The brain dataset was supported by the Foundation for Science and Technology of Mongolia and implemented and collected by colleagues from the Electronics Department of the School of Information and Communication Technology at the Mongolian University of Science and Technology. This dataset consists of more than 3294 minutes of EEG recording files from 122 volunteers participating in 4 types of exercises as described below. Each participant performed 4 different tasks during EEG recording using a 14-channel EMOTIV EPOC X system. During the experiment, the data consisted of one- and two-minute recordings, which were manipulated to exclude both eyes-open and eyes-closed cases. Additionally, instructions on how to make measurements have been developed. These tasks include the following 4 types of exercises:
- Lights on, lights off and normal thoughts – EEG_DATA_COLLECTION-LIGHT
- Thinking about moving the box on the screen left, right, up, down or not moving at all -EEG_DATA_COLLECTION-BOX
- Thoughts of moving the left arm, right arm, left leg, right leg or not moving at all – EEG_DATA_COLLECTION-MI
- Display video - VIDEO– EEG_DATA_COLLECTION-VIDEO
Data during each of the above exercises were recorded by reading the EEG signals at 128 sampling steps per second in a 14-channel system. Each recording was saved as a *.CSV file using Emotiv software and Python libraries. According to the international 10-20 system shown in Figure 1 (AF3, F7, F3, FC5, T7, P7, O1, O2, P8, T8, FC6, F4, F8, AF4, and additional 2 CMS/DRL references at P3/P4 electrodes) for EEG recording from 16 electrodes, each participant was presented with the developed instructions, and their consent was obtained.
Figure 1. Connection points of the brain
The numbers below each electrode name indicate the order in which they appear in the recording. The signals on the record are numbered from 1 to 14, and columns 4-17 of the resulting table when opening the *.CSV file contain the values of the electrodes, while column 20 contains the label values. The total uncompressed file size is 6.84 GB.
Measurement instructions
- Make the participant sit still
- Turn off the phone and keep electrical appliances away from the body
- Sterilize and moisten the electrodes and wear them properly
- Go to the EEG_DATA_COLLECTION-LIGHT folder and create a VOLUNTEER folder a. LON- thoughts are taken 5 times. In this case, change the number of the Lonx.csv file b. LOF- thoughts are taken 5 times. In this case, change the number of the Lofx.csv file
- Types of Data Recorded: • Light on • Light off • Normal (presumably under regular lighting conditions)
2.Brain Signal Recording Methods:
• The first 61 entries (Volunteer_1 to Volunteer_61) involved an interrupted entry method, where brain signals were recorded while the volunteers viewed dynamic images. This implies that the recording might have been started and stopped at certain intervals or events.
• The remaining 61 videos (Volunteer_62 to Volunteer_122) were recorded continuously while the volunteers watched dynamic images. This suggests that there was no interruption in the recording process; it was done continuously throughout the viewing session.
The Volunteer_xx folder contains 12 files with the *.CSV extension. The files lof0.csv - lof5.csv contain data for when the light is off and normal thoughts. However, the files lon0.csv – lon5.csv contain data for when the light is on and normal thoughts. These files (lof0.csv and lon0.csv) should not be used for machine learning.
In general, a file with an index of 0 should not be used because it's where the volunteer records the initial method and saves the test recording.
During the experiment, a volunteer participant sits in front of a screen and observes images that alternate between light and darkness for 6-second intervals to evoke specific thoughts. The diagram for creating the lon1.csv file is shown in the figure below. An image of light appears on the screen for 6 seconds, during which the participant thinks "light on". Afterward, there will be a black screen for 6 seconds, during which the participant can think any thought and record it as "normal". Subsequently, the image will flash for another 6 seconds, totaling 60 seconds of recorded video.
Figure 2. Thoughts to "turn on" the interrupted video light
Figure 3. Interrupted recording thought of "turning off" the light
In general, a file with an index of 0 should not be used because this file is where the volunteer writes the initial recording method and saves the test recording. Here, the letter "c" is added to the file name, and "-" signifies "continues".
During the experiment, a volunteer participant sits in front of a screen and observes images that fade in and out of light for 6 seconds to evoke specific thoughts. The diagram for creating the lonc1.csv file is shown in the figure below.
An image of light appears on the screen for 6 seconds, during which the participant forms the thought "light on". Following that, another 6-second interval with the image of light occurs, during which the participant writes down their thought. A total of 60 seconds is recorded in this manner."
Figure 4. Thoughts of "turning on" continuous recording light
The diagram for creating the lofc1.csv file is shown in the figure below. An image of dimming lights appears on the screen for 6 seconds, during which the participant creates the thought 'lights off.'
Afterward, for another 6 seconds, the image with the fading light will appear, and the thought will be written. A total of 60 seconds is recorded.
Figure 5. The idea of "turning off" the continuous recording light
The lofc1.csv file records normal thoughts, indicated by a blank screen for the duration of 60 seconds.