BCI Competition 2008–Graz data set A

Citation Author(s):
University of Graz
École Polytechnique Fédérale de Lausanne (EPFL)
Graz University of Technology
Submitted by:
kai zhou
Last updated:
Wed, 01/17/2024 - 00:38
Data Format:
Research Article Link:
0 ratings - Please login to submit your rating.


This data set consists of EEG data from 9 subjects. The cue-based BCI paradigm consisted of four di erent motor imagery tasks, namely the imag ination of movement of the left hand (class 1), right hand (class 2), both feet (class 3), and tongue (class 4). Two sessions on di erent days were recorded for each subject. Each session is comprised of 6 runs separated by short breaks. One run consists of 48 trials (12 for each of the four possible classes), yielding a total of 288 trials per session.

At the beginning of each session, a recording of approximately 5 minutes was performed to estimate the EOG inuence. The recording was divided into 3 blocks: (1) two minutes with eyes open (looking at a xation cross on the screen), (2) one minute with eyes closed, and (3) one minute with eye movements. Note that due to technical problems the EOG block is shorter for subject A04T and contains only the eye movement condition.

The subjects were sitting in a comfortable armchair in front of a com puter screen. At the beginning of a trial (t = 0s), a xation cross appeared on the black screen. In addition, a short acoustic warning tone was pre sented. After two seconds (t = 2s), a cue in the form of an arrow pointing either to the left, right, down or up (corresponding to one of the four classesleft hand, right hand, foot or tongue) appeared and stayed on the screen for 1.25s. This prompted the subjects to perform the desired motor imagery task. No feedback was provided. The subjects were ask to carry out the motor imagery task until the xation cross disappeared from the screen at t = 6s. A short break followed where the screen was black again.


To utilize the BCI Competition 2008–Graz dataset A for your research or analysis, you can follow these general steps:


1. **Understand the Dataset**: Familiarize yourself with the structure and contents of the dataset. This includes understanding the format of the data files, the type of EEG recordings, the experimental protocol used to collect the data, and any accompanying documentation or metadata.


2. **Data Preprocessing**: Preprocess the dataset as needed for your specific analysis. This may include tasks such as filtering the EEG signals to remove noise, segmenting the data into trials corresponding to different motor imagery tasks, and extracting relevant features from the EEG signals.


3. **Feature Extraction**: Extract features from the preprocessed EEG data that are relevant to your analysis. This could involve computing spectral features, time-domain features, or other features that capture important characteristics of the EEG signals related to motor imagery.


4. **Model Training and Evaluation**: Train machine learning or statistical models using the extracted features and corresponding labels (e.g., left-hand vs. right-hand motor imagery tasks). You can use various classification algorithms such as Support Vector Machines (SVM), Convolutional Neural Networks (CNNs), or other methods commonly used in BCI research.


5. **Performance Evaluation**: Evaluate the performance of your trained models using appropriate metrics such as accuracy, precision, recall, or F1 score. Cross-validation or other validation strategies can help ensure the robustness of your results.


6. **Results Interpretation**: Interpret the results of your analysis in the context of your research goals. This may involve understanding how well your models perform at classifying motor imagery tasks based on EEG signals and drawing conclusions about the feasibility of using this approach in real-world BCI applications.


7. **Documentation and Reporting**: Document your methodology, results, and conclusions thoroughly. This documentation will be important for reproducibility and for communicating your findings to others in the field.


8. **Optional: Algorithm Development**: If your research involves developing novel algorithms for BCI applications, you may need to go beyond using existing methods and develop new algorithms tailored to the specific challenges of the dataset.


It's important to note that the specific details of how you utilize the dataset will depend on the goals of your research, the specific analysis you want to perform, and the tools and techniques you are familiar with.