Information Technology Reference
In-Depth Information
easily discriminated against another. Again, results were very positive with the
largest distinction recorded between auditory imagery and spatial imagery.
Not only did this latter test minimise the number of electrodes for accurately
reading overall EEG, thus likely reducing interference and preparation time, but it
also narrowed the gap between BCMIs and EEG techniques within other BCI
fields
such as assistive technologies, where patients already accustomed to motor imagery
would need less training.
Importantly, these experiments indicated that subjective choices can elicit
expected brain responses. Unlike the previous experiments with auditory stimuli,
they do not rely on the subject
s expectation or perception of stimuli but allow for a
user to impose a subjective decision that has the possibility of becoming separate
from the meaning within the music being used. This is a crucial step in the leap
towards BCI control of music through neurofeedback.
This element of subjective control aside, the systems discussed in this section
rely on an intrinsic link between the stimuli and resultant music. They are in effect
one and the same, creating the ultimate feedback loop. Attempting to implement
such a BCMI as an interoperable interface with musical systems outside brain-
related activity becomes extremely dif
'
cult when using auditory stimuli as the
driver for generating EEG. Issues of attention become prominent when a user is
required to focus on speci
c sounds to generate EEG, which then have a separate
effect as they produce or affect unrelated music as the result. BCMIs designed
speci
cally for utilising these features, such as the b - soloist and b - conductor ideas,
rely on the use of the stimuli as the driver and the receiver of neurofeedback.
However, to design any systems outside such a tight link, the element of neuro-
feedback can become confused and even lost, as the cause is disengaged from the
effect. To counter this, a compromise in neurofeedback loss is made, heavy user
training is required to reassign unrelated mappings through decision making, or as
noted by Miranda et al. ( 2003 ), higher levels of intelligence are imparted in com-
positional algorithms detracting from cognitive musical control.
10.9
Towards BCI Control of Music
Currently, there are a number of systems offering EEG detection linked to musical
functions commercially available, e.g. WaveRider, g.tec, Emotiv, to name but three.
These systems provide various methods of processing raw EEG that can be mapped
to musical engines, in effect providing the hardware for a BCMI system. At the time
of publication, there are few systems that allow for mapping EEG directly to
musical programs without direct access to APIs and designing bespoke tools;
however, the Emotiv system offers the ability to map raw EEG into open sound
control (OSC) data, and software such as Brainbay and WaveRider provides tools
for mapping EEG to MIDI. We note however that the prices of EEG equipment can
differ enormously. The reader should exercise caution here because cheaper
equipment does not always match the quality of more expensive ones; EEG requires
good quality electrodes and decent ampli
ers.
 
Search WWH ::




Custom Search