data Use Agreement (DUA). To request access, download the DUA template, complete the sections on recipient information, briefly describe your research plan, and send the filled-out form to Dr. Martin Dresler via email.

Please note that simply clicking “Request access” on the RDR collection page will not grant you access to the dataset. You must complete the DUA and submit it to Dr. Dresler via your institute or lab head. Access will be granted once the DUA has been fully agreed upon between the legal departments of Radboudumc and your institute and signed by all parties.

Devices and Modalities

The following image shows the positions of the mentioned devices and their recording modalities.
Recordings from the experimental devices are not a part of the dataset.

Dataset Versions

For transparency and ease of use, the dataset has been released in two versions: Wearanize+ Raw v1.0 and Wearanize+ PlugNPlay v1.0. PlugNPlay would be the ideal version for most projects, while the Raw version allows tracing back to the original data and may provide the opportunity for further analysis. Here are the differences in their contents:

Wearanize+ Raw v1.0

This version/file contains the raw, unfiltered data collected from the participants of the project. See Section 3.1 and Appendix 2 of the reference paper for more details.

Wearanize+ PlugNPlay v1.0

This version/file contains a processed, synchronized, and truncated version of the raw data. To streamline usability and avoid repeating the extensive preprocessing steps, data for each participant was consolidated into a single European Data Format (EDF) file, preserving all metadata and signal properties. PSG-based Manual and automatic sleep scores were also integrated into the EDF files as 'PSG_Manual_score' and 'PSG_USleep_score' at a sampling rate of 1/30 Hz. Time-series signals were labeled according to the convention [device_ID]_[channel_name] and stored with the Float32 datatype (if they are read in Float64, convert them back to Float32 to save space). The PlugNPlay version includes data from 100 participants (out of the total 130) for whom both PSG and Zmax data were available, and manual sleep scoring could be performed.

In most cases, the channel names were kept consistent with the names provided by the associated device. However, they were sometimes modified for clarity or broader compatibility. See PlugNPlay Channel Descriptions or the subject-wise sub-nnn_task-sleep_channels.tsv files for detailed information on specific channels. [device_name]_[channel_a]:[channel_b] indicates that channel_a was referenced to channel_b.

Since EDF is a widely used format in Neuroscience, the data should be readable across different platforms and environments. The PlugNPlay version has been formatted according to the EEG-Brain Imaging Data Structure (EEG-BIDS v1.10.0) specifications. The usability of the EEG signals has been checked with eegFloss, and the outputs have been added to the corresponding file. See Section 3.2 of the reference paper for more details.

Important

Please note that PSG data for Sub115, Sub124, Sub129, and Sub130 were collected with Mentalab, and their channel names differ from those collected with Somnoscreen (used for all other participants). See PlugNPlay Channel Descriptions for the exact channel mappings.

Script Descriptions

read_PlugNPlay_EDF.py

• Provides example code to read individual EDF files and extract information from multiple EDF files of the PlugNPlay version using Python.
• Note: Storing raw signals from all EDF files simultaneously requires substantial memory.

read_PlugNPlay_EDF.m

• Provides example code to read individual EDF files and extract information from multiple EDF files of the PlugNPlay version using MATLAB.
• Note: Storing raw signals from all EDF files simultaneously requires substantial memory.

PlugNPlay_preparation//create_PlugNPlay_parquet.py

• Contains the codes used to create the PlugNPlay version from the raw version of the dataset.
• The PlugNPlay version was first generated in Parquet format, storing data as Pandas DataFrames for efficient processing.
• The Parquet files are also available in the data repository as Wearanize+ PlugNPlay Parquet v1.0. If needed, they can serve as alternatives to the EDF files in Wearanize+ PlugNPlay v1.0. More information and usage instructions of these files are available in extra_scripts//read_PlugNPlay.py and extra_scripts//read_PlugNPlay.m.

PlugNPlay_preparation//parquet_to_EDF.py

• Contains the scripts used to convert the Parquet files into EDF format while preserving all relevant signal information and metadata.

EEG-BIDS_preparation//BIDS_derivative.py

• Prepares the EEG-BIDS–compatible derivative dataset containing global metadata and structure definitions.

EEG-BIDS_preparation//BIDS_info_1.py

• Prepares the EEG-BIDS–compatible dataset with local (file-specific) metadata and participant-level information.

automatic_synchronization//Zmax-Somnoscreen_auto_sync.m

• Provides MATLAB code to automatically synchronize simultaneously recorded overnight sleep data using the Zmax headband and SOMNOscreen plus PSG devices.
• To apply this script to new data, organize the input files (Zmax recording, SOMNOscreen plus recording, and Lights Out/Lights On moments from Zmax) in the same structure as in Wearanize+ Raw v1.0, and update the input–output directories at the beginning of the script.
• See Appendix 1 of the reference paper for more details.

manual_synchronization//Zmax-Somnoscreen_manual_sync.m

• Contains MATLAB code for manual/visual synchronization of Zmax and SOMNOscreen plus recordings based on their respective accelerometer and movement signals.
• See Section 2.3.3 of the reference paper for more details.

manual_synchronization//Zmax-Empatica-Activpal_manual_sync.m

• Contains MATLAB code for manual/visual synchronization of Zmax recordings with Empatica E4 and ActivPAL data, based on their respective accelerometer outputs.

PlugNPlay Channel Descriptions

Channel Name	Description$	Unit	Device	Sampling Frequency (Hz)
ActivPal_ACCX	Accelerometer X axis	ⓖ	ActivPAL	20
ActivPal_ACCY	Accelerometer Y axis	ⓖ	ActivPAL	20
ActivPal_ACCZ	Accelerometer Z axis	ⓖ	ActivPAL	20
Emp_ACCX	Accelerometer X axis	ⓖ/64	Empatica E4	32
Emp_ACCY	Accelerometer Y axis	ⓖ/64	Empatica E4	32
Emp_ACCZ	Accelerometer Z axis	ⓖ/64	Empatica E4	32
Emp_BVP	PPG	Unitless	Empatica E4	64
Emp_EDA	Electrodermal activity	μS	Empatica E4	4
Emp_HR	Mean heart rate derived from BVP	bpm	Empatica E4	1
Emp_TEMP	Skin temperature	°C	Empatica E4	4
PSG_A1	EEG channel A1	µV	SOMNOscreen	256
PSG_A2	EEG channel A2	µV	SOMNOscreen	256
PSG_ACCX	Accelerometer X axis	mⓖ	Mentalab	20
PSG_ACCY	Accelerometer Y axis	mⓖ	Mentalab	20
PSG_ACCZ	Accelerometer Z axis	mⓖ	Mentalab	20
PSG_C3	EEG channel C3	µV	SOMNOscreen, Mentalab	256, 250
PSG_C3:A2	EEG channel C3 referenced to A2	µV	SOMNOscreen	256
PSG_C4	EEG channel C4	µV	SOMNOscreen, Mentalab	256, 250
PSG_C4:A1	EEG channel C4 referenced to A1	µV	SOMNOscreen	256
PSG_CP1	EEG channel CP1	µV	Mentalab	250
PSG_CP2	EEG channel CP2	µV	Mentalab	250
PSG_CP5	EEG channel CP5	µV	Mentalab	250
PSG_CP6	EEG channel CP6	µV	Mentalab	250
PSG_Cz	EEG channel Cz	µV	Mentalab	250
PSG_ECG 2	ECG channel 2	µV	SOMNOscreen	256
PSG_ECG1	ECG channel 1	µV	Mentalab	250
PSG_ECG2	ECG channel 2	µV	Mentalab	250
PSG_EMG	EMG channel reference	µV	SOMNOscreen	256
PSG_EMG_minus	EMG channel 1	µV	SOMNOscreen	256
PSG_EMG_plus	EMG channel 2	µV	SOMNOscreen	256
PSG_EMG1	EMG channel 1	µV	Mentalab	250
PSG_EMG2	EMG channel 2	µV	Mentalab	250
PSG_EOG1	EOG channel 1	µV	SOMNOscreen, Mentalab	256, 250
PSG_EOG1:A1	EOG channel 1 referenced to A1	µV	SOMNOscreen	256
PSG_EOG1:A2	EOG channel 1 referenced to A2	µV	SOMNOscreen	256
PSG_EOG2	EOG channel 2	µV	SOMNOscreen, Mentalab	256, 250
PSG_EOG2:A1	EOG channel 2 referenced to A1	µV	SOMNOscreen	256
PSG_EOG2:A2	EOG channel 2 referenced to A2	µV	SOMNOscreen	256
PSG_F3	EEG channel F3	µV	SOMNOscreen	256
PSG_F3:A2	EEG channel F3 referenced to A2	µV	SOMNOscreen	256
PSG_F4	EEG channel F4	µV	SOMNOscreen	256
PSG_F4:A1	EEG channel F4 referenced to A1	µV	SOMNOscreen	256
PSG_F7	EEG channel F7	µV	Mentalab	250
PSG_F8	EEG channel F8	µV	Mentalab	250
PSG_FC1	EEG channel FC1	µV	Mentalab	250
PSG_FC2	EEG channel FC2	µV	Mentalab	250
PSG_FC5	EEG channel FC5	µV	Mentalab	250
PSG_FC6	EEG channel FC6	µV	Mentalab	250
PSG_FCz	EEG channel FCz	µV	Mentalab	250
PSG_FT10	EEG channel FT10	µV	Mentalab	250
PSG_FT9	EEG channel FT9	µV	Mentalab	250
PSG_GYRX	Gyroscope X axis	mdps	Mentalab	20
PSG_GYRY	Gyroscope Y axis	mdps	Mentalab	20
PSG_GYRZ	Gyroscope Z axis	mdps	Mentalab	20
PSG_MAGX	Magnetometer X axis	µT	Mentalab	20
PSG_MAGY	Magnetometer Y axis	µT	Mentalab	20
PSG_MAGZ	Magnetometer Z axis	µT	Mentalab	20
PSG_Move.	Movement info	mⓖ	SOMNOscreen	4
PSG_O1	EEG channel O1	µV	SOMNOscreen, Mentalab	256, 250
PSG_O1:A2	EEG channel O1 referenced to A2	µV	SOMNOscreen	256
PSG_O2	EEG channel O2	µV	SOMNOscreen, Mentalab	256, 250
PSG_O2:A1	EEG channel O2 referenced to A1	µV	SOMNOscreen	256
PSG_Oz	EEG channel Oz	µV	Mentalab	250
PSG_P3	EEG channel P3	µV	Mentalab	250
PSG_P4	EEG channel P4	µV	Mentalab	250
PSG_P7	EEG channel P7	µV	Mentalab	250
PSG_P8	EEG channel P8	µV	Mentalab	250
PSG_Pos.	Body position infoⓟ	Unitless	SOMNOscreen	4
PSG_Pz	EEG channel Pz	µV	Mentalab	250
PSG_T7	EEG channel T7	µV	Mentalab	250
PSG_T8	EEG channel T8	µV	Mentalab	250
PSG_Manual_score	Manually-identified sleep scoresⓢ from PSG data	Unitless	N/A	1/30
PSG_USleep_score	Automatic sleep scoresⓢ identified by Usleep v2.0	Unitless	N/A	1/30
Zmax_ACCX	Accelerometer X axis	ⓖ	Zmax	256
Zmax_ACCY	Accelerometer Y axis	ⓖ	Zmax	256
Zmax_ACCZ	Accelerometer Z axis	ⓖ	Zmax	256
Zmax_EEGL	Forehead EEG Left channel	µV	Zmax	256
Zmax_EEGR	Forehead EEG Right channel	µV	Zmax	256
Zmax_NOISE	Noise channel	Unitless	Zmax	256
Zmax_OXY_IR_AC	Forehead PPG	Unitless	Zmax	256
Zmax_OXY_IR_DC	Oximetry IR DC component	Unitless	Zmax	256

^$Electrode placement: SOMNOscreen: 10–20 system, Mentalab: 10–10 system.
ⓖ: Gravity (m/s²).
^ⓟLabels: 1: Prone, 2: Upright, 3: Left, 4: Right, 5: Upright (head), 6: Supine.
^ⓢLabels: -1: Unscorable, 0: Wake, 1: N1, 2: N2, 3: N3, 4: REM.

Ethical Statements

This study was conducted in accordance with the Donders Centre for Cognitive Neuroimaging (DCCN) blanket approval, protocol ‘Imaging Human Cognition’ (NL45659.091.14), approved by METC Oost-Nederland (2014/288).

Funding

This work was supported by the Swiss National Science Foundation (SNF), a Vici Fellowship from the Dutch Research Council (NWO), and the European Union’s Horizon Europe Programme (HORIZON-MSCA-2021-PF-01-01) through a Marie Skłodowska-Curie Postdoctoral Fellowship (Grant No. 101066123, GlymphoSleep).

Reference Paper

Sikder, N., Verkaar, L., Paltarzhytskaya, A., Acan, S., Bovy, L., Almazova, T., Krugliakova, E., Rosenblum, Y., Krauledat, M., Dresler, M., & Zerr, P. (2025). Wearanize+: A Multimodal Dataset for Evaluating Wearable Technologies in Sleep Research. Center for Open Science. https://doi.org/10.31219/osf.io/dth8y_v3
Read on ResearchGate

Citation

If you use the Wearanize+ dataset, please cite the reference paper as:

@article{sikder2025wearanizeplus,
  title     = {Wearanize+: A Multimodal Dataset for Evaluating Wearable Technologies in Sleep Research},
  author    = {Sikder, Niloy and Verkaar, Lieuwe and Paltarzhytskaya, Anastasiya and Acan, Selin and Bovy, Leonore and Almazova, Tania and Krugliakova, Elena and Rosenblum, Yevgenia and Krauledat, Matthias and Dresler, Martin and Zerr, Paul},
  journal   = {OSF Preprints},
  year      = {2025},
  doi       = {10.31219/osf.io/dth8y_v3},
  url       = {https://doi.org/10.31219/osf.io/dth8y_v3},
}

and the dataset as:

@dataset{niloy_sikder_2025_wearanize_a_multim,
  author    = {Sikder, N.S. and Verkaar, Lieuwe and Paltarzhytskaya, A. (Anastasiya) and Acan, S. (Selin) and Bovy, L. and Tatiana Almazova and Krugliakova, E. (Elena) and Rozenblum, Y.R. and Krauledat, M. and Dresler, M. and Zerr, P.},
  title     = {{Wearanize+: A multimodal dataset (N=130) with wearable-based at-home overnight sleep recordings}},
  year      = 2025,
  publisher = {Radboud University},
  version   = 1,
  doi       = {10.34973/j6jf-9e62},
  url       = {https://doi.org/10.34973/j6jf-9e62}
}

If you use the provided scripts, please cite the GitHub repository as:

@software{sikder2025wearanizeplus,
  author    = {Sikder, Niloy},
  title     = {Wearanize_plus},
  year      = {2025},
  publisher = {Zenodo},
  doi       = {10.5281/zenodo.14892764},
  url       = {https://doi.org/10.5281/zenodo.14892764},
}

People

Principal investigator: Martin Dresler
Data collection: Niloy Sikder, Lieuwe Verkaar, Anastasiya Paltarzhytskaya, Selin Acan, Elena Krugliakova
Sleep-scoring: Leonore Bovy
Data preparation: Niloy Sikder
Supervision: Matthias Krauledat, Paul Zerr, Yevgenia Rosenblum

For questions, comments, queries regarding data access, or interest in collaboration, please contact Martin Dresler.