Database Open Access

Simultaneous physiological measurements with five devices at different cognitive and physical loads

Marcus Vollmer Dominic Bläsing Julian Elias Reiser Maria Nisser Anja Buder

Published: June 16, 2020. Version: 1.0.0

When using this resource, please cite: (show more options)
Vollmer, M., Bläsing, D., Reiser, J. E., Nisser, M., & Buder, A. (2020). Simultaneous physiological measurements with five devices at different cognitive and physical loads (version 1.0.0). PhysioNet.

Please include the standard citation for PhysioNet: (show more options)
Goldberger, A., Amaral, L., Glass, L., Hausdorff, J., Ivanov, P. C., Mark, R., ... & Stanley, H. E. (2000). PhysioBank, PhysioToolkit, and PhysioNet: Components of a new research resource for complex physiologic signals. Circulation [Online]. 101 (23), pp. e215–e220.


This open database contains physiological measurements from 13 adult and healthy subjects during a standardized experimental setup. The physiology was examined with five different devices (NeXus-10 MKII, eMotion Faros 360°, SOMNOtouch NIBP, Hexoskin Hx1, Polar RS800 Multi), which, in contrast to other studies, were recorded simultaneously. The experiment included a five minute baseline measurement, five minute walking on the treadmill, a five minute cognitive audio test, and a five minute uphill walking on the treadmill. Measurements included electrocardiography (ECG), photoplethysmography, accelerometry, oxygen saturation, respiration, heart rate, heart rate variability, and RR intervals sampled between 1 Hz and 8000 Hz. The dataset was originally generated during an investigation of functionality, accuracy, and usability of several ECG-measurement devices. This resource included the complete raw data files (EDF and HRM) and merged data with a manual expert annotation of heart beats and annotation files with information about the beginning of each experimental phase.


The database was acquired in December 2017 at the Leibniz-Institute for Labour Research at the Technical University of Dortmund (Germany) where healthy volunteers has been confronted with cognitive and physical stressors to simulate standard tasks for research studies with the aim to investigate of functionality, accuracy, and usability of several ECG-measurement devices. The lead investigator Dominic Bläsing (University Greifswald, Germany) designed the study protocol together with Julian Reiser (Leibniz-Institute for Labour Research, Technical University of Dortmund, Germany); Dr. Marcus Vollmer (University Medicine Greifswald, Germany) was responsible for data processing.

The resource consists of physiological recordings from 13 subjects with a sensor placement that enables simultaneous recording with all five devices used. Standard 3-lead ECGs were measured with NeXus, SOMNOTouch, Faros and Hexoskin. The latter uses textile electrodes, whereas others used standard Ag/Ag-Cl electrodes. Polar recorded identified heart beats by using a chest strap with flat electrodes made of conductive rubber. Beside ECG measurements further physiological parameters were tracked: heart rate, heart rate variability, photoplethysmogram, pulse, oxygen saturation, respiration, respiration rate, accelerometer, activity, and pedal frequency. Data from all five device were merged into a single file based on the common ECG measurement and a manual reference annotation of heart beats was generated. The usage of the dataset is manifold, e.g.:

  1. The dataset provides an unique set of simultaneous measurements to study heart beat detection and segmentation algorithms on noisy signals (especially in active phases, captured by acceleration sensors) with a reliable ground truth annotation based on all available signals.
  2. The dataset offers the opportunity to compare heart rate, HRV and accelerometric  measurements/computations among the devices.
  3. The dataset can be used to study R peak adjustment methods for low-resolution ECGs to correct R peak annotations resulting from low sampling rates. NeXus-10 MKII ECG sensor sampled at 8000 Hz can serve as the ground truth for R peak annotations.
  4. The interplay of actively measured respiration (Hexoskin), acceleration (Faros, SOT, Hexoskin) with Pulse/PPT (SOT) and ECG can be studied and methods can be tested to measure respiration indirectly from the PPT or ECG.


Experimental design and data acquisition

This study sample consisted of a cohort of 13 participants (seven female and six male volunteers) measured in December 2017 at the Leibniz-Institute for Labour Research at the Technical University of Dortmund, Germany. For the purpose of investigating the functionality, accuracy, and usability of several ECG-measurement devices, the participants were equipped with five sensor systems at the same time. The sensor placement is illustrated in Figure 1 of reference [1]. Three clinically certified devices were used:

  • NeXus-10 MKII (Mind Media B.V., The Netherlands)
  • eMotion Faros 360° (Mega Electronics Ltd., Finland)
  • SOMNOtouch NIBP (SOMNOmedics GmbH, Germany)

In parallel, two consumer products were attached:

  • Hexoskin Hx1 (Carré Technologies Inc., Canada)
  • Polar RS800 Multi (Polar Electro Oy, Finland)

After participants were equipped with all measurement devices, four tasks at different cognitive and physical load needed to be performed consecutively:

  1. Five minute resting period baseline while standing upright on the treadmill
  2. Five minute walking on the treadmill at a moderate speed (1.2 m/s)
  3. Five minute cognitive auditive 2-back task while standing still on the treadmill
  4. Five minute uphill walking on the treadmill (15% track inclination, 1.2 m/s)

The experiment was conducted in line with the Declaration of Helsinki and written informed consent was signed. The study was approved by the ethics committee of the University of Greifswald (Identifier: BB 171/17).

Data preprocessing

Measurements had be synchronized in preprocessing steps for data alignment as described in previous work (Vollmer et al. 2019 [1]). It includes a general alignment to synchronize heart beats from a resting period by finding similar patterns in all devices and the correction of incorrect and unsteady sampling frequencies by non-linear resampling to a target sampling frequency of 256 Hz (see [1]). RR intervals from Polar were transformed to an indicator function that had gone through the same synchronization steps such that the Polar information appears as a separate signal. A reference annotation of R peaks has been generated by manually screening the aligned ECG in its entirety. Timestamps from the Faros trigger function were manually inspected and relocated by use of movement sensors and the heart rate increase to fix the beginning of each of the four experimental phases.

Data Description

Raw data

Raw EDF+C files from SOMNOTouch, NeXus, Faros, and Hexoskin devices were anonymized using edf-anonymize and includes the complete recording with several minutes prior to the resting period.

File names are named as follows and are stored in the folder 'raw_data':

  • prefix: proband identifier 'x001' to 'x013'
  • suffix: device abbreviation ('FAROS','NEXUS','SOT','HX')

Recorded data for each device:

  • NeXus-10 MKII: ECG ('Sensor-B:EEG'/'Sensor-B:EKG'), Heart rate variability ('[B] HRV Amp.'), 'EDF Annotations'
  • eMotion Faros360°: ECG ('ECG'), Accelerometer ('Accelerometer_X', 'Accelerometer_Y', 'Accelerometer_Z'), Pushbutton marker ('Marker'), Heart rate variability ('HRV'), Device inside temperature ('DEV_Temperature')
  • SOMNOtouch NIBP: Battery ('Akku'), ECG ('EKG', 'EKG 1', 'EKG2', 'EKG 3', 'aVL', 'aVR', 'aVF'), Oxygen saturation ('SpO2'), Photoplethysmogram ('Pleth'), Activity/Accelerometer ('Aktivitt'), Body position ('Krperlage'), Pulse ('Puls'), 'EDF Annotations'
  • Hexoskin Hx1: ECG ('4113:ECG_I'), Heart rate ('19:heart_rate'), Respiration rate ('33:breathing_rat'), Respiration ('4129:resp_thorac', '4130:resp_abdomi'), Activity/Accelerometer ('49:activity', '4145:accel_X', '4146:accel_Y', '4147:accel_Z'), Pedal frequency ('53:cadence'), 'EDF Annotations'

Sampling frequencies for each sensor, preprocessing information and physical dimensions is stored within the EDF files.

Notice: The sensor name of the recorded ECG in the records x001 to x003 is 'Sensor-B:EEG' of NEXUS' EDF files. We corrected the name during the experiment and also maximized the sampling frequency for later recordings (x004 to x013) from 256 Hz to 8000 Hz.

Recorded data from Polar are stored as device specific hrm file:

  • Polar RS800 Multi: RR intervals in ms

Processed data

After processing the raw files of the five devices, the aligned signals were stored into WFDB-readable files, a manual reference annotation of R peaks was generated and event locations were extracted from Faros pushbutton markers and were manually revised. The following files are stored in the folder 'created_data' with the proband identifier as the prefix:

  • *.dat [WFDB-readable files containing all aligned and resampled signals]
  • *.hea [Header information containing the device and signal name of each signal in the dat-file]
  • *.atr [Manual reference beat annotation of R peaks]
  • *.aux [Faros markers and manually revised markers for the start of each experimental phase]

Aligned and resampled signals of all devices are stored in the dat file at 256 Hz. The first four signals in the dat files contains preprocessed ECGs (trimmed moving average filter, zscore normalization). Signals from Hexoskin were imported from binary wav-files (at the time were Hexoskin has not supported an EDF export). Faros markers and manually revised markers are stored with an annotation file with aux file ending.

Data size

Total record length of raw files is 2577.68 minutes:

Record length of raw files
  Total length Minimum Maximum
eMotion Faros360° 564.58' 33.55' 65.63'
Hexoskin Hx1 526.60' 31.43' 58.37'
NeXus-10 MKII 429.84' 29.83' 39.73'
SOMNOtouch NIBP 557.17' 31.97' 60.97'
Polar RS800 Multi 499.49' 31.11' 55.18'

Total record length of processed data (dat files) is 427.80 minutes. Individual records varying between 29.18' and 39.62'. The number of annotated heart beats (atr) is 42413 (2377 to 4256 beats per record).

Usage Notes

No special software is required to use the data. The WFDB software package offers appropriate functions to for the reuse of the data resource.

Sampling frequencies for each sensor, preprocessing information and physical dimensions can be extracted using the open source function read_edf:

[record, signals, start_date, header, header2] = read_edf('x001_FAROS.edf')

Information on experimental phases can be extracted using rdann:

[ann,~,~,~,~,comments] = rdann('x001', 'aux')

ann =


comments =

  5×1 cell array

    {'FAROS_Marker/Rest'   }
    {'Manual/Walking'      }
    {'FAROS_Marker/2-Back' }
    {'Manual/Running'      }


Dominic Bläsing acknowledges financial support by the Federal Ministry of Education and Research of Germany in the project Montexas4.0 (FKZ 02L15A261). Marcus Vollmer acknowledges travel grants from the German Centre for Cardiovascular Research (DZHK), partner site Greifswald.

Conflicts of Interest

The authors have no conflicts of interest to declare.


  1. Vollmer M, Bläsing D, Kaderali L (2019). Alignment of Multi-Sensored Data: Adjustment of Sampling Frequencies and Time Shifts. 2019 Computing in Cardiology Conference (CinC). doi: 10.22489/cinc.2019.031


Access Policy:
Anyone can access the files, as long as they conform to the terms of the specified license.

License (for files):
Creative Commons Attribution 4.0 International Public License

Corresponding Author
You must be logged in to view the contact information.


Total uncompressed size: 1.7 GB.

Access the files

Visualize waveforms

Folder Navigation: <base>
Name Size Modified
ANNOTATORS (download) 136 B 2020-04-21
LICENSE.txt (download) 14.5 KB 2020-06-04 (download) 1.9 KB 2020-04-09
RECORDS (download) 1.7 KB 2020-04-22
SHA256SUMS.txt (download) 10.7 KB 2020-06-16
sensor_placement.png (download) 75.6 KB 2020-04-14