Information Technology Reference
In-Depth Information
12.4
Brain Scanning: Materials and Methods
The brain images were collected using equipment and parameters that are typical in
cognitive neuroscience. The scanner was a Siemens Allegra 3-T head-only scanner
with a head coil. (
field strength.
Contemporary scanners range from 1.5 to 7 T.) A T2-sensitive echo planar imaging
(EPI) pulse sequence was used to obtain blood oxygenation level-dependent
(BOLD) contrasts: TR = 2,000 ms, TE = 30 ms, 36 axial slices, 3
'
T
'
stands for Tesla, a measure of magnetic
3
3mm,
×
×
64
192 mm FOV. That is, each full-brain image took two
seconds to collect, to yield 36 image slices of the brain. Each slice comprised
64
64 matrix in a 192
×
×
64 picture elements, known as voxels or volume pixels. Thus, each image
comprised approximated 150,000 continuously varying voxels.
Subjects heard the second movement of Beethoven
×
s
Seventh Symphony
twice.
The subjects were instructed to attend to the music with their eyes closed. The fMRI
recording began with 30 s without music, then 460 s of Beethoven, then 18 s
without music and
'
finally the same 460 s of Beethoven previously heard. Thus,
each run generated 484 images.
12.5
fMRI Analysis
The raw fMRI scans were first preprocessed following usual procedures for func-
tional neuroimaging. These included correcting for head motion, morphing the
individual brains to conform to a standard anatomical atlas, and spatial smoothing,
which is a procedure that reduces random
fluctuations by calculating a moving
average of each voxel in the context of its spatial neighbours. These preprocessing
steps were implemented using Statistical Parametric Mapping software (Ashburner
et al.
2013
).
Each of the 484 images produced 150,000 voxels, which are very complex for
direct analysis. Instead, the image series were further processed with independent
component analysis, abbreviated as ICA (Stone
2004
). Informally, ICA separates
ensembles of voxels that oscillate in unison. These are uni
ed as supervoxels
representing temporally coherent networks of brain activity. The coloured patches
in Fig.
12.2
are examples of independent components. A total of 25 components
were calculated for the three subjects in the experiment.
In order to select which of these components might be musically signi
cant, the
activity of each component during the
first pass through the Beethoven listening
was compared to that same component during the second pass. If these two seg-
ments of a component time series were correlated, we hypothesised that the activity
was at least partly musically driven, since the stimulus, that is, the music, would be
identical at the corresponding time points in the two passes through the music.
Although 25 independent component time series were identi
ed, only the strongest
15 were selected to in
uence the compositional process. The order of strength of
the selected 15 ICA components is as follows: 25, 15, 14, 8, 5, 10, 11, 18, 6, 2, 4, 1,
