Information Technology Reference
In-Depth Information
10.3.2 Computer-Orientated Systems
In a computer-orientated system, the user adapts to the functions of the computer.
The computer model stays
fixed, and the success of the system relies on the ability
of a user to learn how to perform control over musical events. A performance piece
conceived in 2011 by BioMuse Trio, called
Music for Sleeping and Waking Minds
,
uses this approach. The responses of performers
xed
musical parameters. Controlling their state of mind (or sleep in this case) affects
control over the music. Attempts to control musical systems with alpha waves
using, a technique called neurofeedback, have mostly fallen into this category as the
user is required to learn how to control their EEG in certain ways in order to
produce desired sonic results.
'
brainwaves are mapped to
10.3.3 Mutually Oriented Systems
Mutually orientated systems combine the functions of both user and computer
orientation whereby the two elements adapt to each other. This was the approach
used in Eaton
'
s
The Warren.
Here, the system requires the user to learn how to
generate speci
c commands and features mappings that adapt depending on the
behaviour of the user.
The majority of BCMIs fall into the category of computer-orientated systems.
This allows for
fixed parameters to be built that respond to known user brain
responses. The use of mutually orientated systems allows for two useful things.
Firstly, more sophisticated algorithms derived from EEG behaviour can be mapped
onto music. As the system learns the EEG behaviour of a subject over time, this
information can be used in series with primary mappings and in parallel through
embedding deeper secondary mappings. Secondly, a system where user and com-
puter adapt together increases the likelihood of obtaining accurate EEG as both
elements are effectively calibrated to optimise the system performance.
10.3.4 Brainwave Data for BCMI
There are two types of EEG data used in the systems discussed in this chapter:
event-related potentials (ERPs) and spontaneous EEG. ERPs are
fluctuations of
EEG measured in response to events triggered by external stimuli. ERP data are
time locked to stimulus and are recognised as positive or negative amplitude
de
ections. ERPs are categorised by their response time post-stimuli and are
associated with brain processing of event expectation and perception.
Systems monitoring spontaneous EEG look at ongoing EEG data, often across
multiple frequencies for patterns or trends that correspond to speci
c brain activ-
ities. This can also be time locked to external stimuli, and if so, windows of
corresponding data are captured for analysis.
