An EEG Dataset for Visual Imagery-Based Brain-Computer Interface.

Area of Science:

• Neuroscience
• Human-Computer Interaction
• Biomedical Engineering

Background:

• Advancements in non-invasive brain-computer interface (BCI) technologies are crucial for enhancing human-machine interaction.
• Visual imagery-based BCI (VI-BCI) offer unique pathways for immersive applications by enabling voluntary brain activation.
• Existing electroencephalogram (EEG) datasets are limited, primarily focusing on motor imagery, thus hindering VI-BCI model development.

Purpose of the Study:

• To present a novel EEG dataset specifically designed for visual imagery tasks.
• To address the data homogeneity issue in current VI-BCI research.
• To facilitate the development of robust VI decoding models and explore related research areas.

Main Methods:

• Recorded EEG data from 22 participants performing visual imagery tasks.
• Tasks involved recognizing ten common images across figures, animals, and objects.
• EEG data acquired using 32-channels at a 1000 Hz sampling rate across two sessions per participant.

Main Results:

• A comprehensive EEG dataset for visual imagery tasks was successfully created.
• The dataset comprises data from diverse cognitive tasks, overcoming limitations of existing motor imagery-focused datasets.
• The resource is suitable for investigating neuroplasticity, adaptive decoding, and cross-subject generalization in VI-BCI.

Conclusions:

• The presented EEG dataset is a valuable resource for advancing VI-BCI research.
• It supports the development of more sophisticated and generalizable decoding algorithms.
• This work facilitates the translation of VI-BCI from laboratory settings to real-world applications.