Multimodal signal dataset for 11 intuitive movement tasks from single upper extremity during multiple recording
Ji-Hoon Jeong1, Jeong-Hyun Cho1, Kyung-Hwan Shim1
1Department of Brain and Cognitive Engineering, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 02841, South Korea.
Gigascience
|October 9, 2020
Summary
This study introduces a large, multimodal dataset for intuitive brain-computer interfaces (BCIs), featuring electroencephalography and other biosignals for 11 upper extremity movements to advance BCI technology.
Area of Science:
- Neuroscience
- Biomedical Engineering
- Human-Computer Interaction
Background:
- Non-invasive brain-computer interfaces (BCIs) aim for natural user interaction with robotic systems, but artificial matching hinders effective communication.
- Intuitive decoding in BCIs is crucial for overcoming limitations like few classes and manual command matching.
- Progress in BCIs is hampered by the scarcity of large, uniform datasets for developing advanced decoding methods.
Purpose of the Study:
- To present a comprehensive, large-scale dataset for intuitive brain-computer interface (BCI) research.
- To facilitate the development of advanced BCI decoding algorithms for upper extremity movement tasks.
- To provide a valuable resource for comparing brain activity during real movements versus motor imagery.
Main Methods:
- Collected a large dataset comprising 60-channel electroencephalography (EEG), 7-channel electromyography (EMG), and 4-channel electro-oculography (EOG).
- Recorded data from 25 healthy participants over 3-day sessions, encompassing 11 distinct upper extremity movement tasks.
- Totaled 82,500 trials, including multimodal signals and data from multiple recording sessions for robust analysis.
Main Results:
- Neurophysiological analysis confirmed the dataset's validity, showing distinct sensorimotor de-/activation patterns.
- Observed clear spatial distributions related to both real movements and motor imagery, validating the data's physiological relevance.
- Demonstrated dataset consistency through baseline machine learning classification performance across sessions.
Conclusions:
- The dataset offers multimodal signals for 11 upper extremity movements, supporting research in real movement vs. imagination.
- Enables improvement of decoding performance in brain-computer interfaces (BCIs) through advanced analysis.
- Facilitates the study of inter-session differences, crucial for advancing BCI technology as a Data Note.
Keywords:
brain–computer interfaceintuitive upper extremity movementsmultimodal signalsmultiple sessions
