Information Technology Reference
In-Depth Information
4.6
Towards a Semiotic-Based Brain Computer Interface (BCI)
BCIs provide a link between a user and an external electronic device through his or
her brain activity, independently of the voluntary muscle activity of the subject.
Most often BCIs are based on EEG recordings that allow for non-invasive mea-
surements of electrical brain activity. As substitutional devices, BCIs open inter-
esting perspectives for rehabilitation, reducing disability and improving the quality
of life of patients with severe neuromuscular disorders such as amyotrophic lateral
sclerosis or spinal cord injury (Wolpaw et al.
2002
). Such interfaces, among many
other possibilities, enable patients to control a cursor, to select a letter on a com-
puter screen, or to drive a wheelchair. In addition to medical and substitutional
applications, BCIs as enhancing devices can be used with healthy subjects. For
example, in the
field of video games, BCIs could capture the cognitive or emotional
state of the user through the EEG to develop more adaptive games and to increase
the realism of the gaming experience (Nijholt
2009
). To date, two approaches to
BCI could be highlighted:
“
explicit (or active) BCI
”
and
“
implicit (or passive) BCI
”
(George and L
cuyer
2010
). These two classes of BCI could be linked with the two
approaches inspired from the paradigms of cognitive science (described in
Sect.
4.2
) and the two approaches for sound synthesis (described in Sect.
4.3
).
é
4.6.1 Explicit BCI
The explicit BCI is based on the principles of
operant conditioning
, the basic
learning concept in experimental psychology, which assumes that the probability of
occurrences of an animal or human behaviour is a function of a positive or negative
reinforcement during the subject
s learning process (Micoulaud-Franchi et al.
2013
). Thus, the explicit BCI requires a learning period (George and L
'
cuyer
2010
). In practice, the subject intentionally tries to control his/her cognitive activity
to change his/her EEG activity and control an external electronic device. The EEG
signal is recorded, processed in real time to extract the information of interest (e.g.
spectral power EEG, slow cortical potential or ERP). This information is related to a
cognitive activity that the subject intentionally produces. This information is further
transmitted to the external electronic device using speci
é
c mapping that leads to the
control of the device in the desired direction. The positive reinforcement (and the
success rate) is determined by the capacity of controlling the external electronic
device to achieve a given task.
This configuration fits with traditional neurofeedback therapeutics where the
subject learns to intentionally control EEG through visual or auditory positive
reinforcement, without any control of external device (Micoulaud-Franchi et al.
2013
). In this context, the positive reinforcement could be an increase of a number
of points, an advance of an animation on a computer screen, or a modi
cation of a
sound. When the EEG is related to symptoms of a disease, it has been shown that
neurofeedback techniques can have a therapeutic effect, as is the case with attention
