Loading...

​​Brain-Computer Interfaces Research Starter Guideline

​​

​​Table of Contents

• ​​Fundamentals

• ​​Conventional Methods

• ​​Deep and Hierarchical Methods

​​

​​Fundamentals

• ​​We focus on studying non-invasive electroencephalogram-based brain-computer interfaces.

• ​​EEG: Electroencephalogram (EEG) is an electrophysiological monitoring method to record electrical activity of the brain. 

• ​​Non-invasive brain-computer interfaces (BCIs) are categorized into two types: evoked vs. spontaneous BCIs.

• ​​Evoked BCIs exploit evoked potentials like P300 or steady-state visual evoked potential (SSVEP), mostly induced by an external stimulus.

• ​​Spontaneous BCIs focus on internal cognitive processes such as event-related (de)synchronization (ERD/ERS), like motor imagery (MI).

• ​​For more details, please read Ch. 1 in a book, “EEG Signal Processing.”

• ​​Importantly, please organize read papers in Google spreadsheet.

​​

​​Conventional Methods

• ​​Lotte et al. “A Review of Classification Algorithms for EEG-based Brain-Computer Interfaces: A 10 Year Update”

• ​​Motor Imagery Classification Methods

• ​​Feature Extraction Methods

• ​​Common Spatial Patterns (CSP)

• ​​Filter Bank Common Spatial Patterns (FBCSP)

• ​​Bayesian CSP

• ​​Classifier

• ​​Linear Discriminant Analysis (LDA)

• ​​Support Vector Machine (SVM)

• ​​SSVEP Classification Methods

• ​​Canonical Correlation Analysis (CCA)

• ​​Further, please read Ch. 11.3.3 in a book, “Signal Processing and Machine Learning for Brain-Machine Interfaces.”

​​

​​Deep and Hierarchical Methods

• ​​For MI-based BCI, Schirrmeister et al. proposed various methods.

• ​​Shallow ConvNet

• ​​Deep ConvNet

• ​​For looking deeper, read:

• ​​DeepSleepNet

• ​​ResNet

• ​​Network in Network

• ​​Inception and Xception

​​

​​