EEG-based BCI has been an active research field for over three
decades. With BCI, neural signals could be used to serve
non-muscular communication and control functions that connect
man and machine. Thus, BCI has been applied to build tools that help
people with impaired motor and cognitive functions. There are several
types EEG-based BCI depending on types of neural signals of interest.
These include event-related potentials (ERPs), steady-state visually evoked potentials (SSVEPs), and other oscillatory signals. The drastic development in computing technology and computational methods makes BCI one of front-runner research fields in science and technology that could produce real impacts to our society.
So far, most EEG-based BCI has used commands to connect between man and machine without bi-directional feedbacks (open-loop EEG-based BCI). Achieving closed-loop EEG-based BCI (Fig. 1) still faces many tremendous challenges. Our Bio-inspired Robotics and Neural Engineering (BRAIN) laboratory at VISTEC are developing
closed-loop EEG-based BCI systems that will allow human users to continuously and smoothly interact with and control machines (e.g. robot arms) with high speed and accuracy. To achieve these research goals, we take approaches as summarized in Fig. 2.
By incorporating fundamental knowledge from cognitive science and
neuroscience, we can design behavioral tasks and use patterns of neural
correlates underlying sensory processing and core cognitive functions,
such as attention, working memory, and cognitive control to build various
EEG-based BCI applications. This shows how basic research (Fig. 3) in neuroscience will advance the development and expansion of BCI-based applications.
SIGNAL PROCESSING AND RECOGNITION METHOD
Neural responses could be analyzed by traditional signal
processing techniques (e.g. ensemble averaging, digital
filtering, Fourier transform, wavelet transform) and
multivariate pattern recognition methods (e.g. principle
component analysis, independent component analysis,
support vector machine, linear discriminant analysis). Fig. 4
is an example of simple response signals. Moreover, we
also use other advanced learning algorithms (e.g., deep
learning with long short-term memory, regression,
In additional to basic research in our lab, we are developing
many closed-loop BCI applications. These include control
systems for computer applications, robotic arms, prosthesis
and exoskeleton. Moreover, we are prototyping in-ear EEG
electrodes and around-ear EEG electrodes as alternatives for conventional scalp EEG electrodes, whic could be mor