High-density surface EMG + EEG recordings of isometric ankle dorsiflexion, imaginary movement and GO/NO-GO tasks

  • Authors: Nina Murks, Leon Kutoš, Aleš Holobar, Matjaž Divjak, Ciaran McGeady, Dario Farina, Matej Kramberger

Keywords: surface high density electromyogram (HDEMG), electroencephalogram (EEG), dataset, tibialis anterior, isometric ankle dorsiflexion, GO/NO-GO task, imaginary movement, corticomuscular coherence, motor unit filter, HybridNeuro

SHARING/ACCESS INFORMATION
License: This dataset is available under the Creative Commons Attribution (CC BY) 4.0 license.

Publicly accessible location of the data: https://dk.um.si/IzpisGradiva.php?lang=eng&id=96398

PID: 20.500.12556/DKUM-96398

Recommended citation for this dataset:
Murks Nina, Kutoš Leon, Holobar Aleš, Divjak Matjaž, McGeady Ciaran, Farina Dario, Kramberger Matej. “High-density surface EMG + EEG recordings of isometric ankle dorsiflexion, imaginary movement and GO/NO-GO tasks (HybridNeuro project)”. System Software Laboratory, Faculty of Electrical Engineering and Computer Science, University of Maribor, Slovenia. Digital library of University of Maribor. https://dk.um.si/IzpisGradiva.php?lang=eng&id=96398

DATA & FILE OVERVIEW
This dataset was prepared in the context of the HybridNeuro project (https://www.hybridneuro.feri.um.si/). It contains a collection of simultaneous EMG and EEG recordings from 20 healthy volunteers. The main purpose of data collection was to assess the sensitivity of corticomuscular coherence detection estimated by the proposed motor unit-based EEG filters and compare it to the classical EEG processing techniques. Surface EMG was recorded from the tibialis anterior muscle during isometric ankle dorsiflexions using 13 x 5 electrode array with 8mm interelectrode distance (model GR08MM1305 from OT Bioelettronica, Torino, Italy). EEG was recorded with a 64 channel EEG cap (model EEG-Cap64 from OT Bioelettronica, Torino, Italy). Both EMG and EEG were recorded simultaneously using the same amplifier (model Quattrocento from OT Bioelettronica, Torino, Italy) at 16 bit resolution and 2048 Hz sampling frequency. For each participant, 25 tasks were recorded with total length of approximately 37 minutes. Each task's data is stored in a separate file in Matlab format (.MAT). Total dataset size is 26.9 GB.

File list:
  • README.txt: this text file containing description of the dataset
  • README.pdf: formatted PDF file containing description of the dataset
  • metadata.xml: Dublin Core metadata in XML format
  • compressed (ZIPed) files "S1.zip" to "S20.zip" containing data folders for 20 different participants, each folder contains the same set of 25 files:
    - block1.mat: data for GO/NO-GO block 1
    - block2.mat: data for GO/NO-GO block 2
    - mvc1.mat: data for maximal voluntary contractions (MVC) 1
    - mvc2.mat: data for maximal voluntary contractions (MVC) 2
    - ramp10_1.mat: data for dorsiflexion at 10% MVC, set 1
    - ramp10_2.mat: data for dorsiflexion at 10% MVC, set 2
    - ramp20_1.mat: data for dorsiflexion at 20% MVC, set 1
    - ramp20_2.mat: data for dorsiflexion at 20% MVC, set 2
    - ramp20_imaginary.mat: data for imagined dorsiflexion at 20% MVC
    - ramp30_1.mat: data for dorsiflexion at 30% MVC, set 1
    - ramp30_2.mat: data for dorsiflexion at 30% MVC, set 2
    - ramp40_1.mat: data for dorsiflexion at 40% MVC, set 1
    - ramp40_2.mat: data for dorsiflexion at 40% MVC, set 2
    - ramp50_1.mat: data for dorsiflexion at 50% MVC, set 1
    - ramp50_2.mat: data for dorsiflexion at 50% MVC, set 2
    - rest_eyes_closed_1.mat: data for rest with eyes closed, set 1
    - rest_eyes_closed_2.mat: data for rest with eyes closed, set 2
    - rest_eyes_closed_3.mat: data for rest with eyes closed, set 3
    - rest_eyes_closed_4.mat: data for rest with eyes closed, set 4
    - rest_eyes_closed_5.mat: data for rest with eyes closed, set 5
    - rest_eyes_closed_6.mat: data for rest with eyes closed, set 6
    - rest_eyes_closed_7.mat: data for rest with eyes closed, set 7
    - sin_1.mat: data for dorsiflexion with sinusoidal force profile, set 1
    - sin_2.mat: data for dorsiflexion with sinusoidal force profile, set 2
    - sin_imaginary.mat: data for imagined dorsiflexion with sinusoidal force profile
  • participant_info.csv: table in CSV format with anonimized participant information (ID, gender, age, height, weight)
  • figure1_protocol_steps.png: figure depicting all the steps of the experimental protocol
  • figure2_go_nogo_ramp with visual_cues.png: figure depicting the required contraction levels and visual cues used during the GO/NO-GO task
  • figure3_go_nogo_visuals.png: figure depicting an example of visual feedback provided for the participant during the GO/NO-GO task
  • figure4_mat_file_content.png: screenshot of the Matlab workspace with all the variables loaded from one MAT file
  • figure5_eeg_channels.png: figure depicting relations between EEG channel labels and EEG data matrix row numbers

All files with extension .MAT are in Matlab format, which is a proprietary format but its specifications are open and there are open source routines available for reading and writing:

  • The GNU Octave software, which is an open source alternative to Matlab, can read and write .MAT files.
  • The Python library Scipy can load MAT files.
  • The Matio project on SourceForge is a C library for reading and writing Matlab MAT files.

METHODOLOGICAL INFORMATION
Description of methods used for collection/generation of data:

Description of methods used for collection/generation of data: 20 healthy participants volunteered to take part in the study, including 12 men and 8 women. The men had a mean age of 34.8 ± 8.9 years, a height of 182.5 ± 10.0 cm, and a body mass of 89.6 ± 13.0 kg. The women had a mean age of 30.3 ± 11.2 years, a height of 171.6 ± 5.3 cm, and a body mass of 72.0 ± 11.1 kg. The experimental protocol was conducted at the Faculty of Electrical Engineering and Computer Science, University of Maribor and was approved by the Medical Ethics Committee of the Republic of Slovenia (No. 0120-340/2023/3). Environment: Participants were seated in an ankle dynamometer (Wise Technologies, Ljubljana, Slovenia) equipped with a force sensor (Type 1, 200 kg, AEP Transducers, Cognento, Italy), with the hips, knees, and ankles flexed at 90° angle. The lateral malleolus was aligned with the dynamometer's axis of rotation. Plantarflexion movement was restricted using the dynamometer's fixation system, which applied a blockade to the thigh just proximal to the knee joint. The foot was additionally secured with straps placed over the metatarsal bones.
Participants were instructed to remain relaxed, with the forearms resting on the dynamometer supports. To minimize signal variability, they were asked to keep their head position constant and to avoid clenching their teeth or squeezing their hands. Visual feedback of the produced force and trigger signals were provided via a monitor positioned in front of the participant at a comfortable viewing height. Preparation: All participants were instructed to abstain from alcohol and caffeine on the day preceding and the day of the experimental protocol. They were also asked to wash their hair without using any hair products one day before testing and to shave the right leg over the tibialis anterior muscle. The tibialis anterior muscle was first identified by asking participants to perform a brief ankle dorsiflexion. The skin over the muscle belly was then prepared using an abrasive paste (Spec Medica Everi). Surface electrodes were prepared by placing an adhesive foam and filling the foam cavities with Ten20 conductive paste. The electrode was positioned over the muscle belly and secured using velcro straps. Two reference straps - one serving as the patient reference and the other as the preamplifier reference - were placed around the ankle. EMG signals were displayed in real time on a computer screen for visual inspection. First quality check was performed with the participant in a relaxed state, during which the signals were visible as flat lines on all channels. Participant then performed a slight dorsiflexion contraction, resulting in clearly visible motor unit action potentials across all channels, which allowed for the observation of signal propagation.
Prior to EEG cap placement, head dimensions were measured from ear to ear and from nasion to inion to ensure accurate positioning. The cap was positioned by aligning the Cz electrode at the midpoint of both measurements. All electrodes were then filled with conductive gel (Neurogel, Spec Medica). Using Sessantaquatro amplifier (OT Bioelettronica, Torino, Italy) impedance was monitored and reduced to below 5 kΩ. The reference electrode (CDES003545, Spes Medica) was placed on the collarbone. EEG signal quality was verified by visual inspection for artifacts, including eye blinks, vertical and horizontal eye movements, and teeth clenching. Signal quality was further confirmed during a resting state with eyes closed by identifying the presence of alpha activity.

Instrument- or software-specific information needed to interpret the data: We used the following acquisition settings for the Quattrocento amplifier:

  • acquisition mode: monopolar
  • sampling frequency: 2048 Hz
  • A/D resolution: 16 bits
  • EMG band pass filter: 10 - 500 Hz
  • EEG band pass filter: 0.7 - 130 Hz
The reference force value was set and visualized using the OTBiolab+ software. This software records the generated values at 16 Hz and then interpolates them (using nearest neighbour interpolation) to the desired sampling frequency (2048 Hz in our case). The same interpolation happens with the values that the software reads from the external force sensor. As a result, both data variables (refRequested, refPerformed) were interpolated from 16 Hz to 2048 Hz and have the same length as the rest of the signals. In addition, the reference force value was also sampled from the AUX input of the Quattrocento amplifier at 2048 Hz (refAcquired, refAux, expressed in both volts and as % of MVC). The EEG signals are stored as 64 x N array (N = total number of samples), with each row containing data for one EEG electrode. The relation between row numbers and EEG channel labels is depicted in Figure 5.
Quality-assurance procedures performed on the data: Signal quality was visually checked and adjusted for each participant at the start of the protocol. For EMG we visually checked for flat signal lines on all electrodes during relaxed state and for clearly visible motor unit action potentials across all channels during a slight dorsiflexion. For EEG we visually checked for presence of clearly visible artifacts (eye blinks, vertical and horizontal eye movements, teeth clenching) and for presence of alpha activity during resting state with eyes closed. No additional quality assurance steps were performed after the acquisition.

DATA-SPECIFIC INFORMATION FOR ALL .MAT FILES:
Variable list:

  • adBits: double number. Resolution of the amplifier's A/D converter for data acquisition, in bits.
  • eeg: 64 x N double array. Recorded EEG channels, one channel in one row. Each row contains N samples (double values) representing EEG voltage. N = total number of samples.
  • eegCap: string. Name of the cap model used for recording EEG.
  • eegHPF: double number. High pass filter cutoff frequency for EEG, in Hz.
  • eegLPF: double number. Low pass filter cutoff frequency for EEG, in Hz.
  • emg: 13 x 5 cell array. Recorded EMG signals: 13 rows x 5 columns. Each cell contains a 1 x N vector of samples (double values) representing EMG voltage. N = total number of samples.
  • emgElectrode: string. Name of the electrode model used for recording high density surface EMG.
  • emgHPF: double number. High pass filter cutoff frequency for EMG, in Hz.
  • emgIED: double number. Inter-electrode distance of the surface electrode grid used to record EMG, in millimeters.
  • emgLPF: double number. Low pass filter cutoff frequency for EMG, in Hz.
  • emgMode: string. Name of the amplifier's EMG signal acquisition mode used.
  • emgMuscle: string. Name of the muscle from which surface EMG was recorded.
  • emgSide: string. On which side limb the EMG electrode was mounted: left or right.
  • fSamp: double number. Amplifier sampling rate for data acquisition, in Hz.
  • refAcquired: The exerted force measured by the force sensor, acquired from OTBioelettronica GUI, sampled at fSamp with amplitude in volts.
  • refAux:The exerted force measured by the force sensor, acquired from AUX, sampled at fSamp expresesed in volts.
  • refColorPoints: vector of double numbers. Labelled samples denoting the colours of the circles that were shown to the participant during the GO/NO-GO task: 1-blue, 2-yellow, 3-orange, 4-red, 5-green. Only available in the block1.mat and block2.mat files.
  • refPerformed: The exerted force measured by the force sensor, acquired from AUX, sampled at 16 Hz, interpolated to fSamp and expresesed in % MVC.
  • refRequested: The requested (target) force level, acquired from OTBioelettronica GUI, sampled at 16 Hz, interpolated to fSamp and expressed in % MVC.