Tracks

Competition

Dataset

Performance measurement

Track 1

Few-shot EEG learning

OpenBMI

Classification accuracy (%)

Track 2

Microsleep detection from single-channel EEG

Sleep-EDF

Kappa value

Track 3

Imagined speech classification

Release

Classification accuracy (%)

Track 4

Upper-limb movements decoding in a single-arm

Release

Classification accuracy (%)

Track 5

EEG(+Ear-EEG)-based ERP detection during walking

Release

AUC

Description of Competition Datasets

Track 1

- Measuring performances for few-shot learning using EEG signals
- Two-class (Right hand and left hand) motor imagery tasks
- Training: 10 trials/class, Validation: 10 trials/class, Test: 10 trials/class

Track 2

- Detecting microsleep for driver drowsiness based on subject-independent strategy
- Five sleep stages from single EEG channel
- Training: Sleep-EDF, Validation & Test: drowsiness DB with Karolinska Sleepiness Scale

Track 3

- Measuring classification performances of multi-class imagined speech
- Five-class imagined speech words/phrases for basic communication
(e.g., ‘Hello’, ‘Help me’, ‘Stop’, ‘Thank you’, and ‘Yes’)
- Training: 60 trials/class, Validation: 10 trials/class, Test: 10 trials/class

Track 4

- Measuring performances for session-to-session transfer problem
- Three grasp motions from a single-hand (‘Cylindrical’, ‘Spherical’, and ‘Lumbrical’) 
- Training: day 1, Validation: day 2, Test: day 3

Track 5

- Measuring performances for ambulatory BCIs to solve signal distortion problem
- Measure: Scalp-EEG/Ear-EEG/IMU, Paradigm: ERP, Walking speed: 1.6 m/s
- Training: 180 trials, Validation: 60 trials, Test: 60 trials

Schedule for 2020 International BCI Competition

Events

2020

2021

Mar.

Apr.

May

Jun.

Jul.

Aug.

Sep.

Oct.

Nov.

Dec.

Jan.

Feb.

Competition setup, Announcement

Release training and validation dataset

Code submission

Release test dataset and evaluation

Results notification

Award ceremony