In Study 2, an fMRI experiment was designed to explore the active broad Brodmann area when consumers received and processed ad information. A 2 (with[117]/without[117] spokesperson) * 3 (strong[52]/weak[52]/heuristic[13] argument) within-subjects design was employed for the fMRI experiment.
Participants
Thirty-two participants were recruited from the National Chengchi University in Taipei. Two participants were excluded because of the image interference due to participants’ dental mouthpiece.
Finally, 30 participants (mean age 20.9; 15 male and 15 female) who had normal or corrected-to-normal visual acuity and no known neurological condition, were included in analyses.
After being informed about potential risks, participants provided an informed written consent before participation. The experimental standards were approved by the local ethics committee of National Taiwan University in Taipei.
Materials
The stimuli through E-prime in the fMRI study were 234 ads, including 52 products (in 13 categories) and 52 celebrities selected by the research team. The products and celebrities were first chosen based on the database of Eastern Integrated Consumer Profile, and then tested on a sample of 26 college students. The stimuli were divided into two groups. One group had 117 advertising stimuli including 52 strong arguments, 52 weak arguments, and 13 heuristic arguments with spokesperson. The other group had 117 advertising stimuli including 52 strong arguments, 52 weak arguments, and 13 heuristic arguments but without spokesperson.
19
Procedures
fMRI experiment. The participant’s head was fixed with foam pads during the fMRI scan to minimize head motions. All participants reported no difficulty in viewing stimuli or hearing instructions. Participants who required vision correction used either MRI-compatible contact lenses or MRI-compatible plastic goggles. The experiment was performed in two runs. First run contained 117 advertising stimuli with spokesperson and the other run contained 117 advertising stimuli without spokesperson. During each run, participants were first presented with the instruction and a dummy scan by self-paced, and the 117 trials were presented. Within one trial, a stimulus was presented for 7 seconds following a fixation period with randomly jittered inter-trial intervals of 1, 2, or 3 seconds. Participants were asked to judge whether the advertising concept was persuasiveness or not persuasiveness. (see Fig. 1).
Fig. 1. Study 2: fMRI procedures of experimental design. The experiment was performed in two runs, and 117 trials were presented in each run.
The response time was unlimited until the participant responded the question. The stimuli of the 3 type argument (strong[52]/weak[52]/heuristic[13] argument) of persuasiveness judgment were also randomly distributed in each run. There was a 1-min break between the two runs, and each
20
run took approximately 21 minutes. The total duration of the experiment for each participant was approximately 44 minutes. Participants then completed a post-scan 3-point rating of the persuasiveness judgment on each argument with product, in order to for this research to retrieve the brain–behavior relationship later. Finally, participants were requested to choose dislike and unknown spokesperson from the 52 spokespersons.
Data collection. Scans were performed in a 3-tesla Siemens Megnetom Skyra Siemens MRI
scanner using a 32-channel head coil. Visual stimuli were presented to the participants on a Hitachi CP-SX635 Projector. BOLD echoplanar images (EPIs) were collected using T2*-weighted gradient-echo echoplanar imaging sequences (voxel size, 4*4*3𝑚𝑚𝑚𝑚3 ). Each volume contained 34 transversal slices of 3 mm slice thickness that were oriented parallel to the anterior and posterior commissure (AC–PC) line covering the entire brain (TR=2000 ms, TE=24 ms, flip angle=90°, FOV=256 mm, 64*64 matrix, in plane-resolution=4.0*4.0 mm2). High-resolution T1-weighted structural images were also acquired using the 3D MPRAGE pulse sequence:
TR=1560 ms, TE=3.30 ms, flip angle=15.0°, 256*256 voxel matrix, FOV=256 mm, 192 contiguous axial slices, thickness=1.0 mm, and in-plane resolution: 1.0*1.0 mm2. This study included two runs, and the first two TRs in each functional run were discarded to avoid T1 equilibrium effects.
fMRI data preparation and analysis. Prior to statistical analysis, all images were
reconstructed, aligned, and corrected (in the x and y dimension) for movement artifacts (Woods, Mazziotta, and Cherry 1993). A two-dimensional Gaussian filter (approximately 3 mm at half-height) was applied to enhance signal-to-noise characteristics for each voxel. Signal changes during brain activity were identified using a “block design” that compared average signal amplitude acquired during the activity epochs with average signals acquired during baseline epochs according to a general linear model. An “active” voxel was defined as one in which the average magnetic resonance signal acquired during the stimulation periods was significantly different from the average baseline levels, p < .005, corrected for multiple comparisons based on empirically validated false-positive rates obtained using both resting brain and copper sulfate phantoms (Hirsch et al.
21
2000). This particular analysis procedure was developed to map sensory/motor, language, and visual-sensitive areas for neurosurgical planning using fMRI, and has been validated by conventional mapping techniques such as direct cortical stimulation, somatosensory evoked potentials, and surgical outcome studies.
An active area was defined for each participant as a cluster of at least 5 contiguous voxels each with a false-positive rate, p < .005. To preserve the highest spatial resolution for each participant, an idiopathic strategy was applied for the first stage of data analysis where each participant was processed separately. A modified “forward transform” method was employed to assign labels to the active individual brain areas for each subject where the brain topology was employed as an index to labels of the Human Brain Atlas (Lancaster et al. 2000). Accordingly, the stages of assignment included identification of the brain slice passing through the AC/PC line and location of respective commissures of the axial view; assignment of an atlas plate to each brain slice; location of the vertical AC/PC plane on all T2*-weighted images of brain slices; location of the central sulcus and confirmation of those landmarks on all T1-weighted images of brain slices; assignment of the anatomical labels, Brodmann’s areas, and atlas sectors for each active cluster; and determination of each active cluster volume on the basis of voxel count.
Results
Manipulation checks. The spokespersons were checked by participants’ post-test behavioral
responses to whether they have seen the spokespersons or not. Generally, participants recognized the 52 spokespersons. Moreover, three type arguments (strong/weak/heuristic argument) also were checked by participants’ post-test response. Participants’ persuasiveness judgments for the three type arguments are consistent with the design.
Strong versus weak argument contrast. The hypothesis about the contrast of strong versus weak argument suggested a stronger activation in MPFC and AFC. However, the result did not support hypothesis 2. Instead, the experimental results showed that BA32 (R, Anterior Cingulate)
22
was activated. The anterior cingulate cortex (ACC) is the frontal part of the cingulate, and it consists of BA 24, 32, and 33 (Foxall, 2015). Anterior cingulate cortex (ACC) is involved in rational cognitive functions, such as reward anticipation, decision-making, empathy, impulse control, and emotion (Foxall 2015).
Table 1. Activated Regions in the Strong versus Weak Argument Contrast
Condition Regions Side BA Voxels Z Max MNI Coordinate X Y Z S1+S0-B0 Inferior Frontal Gyrus L 46 4534 6.37 -46 26 16 Inferior Frontal Gyrus R 9 1900 5.29 40 6 26 W1+W0-B0 Inferior Frontal Gyrus L 46 3095 6.06 -46 26 16 Inferior Frontal Gyrus R 9 1125 4.75 42 8 26 S1+S0-(W1+W0) Anterior Cingulate R 32 81245 3.68 22 44 14 Note 1: S = Strong argument, W = Weak argument, 1 = with spokesperson, and 0 = without
spokesperson.
Note 2: BA = Brodmann’s area; Voxels = number of voxels in cluster, only clusters with an extent threshold of p < .05, corrected for whole brain and cluster 10 or greater are presented;
threshold of p < .05, FWE corrected.
S1+S0-B0: BA46 (L, IFG) S1+S0-B0: BA9 (R, IFG)
23
W1+W0-B0: BA46 (L, IFG) W1+W0-B0: BA9 (R, IFG)
S1+S0-(W1+W0): BA32 (R, Anterior Cingulate)
With versus without spokesperson contrast between strong and weak argument. The
present study assumed that the different brain activations between strong and weak argument was stronger in middle frontal gyrus when the advertisement had a spokesperson than no spokesperson’s endorsement. However, the experimental results showed that no activation difference in brain.
Table 2. Activated Regions in the Spokesperson Contrast between Strong and Weak Argument
Condition Regions Side BA Voxels Z
Max
MNI Coordinate X Y Z S1-W1 Middle Frontal Gyrus L 9 44419 4.49 -34 32 34 S0-W0 Middle Frontal Gyrus R 10 40012 3.44 38 46 18 S1+S0-(W1+W0) No Activation Area
Note: BA = Brodmann’s area; Voxels = number of voxels in cluster, only clusters with an extent threshold of p < .05, corrected for whole brain and cluster 10 or greater are presented;
threshold of p < .05, FWE corrected.
24
S1-W1 BA9(L, Middle Frontal Gyrus) S0-W0 BA10(L, Middle Frontal Gyrus)
Heuristic versus strong argument contrast. The comparison of brain area activation
between heuristic versus strong argument could reveal the perceptual difference between these two types of argument. The contrast indicated no activation difference in brain while participants responded to heuristic than strong argument (i.e., H1 vs. S1, H0 vs. S0, and H1+H0 vs. S1+S0).
Thus, hypothesis 4 was not supported.
Discussion
The first fMRI experiment tested participants' brain activations after receiving strong or weak arguments. The most significant result was that participants exposed to the strong arguments had relatively higher activation in Anterior Cingulate (ACC). ACC is the frontal part of the cingulate, which is involved in rational cognitive functions, such as reward anticipation, decision-making, empathy, impulse control, and emotion (Foxall 2015). The result revealed that strong argument did bring in more rational cognitive processes than weak argument. In addition, the stronger activation between strong and weak arguments was not influenced by the present of spokesperson or not. However, the brain activity difference was not significantly different between strong and heuristic argument.
The next fMRI experiment will incorporate the involvement level into the study design as most ELM and HSM studies emphasized.
25