Information Technology Reference
In-Depth Information
the two modalities, one has to decide at which level the individual modalities
should be fused. A straightforward approach is to simply merge the features from
each modality (feature-level fusion). The alternative is to fuse the features at the
decision level based on the outputs of separate single classi
ers (decision-level
fusion) or to use a hybrid method. This issue needs further investigation and new
effective fusion strategies should be developed.
6.7
Brain Responses to Music
It is generally agreed that emotional processing involves activation of wide net-
works of central nervous system (Blood et al.
1999
; Blood and Zatore
2001
). In that
vein, several studies have used brain activity measures to explore emotional
responses during music listening (Koelsch
2005
; Koelsch et al.
2006
). An approach
taken to examine emotional processing in brain are EEG experiments during music
listening. Davidson (
1988
) suggested that the left frontal area is involved in the
experience of positive emotions such as joy and happiness. In contrast, the right
frontal region is involved in the experience of negative emotions such as fear,
angry, and sadness. There are, however, results providing evidence that frontal
asymmetry is related to motivational direction rather than emotional valence
(Harmon-Jones and Allen 1998). Using EEG measurements, Davidson et al. (
1990
)
found substantial evidence for the asymmetric frontal brain activation. Since then,
several EEG studies using various sets of musical stimuli provided support for the
hemispheric specialization hypothesis for emotional valence. That is, musical
stimuli which are considered positive or negative in valence, elicited asymmetric
frontal EEG activity. Schmidt and Trainor (
2001
) investigated patterns of EEG
activity induced by musical excerpts in a group of undergraduates. They found
greater left and right frontal activity during music listening to pleasant and
unpleasant music. Furthermore, they were the
first to show that the overall power of
frontal activity distinguishes the intensity of musical emotion. Moreover, faster
tempi and the major mode produced greater responses in the left hemisphere,
whereas slower tempi and minor mode were associated with greater responses in
the right hemisphere (Tsang et al.
2001
). Sammler et al. (
2007
) investigated elec-
trophysiological correlates during the processing of pleasant (consonant) and
unpleasant (dissonant) music using both heart rate and EEG measurements. In the
EEG, they found an increase of frontal midline theta power for pleasant music in
contrast to unpleasant music. Altenmueller et al. (
2002
) presented musical excerpts
from four different genres to students who provided judgments for each excerpt.
Positively valenced stimuli elicited bilateral fronto
temporal activations predomi-
nantly of the left hemisphere, whereas negatively valenced stimuli elicited bilateral
activations predominantly of the right hemisphere. Females showed greater
valence-related differences than males did. In consequence, the frontal temporal
lobes seem to be involved in emotional evaluation and judgment rather than the
perceptual analysis of emotional information (Heilman
1997
). Flores-Gutierez et al.
-
