BrainSpectralAnalytics/EEG_Mental_Arithmetic_Task

EEG During Mental Arithmetic Tasks

Cognitive Workload Assessment through EEG Recordings

Overview

This dataset contains EEG recordings from 36 healthy participants captured during resting state and mental arithmetic task performance. The recordings provide insights into brain activity dynamics under cognitive workload, specifically during serial subtraction tasks.

Key Features:

• 36 subjects with paired recordings (rest and task states)
• 23-channel EEG system (International 10/20 placement)
• 60-second artifact-free segments per recording
• Performance-based classification (good vs poor performers)
• Preprocessed data with ICA artifact removal

Dataset Summary

The dataset comprises EEG measurements recorded using a Neurocom EEG 23-channel system with silver/silver chloride electrodes positioned according to the International 10/20 scheme. Participants performed mental arithmetic tasks involving serial subtraction of two-digit numbers from four-digit numbers while their brain activity was monitored.

The study design allows for investigation of cognitive workload effects on brain dynamics, making it valuable for research in neuroscience, cognitive psychology, brain-computer interfaces, and machine learning applications in cognitive state classification.

Dataset Structure

Data Organization

The dataset is organized in two folders:

• data/edf_files.zip: Original EDF (European Data Format) files compressed in zip format
• data/csv/: Converted CSV files for easier analysis

Each subject has two EEG recordings:

• Baseline (_1 suffix): 60-second resting state recording with closed eyes
• Task (_2 suffix): 60-second recording during mental arithmetic performance

Available Formats

• EDF Format: Original neurophysiological signal format (data/edf/)
• CSV Format: Converted tabular format for easy analysis (data/csv/)

Important Note: The data provided in both formats represents the original recordings after preprocessing. The only transformation applied was the conversion from EDF to CSV format for accessibility purposes. The preprocessing included high-pass filtering at 30 Hz, 50 Hz notch filtering, and ICA artifact removal as described in the original publication.

Subject Groups

Participants are divided into two performance groups based on task completion:

Group	Label	Subjects	Mean Operations (4 min)	SD
Good Performers	G	24	21 operations	7.4
Poor Performers	B	12	7 operations	3.6

Data Fields

EEG Recordings:

• 23 EEG channels following International 10/20 system
• Sampling rate: Variable (typically 500 Hz)
• Duration: 60 seconds per recording
• Format: EDF (original) and CSV (converted)

Subject Information (subject-info.csv):

• Subject ID: Unique identifier
• Gender: Male/Female
• Age: Age in years
• Job: Occupation/student status
• Date: Recording date
• Count Quality: Performance group (0 = Poor, 1 = Good)

Signal Processing

Recordings underwent preprocessing with a high-pass filter at 30 Hz cutoff frequency and a 50 Hz notch filter, with Independent Component Analysis used to remove artifacts from eye movements, muscles, and cardiac activity.

Collection Methodology

Experimental Protocol

1. Adaptation Phase: 3 minutes for participants to acclimate to experimental conditions
2. Resting State: 3 minutes of EEG recording with closed eyes (last 3 minutes used)
3. Mental Task: 4 minutes of serial subtraction (first minute used in dataset)

Task Description

Participants performed serial subtraction tasks where they were given a 4-digit minuend and 2-digit subtrahend orally and asked to mentally calculate the result without speaking or using finger movements.

Example Task: 3141 - 42 = ?

Participant Criteria

Inclusion:

• Normal or corrected-to-normal visual acuity
• Normal color vision
• No clinical manifestations of cognitive impairment
• No verbal or non-verbal learning disabilities

Exclusion:

• Use of psychoactive medication
• Drug or alcohol addiction
• Psychiatric or neurological complaints

Equipment

• System: Neurocom EEG 23-channel (XAI-MEDICA, Ukraine)
• Electrodes: Silver/silver chloride
• Reference: Interconnected ear electrodes
• Environment: Dark soundproof chamber

Common Applications

• Cognitive workload assessment
• Mental state classification
• Brain-computer interface development
• Cognitive neuroscience research
• Feature extraction from EEG signals
• Performance prediction models
• Attention and focus studies
• Educational neuroscience applications

Important Considerations

⚠️ Preprocessed Data: All recordings are artifact-free segments after ICA processing

⚠️ Performance Grouping: Classification based on accuracy within 20% of correct answer

⚠️ Limited Duration: Each recording is only 60 seconds

⚠️ Single Task Type: Only serial subtraction tasks included

⚠️ Sample Size: 36 subjects total (consider for statistical power)

Citation

Original Publication:

Zyma I, Tukaev S, Seleznov I, Kiyono K, Popov A, Chernykh M, Shpenkov O. Electroencephalograms during Mental Arithmetic Task Performance. Data. 2019; 4(1):14. https://doi.org/10.3390/data4010014

BibTeX:

@article{zyma2019eegmat,
  author = {Zyma, Igor and Tukaev, Sergii and Seleznov, Ivan and Kiyono, Ken and Popov, Anton and Chernykh, Mariia and Shpenkov, Oleksii},
  title = {Electroencephalograms during Mental Arithmetic Task Performance},
  journal = {Data},
  volume = {4},
  year = {2019},
  number = {1},
  article-number = {14},
  issn = {2306-5729},
  doi = {10.3390/data4010014},
  url = {https://www.mdpi.com/2306-5729/4/1/14}
}

PhysioNet Citation:

@article{physionet,
  author = {Goldberger, Ary L. and Amaral, Luis A. N. and Glass, Leon and Hausdorff, Jeffrey M. and Ivanov, Plamen Ch. and Mark, Roger G. and Mietus, Joseph E. and Moody, George B. and Peng, Chung-Kang and Stanley, H. Eugene},
  title = {{PhysioBank, PhysioToolkit, and PhysioNet: Components of a New Research Resource for Complex Physiologic Signals}},
  journal = {Circulation},
  volume = {101},
  number = {23},
  pages = {e215--e220},
  year = {2000},
  doi = {10.1161/01.CIR.101.23.e215}
}

Acknowledgments

• Data Collection: Igor Zyma, Sergii Tukaev, Ivan Seleznov
• Institution: National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute", Department of Electronic Engineering
• Ethics Approval: Bioethics Commission of Educational and Scientific Centre "Institute of Biology and Medicine", Taras Shevchenko National University of Kyiv
• Dataset Hosting: PhysioNet
• This dataset is made available on Hugging Face for easier accessibility and integration with modern machine learning workflows

License

This dataset is available under the Open Data Commons Attribution License v1.0 (ODC-By). You are free to share, create, and adapt the data as long as you attribute the original creators.

Additional Resources

• PhysioNet Dataset Page
• Original Paper (Open Access)
• Kaggle Version
• International 10/20 System Information

Contact

For questions about the original dataset:

• Contact: Ivan Seleznov ([email protected])
• Institution: National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute"

For questions about this Hugging Face version, please open an issue in the repository.