Electroencephalography (EEG) brain computer interface (BCI) for mental states detection

• Main Author: Aung, Aung Phyo Wai
• Other Authors: Guan Cuntai
• Format: Thesis-Master by Research
• Language: English
• Published: Nanyang Technological University 2023
• Subjects: Engineering::Computer science and engineering::Computing methodologies::Artificial intelligence; Engineering::Computer science and engineering::Computer applications::Social and behavioral sciences
• Online Access: https://hdl.handle.net/10356/166652
• Institution: Nanyang Technological University

Brain Computer Interface (BCI) enables a new dimension for Human Computer Interface, by allowing people to interact directly through their brain signals without conventional pathways. EEG, the most prevalent BCI sensing modality, enables to measure brain activities in various form-factors suitable for application needs. Regardless of shallow or deep modelling, robust decoding of mental states from EEG signals requires calibration tasks to train optimal classiffier models. The lack of ground-truth, only surrogate calibration task, resulted in sub-optimal or poor EEG decoding performance. In this thesis, I proposed generic EEG processing framework covering from calibration, offline modelling to online mental states detection. Then, I investigated attention calibrations under different experiment designs using multiple subjects to understand how different stimuli parameters and tasks influence the attention decoding. Finally, I designed visual search and white noise visual-audio calibration paradigms to further improve the EEG decoding accuracy in attention recognition using wearable EEG devices.