Segmented EEG Data from 12 Subjects during Left and Right Bicep Flexion

Citation Author(s):
Joshua Myszewski, Rezwan Sheikh, Thomas Reina, Eric Bergendahl, Mohammad Rahman
Submitted by:
Joshua Myszewski
Last updated:
Tue, 11/12/2019 - 10:38
DOI:
10.21227/H2F66Q
Data Format:
MATLAB .MAT
License:
Creative Commons Attribution
1419 Views
Categories:
Biomedical and Health Sciences
Neuroscience
Biophysiological Signals
Brain
Signal Processing
Digital signal processing
Keywords:
Electroencephalography, Machine Learning, Feature extraction
0
0 ratings - Please login to submit your rating.
ACCESS DATASET CITE SHARE/EMBED

Abstract 

Electroencephalography (EEG) signal data was collected from twelve healthy subjects with no known musculoskeletal or neurological deficits (mean age 25.5 ± 3.7, 11 male, 1 female, 1 left handed, 11 right handed) using an EGI Geodesics© Hydrocel EEG 64-Channel spongeless sensor net. All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by the Ethics Committee of the University of Wisconsin-Milwaukee (17.352). The subjects were instructed to flex their right bicep while holding a two pound weight every six seconds based on a visual timer displayed on a computer screen. After one minute the subject repeated this process again with their left bicep. This process was repeated for a total of six trials. 

Usable Data

The electrodes being used were C1, C2, C3, C4, C5, C6, FC1, FC2, FC3, FC4, CP1, and CP2 which are named according to the American Clinical Neuroscience Society’s 10-10 system for electrode naming nomenclature guidelines, a more specific breakdown of how these names correlate to the data is in the instructions for this data set. These were the only electrodes used in this acquisition, data from unlisted electrodes is not usable. 

Pre-Processing

After acquisition the signal data for each trial was then epoched and segmented from -2500 ms to +1500 ms after each time the subject flexed their bicep, resulting in ten distinct signal segments each of left and right bicep flexion data per trial, resulting in ~240 signal segments per subject. 

Instructions: 

 each trial was segmented from -2500 ms to +1500 ms, where 0 is the onset of muscle flexion. 

Each column of the data is a sample with a 1 kHz sampling rate. 

a segment labeled LFLE is a segment of right arm flexion.

A segment labeled RFLE is a segment of left arm flexion.

Included is eample feature extraction data for each subject. 

For reference, Subject 1 was the left handed subject. 

The Feature extraction methods used are as follows:

  • FeatureFA; Power spectral density features from 8-30 Hz from all electrodes.
  • FeatureFS; Power spectral density features from 8-30 Hz from the C4 electrode.
  • CoeffFinalA; Discrete Wavelet Transform Coefficients from Decomposition levels A5, D4, D5 from all electrodes.
  • CoeffFinalS; Discrete Wavelet Transform Coefficients from Decomposition levels A5, D4, D5 from the C4 electrode.

 

Channel Locations

The channel location file provided by the manufacture is included with the dataset, this file type (.sfp) can be used by a variety of programs including EEGLAB and BESI.

For the matlab data each row corresponds to a particular electrode number. These electrode numbers are related to the ACNS 10-10 System as follows, i.e. row 16 corresponds to the C1 electrode):

  • C1 -16
  • C2 - 51
  • C3 - 20
  • C4 - 50
  • C5 - 22
  • C6 - 49
  • FC1 - 7
  • FC2 - 54
  • FC3 - 15
  • FC4 - 53
  • CP1 - 21
  • CP2 - 41

Comments

Thank you for sharing these data.
1. The EEG sensor net has 64 electrods, and only 12 of them are used. Does this mean only these 12 electrodes were used during the experiment? In the EEG data, there are 64 channels. Are the data in other channels except the 12 electrodes uncorrect and useless?
2. Did you filter noise and remove artifacts on these data? Could you introduce the preprocessing steps for this dataset?

Submitted by Deqing Wang on Tue, 05/15/2018 - 06:12

1. Yes, only the 12 electrodes listed are usable from this data.

2. no filtering or artifact removal was conducted on this data set, the only pre processing was epoching the data from -2500 ms to +1500 ms after each event of bicep flexion so that each data set contains the discrete segments. 

I've edited the data set to make this more clear in the instructions. 

Submitted by Joshua Myszewski on Tue, 05/15/2018 - 18:15

Could you provide the channels locations information file, which can be used to draw topography. It is also MATLAB 'mat' file for better. Thank you.

Submitted by Deqing Wang on Thu, 05/17/2018 - 04:45

I added the channel location information that was provided by the manufacturer as a file within the dataset. 

Submitted by Joshua Myszewski on Thu, 05/17/2018 - 20:14

Dear sir,

            I need this data set for exploring EEG signal processing.

 

Thanks & Regards

Tarun Karak

Submitted by TARUN KARAK on Tue, 01/28/2020 - 12:00

I need this data set for exploring EEG signal processing.

Submitted by HE MENGJIA on Wed, 03/09/2022 - 10:28

Dataset Files

LOGIN TO ACCESS DATASET FILES

 

QUESTIONS?