Decoupling of desire and salivation over repeated chocolate consumption and the moderating role of food legalizing

This study aimed to investigate the responses in desire and salivation during repeated chocolate consumption, and examining how these responses are influenced by a relaxed relationship with food, or ’ food legalizing ’ . Salivation is often used as a proxy for desire, though evidence for this correlation is mixed. We hypothesized that both desire and salivation would decrease with repeated chocolate intake. Additionally, research has suggested that eating styles may affect habituation rates. We proposed that individuals with the food legalizing trait would habituate more rapidly to chocolate, providing an alternative mechanism to reward sensitivity associated with restrained eating. Fifty healthy-weight individuals participated in the study, consuming five blocks of chocolate (each 4 g, 22 calories) over trials. After the trials, participants were allowed to eat as much chocolate as they desired. The results showed that salivation was not correlated with self-reported desire over repeated chocolate consumption. While desire decreased with repeated intake and predicted ad libitum consumption, salivation increased and did not predict ad libitum consumption. Furthermore, food legalizing moderated the rate of responding in terms of the desire to eat but did not affect salivary flow. These findings suggest that salivation is not a reliable physiological measure of desire when eating chocolate. Instead, salivation appears to reflect sensitization to the sensory characteristics of chocolate and is less predictive of subsequent ad libitum chocolate consumption than self-reported desire. Lastly, having a carefree relationship with chocolate may help regulate its consumption, highlighting the potential benefits of a relaxed attitude toward food.


Introduction
Chocolate can evoke both desire and conflicting emotions; while some see it as a treat, others perceive it as a threat (Kuijer & Boyce, 2014;Small et al., 2001).Admittedly, desires are not inherently bad; in fact they are part of being human and motivate behavior for a stimulus or experience that is anticipated to be rewarding (Papies & Barsalou, 2015).However, the desire for foods like chocolate may be problematic for those trying to limit intake, especially as consumption can perpetuate itself and desire may arise in the absence of hunger (Burger, 2017;Kavanagh et al., 2005).
Because the desire to eat chocolate may arise in the absence of hunger, and chocolate may even disrupt appetite regulation, habituation in addition to satiation may be an important factor for eating cessation (Erlanson-Albertsson, 2005).Habituation is a general process in which responses to a repeatedly presented stimulus decrease over time (Rankin et al., 2009).Repeated presentations of food decrease responding for that food and signal termination of eating (Epstein et al., 2009;Epstein & Carr, 2021).Habituation may occur independently of satiation, as even repeatedly imagining consumption of chocolate or other food can be sufficient to reduce responding (Epstein et al., 1993;Haasova et al., 2016;Kiortsis et al., 2018).Yet, in the event of eating chocolate, habituation and satiation may act in concert and lead to reduced responding over time.
A slower rate of reduction in responding to a food stimulus is associated with greater ad libitum food intake.In general, higher cognitive and physiological reactivity is associated with higher ad libitum food intake (Boswell & Kober, 2016).For example, Temple et al. (2008) found that increasing food variety prolongs the habituation process as indicated by higher reactivity in the last trial, resulting in higher ad libitum food intake.Moreover, distracting participants can slow the rate of habituation leading to higher reactivity in the last trial that was associated with greater ad libitum food intake (Temple et al., 2007).Understanding the relationship between reactivity and ad libitum consumption can help guide interventions to prevent overeating.Therefore, we propose the following hypothesis: H2: Low self-reported desire to eat (H2a) and low levels of salivation (H2b) in the last trial are associated with lower subsequent ad libitum food intake.
Variations in habituation rates among individuals have been observed and can in part be attributed to their eating styles (Epstein et al., 2020).Others have suggested that restrained eating may lead to slower habituation due higher reward sensitivity, but no such association was found (e.g., Ahern et al., 2010;Epstein et al., 2020).Theory suggests that individuals with a carefree attitude towards food, known as 'food legalizing', may habituate more rapidly to chocolate specifically due to less eating-related guilt (Palascha et al., 2020b).According to habituation theory, the emotional intensity of stimuli influences the rate of habituation, with emotional stimuli expected to persist longer than purely sensory ones (Epstein et al., 2009).While emotional eating, or eating in response to emotions, is not directly associated with habituation, self-reported and physiological arousal have been found to negatively impact the rate of habituation (Epstein et al., 2020;Epstein, Mitchell, et al., 1993).Understanding the role of food legalizing is important, because it can help to inform dietary interventions to manage overeating by targeting factors that influence individual differences in habituation rates.Therefore, we propose the hypothesis: H3: Individuals high in food legalizing are expected to habituate faster to chocolate, leading to a quicker decline in self-reported desire to eat (H3a) and salivation (H3b) when compared to people low in food legalizing.

Participants
The study enrolled female participants with a body mass index (BMI) of 30 or less, xerostomia scores of 15 or lower, a history of prior chocolate consumption, no eating disorder history, non-smoking status, and a minimum chocolate liking rating of 5. Sample characteristics are displayed in Table 1.Participants were recruited through social networks and the Wageningen University e-mail list.In exchange for their time, participants received a €5 gift voucher.Assessing the eligibility of participants took place two weeks prior to the experiment (Fig. 1.).All participants provided written informed consent prior to the eligibility assessment and prior to participating in the experiment.

Materials
Milk chocolate (540 kcal/100 g) produced by Albert Heijn B.V. (Zaandam, the Netherlands) was used in the study.Half of a chocolate bar, broken down into blocks ( ± 22 kcal, ± 4 g), was presented on a small plate as the stimulus.

Salivation
To measure a baseline of salivation participants were asked to first rinse their mouth with water, swallow, look at a wall for one minute, keep their lips sealed, to not move their tongue, and to not swallow (Papies et al., 2022).Participants then spat all the collected saliva into a paper cup.The paper cup was weighed on a Precision Scale, model 75390/3001428, Schou Company A/S (Kolding, Germany) beforehand.This paper cup with saliva was then weighed again, and the difference served as a baseline measure of their salivation.In the subsequent trials this process was repeated while looking at the chocolate for a minute and prior to eating.A cup of water was provided for participants to rinse  their mouths with between trials.More water was provided when asked for by the participant.

Desire to eat
The desire to eat was assessed with a single item adapted from Papies et al. (2022); 'How much would you like to eat the chocolate in front of you right now?' to which participants could respond on a 7-point scale (1 = 'Not at all' to 7 = 'Very much').

Food legalizing
A subscale of the Multidimensional Internally Regulated Eating Scale was used to measure food legalizing (ω = .68)(Palascha et al., 2020a).It consisted of three items; 'I am relaxed about my relationship with food', 'I do not feel guilty if I occasionally overeat', and 'I can eat all foods that I like without guilt', to which participants could respond on a 7-point scale (1 = 'Completely untrue for me' to 7 = 'Completely true for me').After deleting the first item, the internal consistency of the scale increased to r sb = .72.Thus, we proceeded with items two and three, which were averaged to create a scale indicating food legalizing.A higher score indicated a relationship with less eating-related guilt.

Food intake
The chocolate was pre-weighed and post-weighed on a Precision Scale, model 75390/3001428, Schou Company A/S (Kolding, Germany) to determine how much the participants had eaten during the experiment and the difference in weight (grams) functioned as the outcome variable for food intake.

Eating-related guilt
Three questions about eating-related guilt (ω = .86)were adapted from a study by de Witt Huberts et al. (2013).To assess guilt specifically related to eating, three questions addressed how guilty participants felt about having eaten the chocolate: 'How guilty do you feel about eating the chocolate?';'How guilty do you feel about the amount of chocolate you have eaten?';'How guilty do you feel about eating?'.The responses were rated on a 7-point scale ranging from 1 = 'Not at all' to 7 = 'Very much'.The three items were averaged into a scale indicating eating-related guilt where a higher score indicated a higher level of eating-related guilt.

Satiation, hunger, and thirst
Satiation, hunger, and thirst were each measured with one singleitem measure, 'How full do you feel', 'How hungry do you feel?', and 'How thirsty do you feel?' to which participants could respond on a 7point scale (1 = 'Not at all' to 7 = 'Very much').

Body Mass Index
BMI was calculated with help of the self-reported height and weight of the respondents during the screening of the participants (BMI = kg/ m 2 ).

Xerostomia inventory score
The shortened version of the Xerostomia Inventory Score (ω = .78)has been used to exclude participants with dry mouth (Thomson, 2015).The scale was slightly adapted so that participants responded to a 5-point scale (1 = 'Never', 2 = 'Hardly ever', 3 = 'Occasionally', 4 = 'Frequently', 5 = 'Always') instead of a 3-point scale to avoid excluding too many participants.The items were summed into a total score where a higher score indicated a dryer mouth.

Procedure
Two weeks before the eating session, an online screening survey was distributed for which participants provided informed consent.Data on inclusion and exclusion criteria, as well as data for the trait food legalizing, were collected.Participants were given instructions prior to partaking in the study; do not consume (much) alcohol the night before participation, drink and eat as usual, show up pleasantly satiated, and not too hungry or thirsty.On the day of the eating session, participants were first asked to provide written informed consent.Participants were asked how hungry/satiated/thirsty they were and to provide a baseline measure of desire and salivation (Fig. 2).In anticipation of the first bite, participants were asked to first report their level of desire to begin eating before spitting the saliva into the corresponding paper cup.After the first bite and in the following trials, participants repeated the procedure of the first trial in anticipation of the next portion.Each trial lasted about two minutes and involved the consumption of one block of chocolate containing around 22 kcal.The consumption of five small blocks (110 kcal) in 10 min made the onset of satiation unlikely to have had a significant influence on responding.After the fifth trial, participants were allowed to eat as much of the chocolate as they desired.After eating, participants were asked to self-report on feelings of guilt they might have experienced while eating the chocolate.

Data analysis strategy
Participants' desire and salivation are nested within each individual and are interdependent due to repeated chocolate consumption and evaluation in anticipation of the next portion.To account for this, we used linear mixed-effects models1 to test our hypotheses (Gałecki & Burzykowski, 2013;Quené & Van Den Bergh, 2004).We conducted all analyses in R (version 4.3.1),specifically using the package lme4 (Bates et al., 2015;R Core Team, 2023).The significance level was α = 0.05 for all analyses.
To test the hypotheses, linear mixed-effects models were fitted with restricted maximum likelihood estimation.In the first step, the main predictor, trials, was added to both models to test H1a and H1b.In a second step, an interaction term, trials*trials, was introduced to test if the data was better explained by a second-order polynomial of the trials term.In the third step, we added food legalizing, a two-way interaction term, trials * food legalizing, and three-way interaction, trials * trials * food legalizing, to test H3a and H3b.To test H2a and H2b, two Pearson correlations were conducted with the desire to eat and salivation in the last trial as independent variables and total food intake as the dependent variable.Other, Pearson correlations were conducted to assess where desire and salivation were correlated in each respective trial, and additionally between age, hunger, satiation, BMI, food legalizing, and eating-related guilt.

Results
We used a linear mixed-effects model to investigate the effect of trials (fixed effect) on the desire to eat, allowing for random slopes and intercepts (Table 2).The analysis revealed that the desire to eat was initially affected by the trials, but this effect was not statistically significant.However, when introducing a second-order polynomial of the trials term, the effect became significant, providing evidence for H1a.This indicates a non-linear change in the desire to eat over trials, with an initial increase from baseline, followed by a decrease in the middle trials, and finally returning to baseline in the last trial (Fig. 3. A).In the third step, food legalizing was added, as well as a two-way interaction (trials * food legalizing) and as a three-way interaction (trials * trials * food legalizing) in the model.The effect of food legalizing on the desire to eat was not significant, nor was the two-way interaction.However, the three-way interaction was significant, providing evidence for hypothesis H3a.This suggests that higher levels of food legalizing are associated with steeper initial increases, followed by steeper decreases in the desire to eat, eventually falling below baseline (Fig. 4. A).
We used a linear mixed-effects model to test the effect of trials (fixed effect) on salivation, allowing for random intercepts (Table 3).The analysis showed that the trials had a significant effect on salivation.When introducing a second-order polynomial of the trials term, this effect remained significant but did not support H1b.This indicates a non-linear change in salivation over trials, with an increase from baseline that remained elevated without returning to baseline (Fig. 3. B).In the third step, food legalizing was added, as well as a two-way interaction (trials * food legalizing) and as a three-way interaction (trials * trials * food legalizing) in the model.Neither the effect of food legalizing nor the two-way and three-way interactions were significant, failing to provide evidence for hypothesis H3b.Thus, food legalizing did not significantly moderate salivation over repeated chocolate intake (Fig. 4

Discussion
Our study aimed to test the influence of repeated chocolate intake on the desire to eat and its presumed physiological correlate salivation.We expected a reduction in responding in both desire and salivation over repeated chocolate intake.Additionally, we wanted to test whether food legalizing moderated the rate of responding to chocolate.The study was designed in a way to reduce the potential impact of satiation as much as possible (e.g., the amount of chocolate consumed over trials and the duration of the trials), however complete mitigation may prove difficult with actual food intake.In any case, if the decrease in desire and salivation is at least in part due to habituation, food legalizing was expected to influence its rate due to eating-related guilt.Last, we expected that desire and salivation in the last trial would predict ad libitum chocolate intake.To not reduce individual differences in salivation and desire in the last trial thereby obscuring effects on subsequent ab libitum intake, we did not introduce a novel food stimulus or other distractor to induce dishabituation at the end of the trials.
In this study, salivation was not a significant physiological correlate for desire when eating chocolate, which is in line which some studies (e. g., Nederkoorn et al., 2000;Nederkoorn & Jansen, 2002;Nirenberg & Miller, 1982;Sahakian, 1981;Van Gucht et al., 2008;Wooley & Wooley, 1981), but also different from previous research (e.g., Booth & Fuller, 1981;Keesman et al., 2016;Legenbauer et al., 2004;Meule & Hormes, 2015).While various methods for salivary flow appear to be correlated (Nederkoorn, Smulders, & Jansen, 1999b;White, 1977), results may be influenced by methodological differences.For instance, Meule & Hormes (2015) used cotton dental rolls to collect saliva, whereas in the current study participants were asked to spit collected saliva into a cup similar to Keesman et al. (2016) and.Nonetheless, the current study suggests that when eating chocolate, salivation is mostly a cephalic phase response to optimize digestion and is not necessarily experienced as the desire to eat.Although some individuals may have noticed their increased salivation and attributed this occurrence to desire, the lack of a correlation is consistent with the notion that there seems to be no physiological fingerprints for emotions (e.g., Barrett, 2017;Berridge, 2018).Instead, salivation is more likely to be a specific instance of desire that is unique to each individual.
Repeated intake of chocolate resulted in non-linear change, with an increase from baseline and a decrease at later trials before returning to baseline, providing evidence for our first hypothesis (H1a).An initial increase in responding from baseline that is typical of habituation can be observed (Epstein et al., 2009;Groves & Thompson, 1970).A deliberate but obvious limitation of our study design is that we cannot completely rule out the possibility that this decrease in responses is due to habituation or satiation.With the actual repeated consumption of a food, a reduction in responding can be attributed to both processes, but to what extent satiation may have exerted an influence has been controlled for.Satiation was controlled for by consuming only small amounts of chocolate (4 g, 22 kcal per block) between trials and by the relatively short timeframe in which all trials were conducted.A mere 10 min seems insufficient to induce physiological signals of satiation, especially at such low intake.Yet, desire to eat levels still returned to baseline in the last trial, which is not surprising as repeatedly imagining chocolate

Table 2
Linear mixed-effect models for desire to eat over trials and moderated by food legalizing.
Step 1 Step 2 Step consumption is enough for habituation to set in (Kiortsis et al., 2018).Although our hypothesis did not state that the observed decrease in responding was solely due to habituation, the fact that habituation occurs independently of caloric intake and that our study design controlled for satiation suggests that habituation did play a role (Epstein et al., 2009).Salivation, however, did not display a similar significant reduction in responses over repeated chocolate intake (H1b).Instead, salivation initially increased, after which it remained elevated and did not return to baseline over trials.The increase in salivary flow from baseline is in line with previous research (e.g., Keesman et al., 2016;Morquecho--Campos et al., 2020;Nederkoorn et al., 2000;Nirenberg & Miller, 1982;Proserpio et al., 2017;Sahakian, 1981;Wooley & Wooley, 1981).The following increase in salivation can be explained by the transition from visual to gustatory exposure (Morquecho-Campos et al., 2020).However, that salivation did not decrease within the allotted trials is unexpected, as another study involving eating did show such a reduction in salivary flow (Epstein, Caggiula, et al., 1993).In other words, the sustained elevation of salivation over trials cannot be explained by the act of eating and the cephalic properties of salivation.The same study by Epstein et al. (1993) had two groups; one group ate portions of 32 calories repeatedly, whereas the other ate portions of 320 calories whilst controlling for sensory characteristics.The authors did not find a significant difference in the rate of habituation as expressed in salivation between 32 and 320 calories.Therefore, caloric intake is unable to explain the lack of reduction in salivation in our study.This may suggest that the sensory characteristics of chocolate are responsible for the sensitization of salivation over repeated intake, which is in line with research that suggests that highly-palatable high-sugar foods may serve to perpetuate consumption in a feed-forward manner (Burger, 2017;Epstein et al., 2009).
Different response systems exhibiting different rates of reduction in response to a stimulus is consistent with theory and has been demonstrated before (Groves & Thompson, 1970;Jordan et al., 2000).Yet, it was unexpected that there was a discrepancy between salivation and desire, as another study did display a synchronous reduction in salivation and motivated behavior, hence desire (Epstein et al., 2003).
The decoupling between desire and salivation can also be explained by the finding that motivational systems are functionally separated over chocolate consumption (Small et al., 2001).On one hand, individuals

Table 3
Linear mixed-effect models for salivation over trials and moderated by food legalizing.
Step may desire to refrain from consuming chocolate; however, on the other hand, the rewarding nature of chocolate prompts the body to prepare for continued consumption by increasing salivation.
If the decrease in desire and salivation after repeated chocolate consumption was partly due to habituation, we expected a faster rate of habituation for individuals high in food legalizing.The current study confirms that food legalizing moderates self-reported desire (H3a), but not salivation, over repeated chocolate intake (H3b).Fig. 4A illustrates that individuals high in food legalizing exhibit steeper habituation curves compared to those lower in food legalizing.This novel insight suggests that people with a relaxed relationship with food, free from eating-related guilt, habituate more quickly to repeated chocolate consumption.Previous studies hypothesized that habituation rates were slower for those with restrained eating styles due to higher reward sensitivity, but found no such association (e.g., Ahern et al., 2010;Epstein et al., 2020).Our findings suggest that it is not necessarily increased reward sensitivity but rather emotions that may inhibit habituation.While emotional eating (i.e., eating in response to emotions) was not associated with habituation, our results indicate that emotions specific to a food stimulus are related to habituation (Epstein et al., 2020).These findings align with habituation theory, which postulates that the emotional intensity of stimuli influences the rate of habituation, with emotional stimuli expected to outlast sensory ones (Epstein et al., 2009).Thus, similar to self-reported and physiological arousal, emotions specific to a stimulus can negatively impact the rate of habituation (Epstein, Mitchell, et al., 1993).
Lastly, per our hypothesis (H2a), the analysis demonstrated that selfreported desire to eat in the last trial can significantly predict total food intake, whereas most recent salivation levels could not (H2b).The findings partly contradict previous research where both cognitive and physiological reactivity predicted food intake (Boswell & Kober, 2016;).In line with dual-process theory of habituation, these findings may imply that cognitive response systems can trump physiological ones (Groves & Thompson, 1970).The two inferred mechanisms that underlie sensitization and habituation are assumed to interact and together yield the final common behavioral outcome, that is, food intake.Even though desire does not always lead to motivated behavior, in this case, participants seemed to have behaved according to their self-reported experience of desire (Papies & Barsalou, 2015).

Limitations and future research
Although it was not our hypothesis that a reduction in responding was purely due to habituation, further proof for this could be provided by conducting a study where ad libitum intake differences do not matter, thus introducing a novel food stimulus at the end.The decision to conduct six trials was made to reduce the burden on participants in terms of collecting saliva.However, that the onset of habituation to chocolate as bodily conveyed by salivation has not been observed within six trials does not rule out the possibility that saliva production does not decrease after more trials.It is possible that following instructions during trials may have led to dishabituation of salivary flow.An interesting observation of this study is the lack of participants who score very low on food legalizing.These participants either did not respond to the invitation e-mail or withdrew at the very last minute.This behavior appears appropriate for someone who does not have a flexible relationship with food.Lastly, because the study was conducted in a classroom, it is likely that the situation was incongruent with an average consumption experience, influencing the desire to eat and salivation in participants (Papies et al., 2022).Especially those that did score low on food legalizing might have felt observed by partaking in the study, which may have influenced their cognition, physiology, and actual behavior.

Conclusion
When it comes to eating chocolate, this study suggests that salivation is not a physiological indicator of desire.The desire to eat reduced significantly over repeated chocolate intake after an increase from baseline due to a combination of satiation and habituation, but likely largely due to the latter.Salivation increased and may reflect sensitization to the sensory characteristics of chocolate over repeated intake.Self-reported desire, but not salivation, was predictive of ad libitum chocolate intake.Food legalizing moderated the rate of responding in the desire to eat indicating that having a carefree relationship with food that is not characterized by eating-related guilt may be beneficial for habituation to chocolate.

Ethical statement
Informed consent was given by participants prior to the screening and before actual participation in the study.Moreover, the Social Sciences Ethical Committee of Wageningen University has approved this study before the collection of data.The study was preregistered at 10.17605/OSF.IO/4PS5X.

Fig. 3 .
Fig. 3.A = Average observed and predicted desire to eat (1 'Not at all' -7 'Very much') ± SEM per trial is shown.B = Average observed and predicted salivation (g) ± SEM per trial are shown.

Fig. 4 .
Fig. 4. A = Food legalizing moderated the desire to eat over trials.B = Food legalizing did not moderate salivation over trials.Down-transformation of food legalizing is for visual purposes only.

Fig. 5 .
Fig. 5. = Desire to eat in the last trial and ad libitum food intake.

Table 1
Characteristics of the 50 participants in the current study.