Experiment 5

3.5.2.1. Method Participants

Twenty-two University of Lyon students (16 females, 6 males; mean age = 23 ± 2.71 years) took part in this experiment. Before starting the experiment, participants completed a consent form. They were individually tested and paid sixteen Euros for their participation. Exclusion criteria included olfactory and/or neurological diseases. All participants reported a normal sense of smell. The study was conducted in accordance with the Declaration of Helsinki and experimental procedures were approved by the local ethical committee.

Stimuli

Sixteen odorants (provided by Firmenich, SA) were selected on the basis of their ratings of pleasantness, familiarity, and intensity obtained from previous studies (Coppin et al., 2010;

Delplanque et al., 2008). Odorants were diluted in odorless dipropylene glycol to obtain a roughly similar average intensity (see Delplanque et al., 2008 for further details). Solutions (4 ml) were injected into the absorbent core of cylindrical felt-tip pens (14 cm long, inner diameter 1.3 cm). The use of these devices (provided by Burghart, Germany) avoids any olfactory contamination of the environment. Each odorant was coded by a random three-digit code.

2.00) 1.65) 2.15) 1.62) 2.42) 2.17)

Sniffing measurement

Sniffing data was measured with an airflow sensor (AWM720, Honeywell, France) connected to a nasal cannulae positioned in both nostrils and was continuously recorded during the experimental sessions. The physiological signal was digitally recorded at 100 Hz and stored in a computer. Sniffs were pre-processed by removing baseline offsets, and aligned in time by setting the point where the sniff entered the inspiratory phase as time zero. Sniff duration and sniff inspired volume were calculated for every odor trial. Both the volume integration and sniff duration ended at the first data point where the sniff returned to zero flow. To allow cross-subject comparison, each participant’s airflow values were divided by the maximum value within that participant, resulting in normalized values.

Procedure

Testing was performed in an experimental room designed specifically for olfactory experiments, and proceeded in three steps: (1) First Rating, (2) Choice, (3) Second Rating.

In the first step, we assessed individual’s ratings of pleasantness, familiarity and intensity for sixteen of the twenty-six odors. After each odorant presentation, participants were asked to verbally indicate a continuous value from 0 to 10 to rate its pleasantness, from

“very unpleasant” to “very pleasant”, the familiarity, from “not familiar at all” to “very familiar”, and the intensity, from “not perceived” to “very strong”. Participants were informed that they could use decimal numbers.

In the second step, eight pairs of smells were created on the basis of the initial pleasantness ratings of each individual participant. These pairs were divided into two groups of four: 1) The first group consisted of four pairs of odors that the participant had rated as similarly pleasant (i.e., difficult choices; mean rating differences = 0.21 ± 0.60 on the 11-point subjective scale); 2) The second group contained four pairs of odors previously rated differently for pleasantness (i.e., easy choices; mean rating differences = 3.16 ± 2.13 on the 11-point subjective scale). For each pair, participants were required to choose which odor they preferred.

Third, after these choices had been made, participants were again requested to assess the pleasantness, familiarity and intensity of the sixteen odors. During the entire experiment, the order in which odors or pairs of odors were presented was randomized. For each odor presentation, participants were instructed to smell the odor as naturally as possible during one inspiration.

Data analysis

Post-choice pleasantness rating modulation is typically reported when the choice is difficult. Consequently, in this manuscript, we present only the results obtained in the difficult condition. In the context of this experiment, a choice is labelled difficult when the difference between the pre-choice pleasantness ratings of the two paired odors is small (mean rating differences = 0.21 ± 0.60 on the 11-point subjective scale).

Moreover, as already described in the previous section (3.4), it is currently unclear whether the psychological processes that underlie postchoice hedonic modulation for chosen versus rejected stimuli differ. To control for this, we performed repeated measures analyses of variance (ANOVAs) for chosen and rejected smells separately.

Regarding the sniff, we were interested in its main parameters, namely its duration and its volume (Mainland & Sobel, 2006). We also reported the upward phase of the inspiratory volume, as it might reflect particularly early processes, occuring before the maximum sniff flow is reached.

3.5.2.2. Results Chosen stimuli

First, we checked whether an increased self-reported preference was observed between the two rating sessions, as classicaly reported in the free-choice paradigm (Brehm, 1956), for chosen stimuli.

To do so, we performed a repeated measures Analysis of Variance (ANOVA) on the pleasantness scores with the Rating (Rating 1, Rating 2) as a within-subject factor.

Pleasantness ratings were significantly increased in the second rating session in comparison to the first one for chosen odors, [F (1,21) = 5.91, p = .024, η² = .22] (see Figure 7).

Second, we checked whether an increase in sniff duration and volume was observed between the two rating sessions. No significant effects were observed on sniff duration and total volume [respectively, Fs (1,20) = 1.61, 1.97, ps = .219, .176]. There was a statistical trend for the upward phase of the inspiratory volume [F(1,21) = 3.28, p = .084, η² = .14], showing that for the upward phase of the inspiratory volume, during Rating 2 in comparison to Rating 1, there was a trend for decrease.

Rejected stimuli

We applied the same methodology to rejected stimuli, with the hypothesis that a decreased preference should be observed between the two ratings sessions.

Pleasantness ratings were significantly decreased between the two ratings [F(1,21)

= 6.00, p = .023, η² = .22] (see Figure 7).

Regarding sniff parameters, no significant effect was observed for sniff volume [F(1,21) = 1.42, p = .246]. Again, a statistical trend emerged for the upward phase of the inspiratory volume [F(1, 21) = 3.38, p = .080, η² = .14]. However, this decrease was not statistically different from the decrease found with chosen smells, as evidenced by a repeated measure ANOVA on the difference between Rating 1 and Rating 2, with Choice (Chosen, Rejected) as a within-subject factor [F(1,21) = 0.005, p = .945].

Sniff duration, on the other hand, was significantly decreased after choice in comparison to before choice [F(1,21) = 5.89, p = .024, η² = .22]. The repeated measure ANOVA on the difference between Rating 1 and Rating 2 with Choice (Chosen, Rejected) as a within-subject factor revealed that this decrease tended to be larger for rejected than for chosen smells [F(1,21) = 3.65, p = .007, η² = .15]⁸.

a.

8Note that if one used a one-tailed t-test instead, which one might find appropriate given our strong a-priori hypotheses regarding the direction of the difference, the decrease in sniff duration was significantly larger for rejected smells than for chosen ones [p = .035].

b.

Figure 7. a: Mean pleasantness ratings of odors that were presented in the difficult-choice condition, as a function of phase of the experiment (prechoice or postchoice) and whether the odor was chosen or rejected during the choice phase. b: Sniff duration for those same smells.

Error bars represent standard errors of the mean.

Influence of pleasantness on choices.

We checked whether the pleasantness of the odor before the choice varied as a function of the participant’s choice. The repeated measures ANOVA with the factor choice (chosen, rejected) performed on pleasantness scores before the choice was not significant [F(1,21) =1.45, p = .242] – odors evaluated as the most pleasant before choice were not chosen significantly more often during the choice phase.

Familiarity and intensity ratings.

The mean familiarity and intensity ratings for all odors are reported in Table 4.

There was a difference between the familiarity ratings of the chosen and the rejected odors prior to difficult choices [F(1,21) = 12.12, p = .002, η² = .37], unfamiliar smells tending to be chosen, but the familiarity ratings taken after choice did not differ significantly between the chosen and the rejected smells [F(1,21) = 0.06, p =.81]. Intensity ratings between chosen and rejected odors were neither significantly different just before choice [F(1,21) = 2.67, p = .117], nor after choice [F(1,21) = 0.77, p = .390].

3.5.2.3. Discussion

In this experiment, we replicated results from the free-choice paradigm, i.e. after making a choice between two similarly rated smells, participants rated the chosen one as more pleasant, and the rejected one as less pleasant (see section 3.2). Note that we replicated these results using a rating method that is different from the one we previously used. In the experiments reported in sections 3.2-3.4, participants rated odors’ pleasantness on a computer screen. Using a mouse, they moved a vertical marker across a horizontal line from “very unpleasant” (left on the scale = 1) to “very pleasant” (right on the scale = 10) and clicked to indicate their rating. In this fifth experiment, participants rated odors’ pleasantness by verbally indicating a continuous value from 0 to 10. The change of pleasantness ratings in the free-choice paradigm consequently seems robust to the rating method.

This preference modulation goes along with a statistical trend for a higher-magnitude decrease in sniff duration for previously rejected smells in comparison to chosen ones. Sniff duration goes down overall in the course of the experiment – taking chosen and rejected smells, difficult and easy pairs all together, sniff duration goes from an average of 1.68 s during Rating 1 from an average of 1.56 s during Rating 2, duration significantly different [F(1,21) = 5.48, p = .029, η² = .21]. This result is not surprising as sniff duration has been identified as decreasing with repeated exposures (see e.g. in animals, Wirth, Ferry, & Di Scala, 1998). However, this decrease tended to be larger for rejected smells than for chosen ones. As described in the introduction of this section, sniff duration is decreased for unpleasant smells in comparison to pleasant ones (Bensafi et al., 2003; Frank, Dulay, &

Gesteland, 2003; Mainland & Sobel, 2006). These results, though marginally significant, are consistent with the hypothesis that sniff patterns are a function of pleasantness ratings, and can be used to measure pleasantness ratings modulations, at least for rejected smells.

The modulation of this psychophysiological correlate is largely involuntary and hence unlikely to be due to experimenter demand effects. When viewed together with other experimental results (Coppin et al., 2010; Jarcho, Berkman & Lieberman, 2011), this is consistent with the suggestion that preference modulation following choice is of an implicit, rather than explicit nature.

We did not find any significant statistical modulation in sniff parameters for chosen odors. Preference modulation following choice depends on the pleasantness spectrum of the stimuli used – a choice between a-priori pleasant stimuli is assumed to lead to a large decrease in hedonic evaluation of the rejected stimulus (Shultz et al., 1999). Our sample of smells was rather pleasant throughout (see Table 4). If the modulation in sniff parameters follows the

same rules as the modulation in pleasantness ratings, one may speculate that the pleasantness homogeneity in our sample of odors might be one reason explaining the absence of significant modulations in sniff parameters for the chosen stimuli.

However, our results are inconclusive regarding a true modulation of preference by choice, as no control of the true impact of choice on preferences was run in this experiment (see Chen & Risen, 2010 for further details). We consequently conducted a second experiment aimed at controlling for this potential flaw of the free-choice paradigm. Rather than complex, midly pleasant to pleasant, odorants used in this experiment, we moreover used monomolecular odorants, which are rather unpleasant (see Table 5).

In Experiments 3 and 4 (described in section 3.4), we used two sequences of measurement: the classical RCR (Rating-Choice-Rating) sequence and a control RRC (Rating-Rating-Choice) sequence to control for Chen and Risen’s (2010) criticisms. In the RRC sequence, choice can reveal but not affect preferences, since they occur after the second rating has already taken place. However, note that mere exposure effects (Zajonc, 1968) appear as a potential confound while using this control, because during the second rating in RCR, participants will have seen the stimuli one time more than in the second rating in the RRC sequence (see section 3.4. for details).

We consequently used a different approach in the following experiment. More precisely, we used three sequence orders here: Choice-Rating (RCRR), Rating-Rating-Choice-Rating (RRCR) and Rating-Rating-Rating-Choice (RRRC). We believed that these three sequence of measurements allow controlling for both Chen and Risen’s (2010) criticisms and mere exposure effect. Our reasoning was that the impact of choice on preferences could be tested by comparing preference modulation between the first and the second rating in RCRR (where a choice has occured inbetween ratings) in comparison to RRCR and RRRC (where no choice has occured inbetween ratings 1 and 2). The impact of choice on preferences could also be tested by comparing the preference modulation between the first and the third rating in RCRR and in RRCR in comparison to RRRC (applying the same reasoning regarding the presence vs. absence of choice in between these two ratings).

Furthermore, comparing preference modulation between the first and the second rating in RCRR and between the first and the third rating in RRRC make it testable whether the mere exposure effect plays a role, as in both cases, stimuli have been seen an equal number (three) of times.