Perceived Quality in Game Audio - Game Sound Technology and Player Interaction: Concepts and Developments - page 157

ventral streams are also activated. Although the
involved neurons are locally distinct, we perceive
one singular object, not separate rolling , red color ,
or round shape .
Until now, it is unclear how the processing of
multiple characteristics of a single object is orga-
nized. A number of theories have been suggested
to explain this binding problem, and the explora-
tion of binding in the visual system has become a
heavily discussed topic. According to Goldstein
(2002), the most prominent theory, suggested by
Singer, Engel, Kreiter, Munk, Neuenschwander,
and Roelfsema (1997), assumes that visual ob-
jects are represented by groups of neurons. These
so-called cell-assemblies are activated jointly,
producing an oscillatory response. This way,
neurons belonging to the same cell-assembly can
synchronize. Whenever the reaction to stimuli
is synchronized, this means that the respective
cortical areas are processing data coming from
one single object or context.
Yet, this binding by synchrony theory has
left doubts with respect to the interpretation and
processing of the synchrony code. For example,
Klein, König, and Körding (2003) postulate that
“many properties of the mammalian visual system
can be explained as leading to optimally sparse
neural responses in response to pictures of natural
scenes” (p. 659). According to Goldstein (2002),
many others argue that binding can be explained by
(selective) attention. Attention is discussed below.

stimuli perceived by our senses. There is actu-
ally no need to consciously further evaluate the
different percepts in terms of relevance, because
they usually complement (and not contradict)
each other.
In order to actually verify any naturally given
order of significance of the perceived stimuli, it is
necessary to present the human perceptual system
with contradictory sensory information and see
what the generally dominating modality is—if
there is any. There have been a number of scien-
tific efforts to explain in a perceptual relevance
model how the human perceptual system weighs
the different contradicting percepts.
Two such models have been proposed to de-
scribe how perceptual judgments are made when
signals from different modalities are conflicting.
One of these models suggests that the signal that
is typically most reliable dominates the competi-
tion completely in a winner-take-it-all fashion: the
judgment is based exclusively on the dominant
signal. In the context of spatial localization based
on visual and auditory cues, this model is called
visual capture because localization judgments
are made based on visual information only. The
other model suggests that perceptual judgments
are based on a mixture of information originating
from multiple modalities. This can be described as
an optimal model of sensory integration which has
been derived based on the maximum-likelihood
estimation (MLE) theory. This model assumes
that the percepts in the different modalities are
statistically independent and that the estimate of a
property under examination by a human observer
has a normal distribution. In engineering literature,
the MLE model is also known as the Kalman Filter
(Kalman & Bucy, 1961).
Battaglia, Jacobs, and Aslin (2003) report that
several investigators have examined whether hu-
man adults actually combine information from
multiple sensory sources in a statistically optimal
manner (that is, according to the MLE model).
They explain:

Dominance of single Modalities

Very often the dominance of visual stimuli over
other modalities is accepted naturally as a given.
In fact, looking at our everyday experiences we
might be inclined to accept this posit without fur-
ther discussion: because “seeing is believing”, we
often think that we tend to trust our eyes more than
the other senses. Yet, this appraisement is often
due to the fact that in the real world we seldom
have to face contradictions in the multi-modal