Information Technology Reference
In-Depth Information
positive reinforcement as de
ned by the operant-conditioning model. In fact, the
aim of the feedback is not to inform the subject about the cognitive strategies that
he/she develops during the learning process, but to directly in
uence the brain
activity (and thus the EEG). Any kind of feedback cannot be used, but only those
with the desired effect on the brain and the cognitive activity in order to enhance the
interaction and the intuitiveness of the system.
Therefore, in the context of sound synthesis, a control strategy involving the use
of explicit or implicit BCI would necessitate different mapping strategies. From a
conceptual point of view, we stress that explicit and implicit BCI involve different
levels of semiotic relation, i.e., the relation between the feedback and the meaning
that the subject attributes to a sound. These two scenarios are discussed in the
following paragraphs.
In the case of explicit BCI as de
ned above, the subject would have to control
his/her cognitive activity to change his/her EEG and thus to control a speci
c
parameter of the sound synthesizer. No semiotic relation between the EEG, the
effect of the synthesized sound on the EEG, and the sound perception is therefore
needed. In other words, the subject has to do something that is not necessarily
related to the semiotics of the perceived synthesized sound to control the synthe-
sizer. More so, an external algorithm is used to interpret the information of interest
extracted from the EEG and to control the electronic device. For example, paying
attention to a target to produce a P300 component that will be processed by the BCI
and arbitrarily associated with a control parameter according to the output of the
algorithm and to a success rate (Fig.
4.2
). This situation that necessitates a certain
expertise acquired during a learning period seems to be quite close to sound syn-
thesis based on the physical or signal modelling of sound vibrations (Sect.
4.3
).
In the case of implicit BCI as de
ned above, the aim would be to enhance the
quality and the intuitiveness of the sound synthesizer by taking into account the
EEG induced by the sound. Thus, a strict semiotic relation between the EEG and
the in
uence of sounds on the EEG should be known. In other words, we need to
understand the neural bases of sound semiotics (
in
Fig.
4.2
) to implement this information in an implicit BCI process dedicated to the
sound synthesizer. We propose to call it
“
electrophysiological data
”
. In this context, the
results obtained from previous EEG experiments presented in Sect.
4.5
constitute
an interesting starting point for the design of such a mapping strategy. This
approach seems to be quite close to sound synthesis based on the modelling of
perceptual effects, which does not necessitate a learning period (Sect.
4.3
). This
intuitive control implies that perceptual and cognitive aspects are taken into
account in order to understand how a sound is perceived and interpreted. As shown
in Fig.
4.2
, a loop is thus designed between perception and action through the
intuitive control of the sound synthesizer (Sect.
4.2
). Implicit BCI offers the
possibility of a second loop, between the sound effect on the EEG and the sound
synthesizer that is likely to optimize the sound effect on both the perceptual
judgment and the Implicit BCI.
“
semiotic-based BCI
”
