Information Technology Reference
In-Depth Information
de
cit disorder with hyperactivity (Micoulaud-Franchi et al.
2011
) or epilepsy
(Micoulaud-Franchi et al.
2014
).
4.6.2 Implicit BCI
In contrast with explicit BCI, the implicit BCI is not based on the principle of
operant conditioning. The feedback in implicit BCI is used to optimize the inter-
action with an external device by directly modulating the brain activity and the
cognitive activity of the subject (George and L
cuyer
2010
). Implicit BCI does not
require a learning period. In practice, the subject does not have to try to control
intentionally his EEG. The EEG signal is recorded, processed in real time to extract
the information of interest (e.g. power spectral EEG or ERP) corresponding to the
subject
é
s cognitive activity, and transmitted to the external electronic device to
modulate and optimize the interaction between the device and the user.
This con
'
guration
fits with some non-traditional neurofeedback therapeutics that
do not require speci
c cognitive tasks and are supposed to directly modulate the
brain activity of the subject in order to optimize brain dynamics, although this
remains largely hypothetical. Thus, unlike traditional neurofeedback approaches
presented in the previous section, these non-traditional neurofeedback approaches
have a very low level of therapeutic and clinical evidence (Micoulaud-Franchi et al.
2013
).
4.6.3 Towards an Intuitive Control Using Semiotic-Based BCI
From the two approaches inspired by previous theoretical frameworks from cog-
nitive neuroscience (Sect.
4.2
), we propose a prospective view on a sound synthesis
control strategy based on BCI. We re
ect on whether EEG BCI would be helpful to
increase the intuitiveness of control with the sound synthesizer. For a didactic
perspective, we suggest to describe explicit and implicit BCI, respectively, from the
representational-computational and from the enactive points of view.
We stress that in the explicit BCI, the user controls the external electronic device
(positive reinforcement) as if it was an external object. In some way, there is a gap
between the information of interest extracted from the recorded EEG activity and
the positive reinforcement. The information feedback could be given to the subject
by any kind of signal. The positive reinforcement mainly is useful for the learning
process and for determining a success rate and is close to an error function
(Sect.
4.4.1
). We think that in many cases, explicit BCI does not permit to create
recurrent sensorimotor patterns (from the enactive point of view) that enable action
to be guided by the direct perception of the stimulus, which could be a limitation in
the intuitiveness of BCI controllability.
We stress that in the Implicit BCI, the user and his/her brain is involved in an
enactive process. In some way, there is a direct link between the information of
interest extracted from the recorded EEG and the feedback. This feedback is not a
