The best magazine
Methods
Participants
We tested 32 participants (18 female, mean age = 24.94 years, s.d. = 3.70), 16 in a hypnosis group and 16 in an autosuggestion group. Inclusion criteria were: right-handedness, normal or corrected-to-normal vision, no history of eating disorders or other psychiatric or neurological disorders, no medication which may influence brain activation, no current diet, liking of both sweet and salty snacks, and upper medium to high hypnotic suggestibility (prescreened using the Harvard Group Scale of Hypnotic Susceptibility: Form A; HGSHS:A; required score: 7–12; Shor and Orne, 1962). Participants of the two groups were matched in terms of age, sex and hypnotic suggestibility (Supplementary Table S1 http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1). Participants received monetary compensation for participation and gave written informed consent. The study was approved by the local ethics committee.
Procedure
Participants arrived at the laboratory hungry, as they had been asked to not eat for 4 h before the experiment (Figure 1A). The 16 participants of the hypnosis group were then hypnotized in a one-to-one setting by a professional hypnotist (H.K.). During hypnosis, participants were suggested to open their eyes and to look at a color on a monitor (blue or green, counterbalanced across subjects). They were suggested that this color would be associated with a strong feeling of disgust regarding either sweet or salty snacks (counterbalanced across subjects). This is a posthypnotic suggestion as it is activated when the posthypnotic color cue is encountered 'after' hypnosis in a normal state. Note that hypnosis did not involve induction of amnesia; participants were therefore aware of what had been suggested to them. We refer to those snacks (i.e. sweet or salty) that were associated with disgust as 'target snacks' and to the others as 'non-target snacks'. The 16 other participants were instructed—without hypnosis—to associate the color cue (blue/green) with disgust regarding either sweet or salty snacks. They were given as much time as the hypnotized subjects to make the association by themselves, and they were also sitting in front of the monitor with the color while engaging in the autosuggestion. Participants were free to use their own strategy to make this association, which we inquired about after the experiment (Supplementary Table S2 http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1).
(Enlarge Image)
Figure 1.
Methods. (A) Procedure: hungry participants were either hypnotized or they used autosuggestion in order to feel disgust regarding either sweet or salty snacks upon seeing either a blue or a green color cue. Afterwards, they carried out an auction on sweet and salty snacks shown on an either blue or green background in the fMRI scanner. They could buy a real snack at the auction. (B) Experimental auction. (C) Overview of the trial types. Note that in this figure, example stimuli have been modified so that the brands of the snacks are not recognizable.
After the intervention, we assessed participants' decision-making and brain activation using functional magnetic resonance imaging (fMRI) during an auction on unhealthy snacks (Figure 1B). Note that participants in the hypnosis group were in a normal state during scanning as hypnosis was finished beforehand. After scanning, participants completed a questionnaire about their experience. The procedures are described in more detail in the Supplementary Data http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1.
Task
The auction was a variant of a Becker–DeGroot–Marshak method (Becker et al., 1964; Plassmann et al., 2007) and took ~35 min. Participants saw pictures of sweet and salty snack items (e.g. chocolate bars, chips) and could bid between 0€ and 2.50€ for each of them (in 0.50€ steps). Snacks were shown on an either blue or green background. That is, during two out of the four runs of the auction, snacks were shown on the color cue associated with disgust (cueON-runs) and during the two other runs snacks were shown on the neutral color (cueOFF-runs). There were, hence, four trial types: targets-cueON, non-targets-cueON, targets-cueOFF, and non-targets-cueOFF (Figure 1C). Runs were separated by math problems for distraction. Our stimulus set comprised 50 high-resolution pictures of appetitive snack items (25 sweet and 25 salty) that were highly familiar in Germany. For each run, 40–43 stimuli were randomly selected for presentation out of the stimulus set. Approximately, half of the snacks in each run were target snacks and half were non-target snacks (i.e. half were sweet and half were salty), presented randomly. No stimulus was shown more than once during one run.
The auction was set up such that participants would treat each decision as a real decision and as the only one that counts (see Supplementary Data http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1). To incentivize honest responses, participants could buy a real snack on the auction. If participants did not buy a snack, they had to stay hungry for another half hour after leaving the scanner.
Postexperimental Questionnaire
The postexperimental questionnaire differed partly between groups. The following questions were overlapping (translated from German): 'How much disgust did you experience for salty snacks on blue background?' (the same question reoccurred three more times, with sweet on blue, salty on green, and salty on blue); 'To what degree did you feel physical disgust?', 'How automatic vs. self-controlled did the disgust appear to you, in case that you felt disgust?', 'Please be honest: during the experiment, did you sometimes just pretend to feel disgust?' and 'Did you often consciously think about the meaning of the color cue during the experiment and consciously/voluntarily recalled the association with disgust?' Answers were given on Likert-scales from 1–7. Some open response questions about participants' experiences were also included (e.g. regarding the strategies that participants in the autosuggestion group used).
Analysis of Behavior and Self-report
Mixed analysis of variances (ANOVAs) were used to analyze bids, mean reaction times (RTs) for bids and postexperimental disgust ratings. ANOVAs included the within-subject factors snack type (target/non-targets) and cue (cueON/cueOFF) and the between-subject factor group (hypnosis/autosuggestion). Prior to the analysis, bids were log-transformed in order to meet the assumption of normality (see[ Supplementary Data http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1 for more information). To compare the two groups regarding the remaining self-report questions, we used independent t-tests where data were normally distributed and non-parametric Mann–Whitney tests where data were not normal. For ANOVAs, η is reported as a measure of effect size. It represents the proportion of variance in the dependent variable explained by an effect. Effect sizes r for t-tests and Mann–Whitney tests were calculated following Rosenthal (1991), interpretation of r: 0.10: small, 0.30: medium, and 0.50 large (Cohen, 1988).
fMRI Data Acquisition and Preprocessing
Data were collected using a 3T Siemens Tim Trio MRI scanner with a 12-channel head coil. For each run, 185 functional images including 33 axial slices were acquired in descending order using a T2*-sensitive one-shot gradient-echo echo-planar imaging (EPI) sequence. The following parameters were used: repetition time = 2 s, echo time = 25 ms, field of view = 24 cm, matrix size = 64 × 64, voxel size = 3 × 3 × 3 mm and inter-slice gap = 0.75 mm. Functional images were realigned and unwarped based on fieldmaps, slice-time-corrected, spatially normalized to the standard Montreal National Institute (MNI) EPI template and smoothed using an isotropic 8 mm full-width half-maximum Gaussian kernel (see also Supplementary Data http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1). Coordinates reported in this article are MNI-coordinates.
fMRI Data Analysis
We calculated general linear models on the single-subject level. In a first model, we modeled target trials, non-target trials and missed trials as box-car functions of 3 s, convolved with the hemodynamic response function. Depending on the run, the target and non-target regressors encoded 'targets-cueON' and 'non-targets-cueON' or 'targets-cueOFF' and 'non-targets-cueOFF'. Motion parameters were included as regressors of no interest. High-pass temporal filtering (128 s) was applied. A second model was set up in the same way as the first model but it included the parametric modulator bid size. This served to identify regions correlating with bids. Relevant contrasts were calculated on the first level for each subject separately. On the second level, we used one-sample t-tests for determining the effects for all subjects together and independent t-tests for the group comparisons. We restricted our main analyses to voxels within our regions of interest (ROIs) in vmPFC and precuneus (Supplementary Figure S1 http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1). We corrected the results using family-wise error (FWE) correction at P < 0.05 within ROIs using small-volume correction. For completeness, we also report other relevant effects within ROIs at a liberal threshold of P < 0.001, uncorrected. Moreover, we conducted exploratory whole-brain analyses at P < 0.001, uncorrected.
Regions of Interest
For the two areas of interest—vmPFC and precuneus—we created a priori ROIs (see Supplementary Data http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1). These regions served to spatially restrict the main analyses and for small-volume alpha error adjustment. For vmPFC, we created a probabilistic ROI that takes into account the coordinates of several previous studies on valuation (Supplementary Figure S1A and Supplementary Table S5 http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1). For precuneus, we used the peak coordinate fromCojan et al. (2009) with a 15 mm sphere around it (Supplementary Figure S1B http://scan.oxfordjournals.org/content/9/9/1281/suppl/DC1).
Source: ...