fMRI Data Acquisition and Analysis
Brain activation of participants was measured using a 3 Tesla magnetic
resonance scanner (Magnetom Trio A, Tim System, Siemens Medical Systems;
Erlangen, Germany). For the functional analysis, 110 volumes of
echo-planar-images were acquired using a T2*-weighted sequence (TE = 30
ms, flip angle = 90°, matrix = 64 x 64, FOV = 192 mm, TR = 2.7 s). Each
volume comprised 40 axial slices with 3 mm thickness and an interslice
space of one millimeter creating a voxel size of 3 x 3 x 3 mm. In
addition, an anatomical scan with high resolution was acquired using a
T1-weighted MPRAGE sequence with a voxel size of 1 x 1 x 1 mm. For data
preprocessing, the first four volumes were discarded to secure
steady-state tissue magnetization. Preprocessing and data analysis was
performed using the software SPM8 (Wellcome Trust Centre for
Neuroimaging, University College London). Data were realigned to
minimize effects of body movement. Realigned data then were normalized
and transposed to the Talairach space (Talairach and Tournoux, 1988)
using the anatomical image that was co-registered with a T1-template and
the mean image of the realigned data. After normalization, the images
were smoothed with a Gaussian kernel of 8 mm full width at half-minimum
(FWHM). Preprocessed data were used for first level analysis where the
onset and duration of the decision-screens were taken to assess the
activation during the three different stimulus conditions (high, low,
and mixed risk). In addition, we added two parameters in the parametric
SPM model for every stimulus condition for the potential gain and the
potential loss. These parameters were calculated as the sum of the
potential wins the sum of potential losses of these two options. For the
first-level analysis, contrasts of predictor estimates (beta-weights)
were defined for each risk level (i.e.: high risk, mixed risk, and low
risk). The expected blood oxygen level-dependent (BOLD) signal changes
were modeled using a canonical hemodynamic response function. The
resulting contrast images of each participant were used for second-level
ANOVA calculations (group level). At the end we got group level
predictor estimates (beta-weights) which we used for calculating
contrasts for comparison between risk and ambiguity, mixed-risk trials
in comparison to high- and low-risk trials with and without ambiguity
and for the parametrical results of potential gains or losses. The
analysis focused on the time of the decision-making. Results of the
analysis within each Region of Interest (ROI) were regarded as
statistically significant when t-values were < 0.001
uncorrected. ROIs were defined using the Talairach client software
(Talairach Project, International Consortium for Brain Mapping). In
order to prevent false positive activations, results are reported only
for brain areas which showed z-value higher than 3.09 – i.e., p
< 0.001, uncorrected – and a volume greater than 180 mm³ -
i.e., five voxels with a spatial resolution of 3 x 3 x 4 mm (here we
opted for four millimeters, because the thickness of measured slices was
three millimeter plus a distance between slices of one millimeter).