Compared to self-immersion, mindful attention reduces salivation and automatic food bias

Immersing ourselves in food images can sometimes make it feel subjectively real, as if the actual food were right in front of us. Excessive self-immersion into mental content, however, is a hallmark of psychological distress, and of several psychiatric conditions. Being aware that imagined events are not necessarily an accurate depiction of reality is a key feature of psychotherapeutic approaches akin to mindfulness-based interventions. Yet, it is still largely unknown to what extent one’s engagement with mental content, considering it as real, biases one’s automatic tendencies toward the world. In this study, we measured the change in subjective realism induced by a self-immersion and a mindful attention instruction, using self-reports and saliva volumes. Then, we measured behaviorally the impact of subjective realism changes on automatic approach bias toward attractive food (FAB) using an approach–avoidance task. We found a reduction in saliva volume, followed by a reduction in FAB in the mindful condition compared to the immersed condition. During the immersed condition only, saliva volumes, state and trait measures of subjective realism, and food craving traits were positively correlated with FAB values, whereas meditation experience was negatively correlated to it. We conclude that mindful attention instructions can de-automatize food bias.


1.2) In-house Questionnaire Validations:
We applied a confirmatory factor analysis to our in-house Dereification-as-a-Trait and Post-State questionnaires by evaluating the robustness and orthogonality of these constructs. We used a Maximum Likelihood Factor extraction method 3 , and then rotated these factors by using an Oblique Oblimin Rotation method 4 . We extracted as many factors as the number of constructs that we proposed in each questionnaire. As presented below, we confirmed that the number of factors was consistent with the number of constructs for each questionnaire, and that each scale and sub-scale/construct was internally consistent.

1.3) Results of the Questionnaire Validations:
Trait Questionnaire: The Trait Dereification Questionnaire was subject to a factor analysis and internal consistency analysis. Factor analysis, using parallel analysis, optimal coordinates, and acceleration factor as extraction methods, suggested the extraction of two factors. However, this solution was not significant (χ 2 =101.2, df=89, p=0.18). Since the original hypothesis considered one factor, factor analysis was re-run extracting only one factor. Items with factor loadings under 0.35 were dropped, leading to a questionnaire of ten questions. These ten items, extracted as one factor, were significant (χ 2 =74.

Pre-State Questionnaire:
Instructions: In the next section, please answer from 1 to 5 the option which best reflects your experience at the moment. With 1 being not at all and 5 very much 1) Are you happy right now?

Post-State Questionnaires:
Instruction: From item 1 to 12, the instructions are as follows: Using the 1-5 scale shown below, please indicate the extent to which you agree or disagree with the following statement regarding your experience during the task: 1-strongly disagree, 2-disagree 3-neither agree nor disagree 4-agree, 5-strongly agree.
From item 12 to 20, the instructions are the following: Using the 1-5 scale below, please indicate the statement which most accurately represents your experience at the moment. 15) How vivid were the food items for you? (How "real" did they feel?)

16) How immersed in the food images were you?
¿Qué tan inmerso te sentiste en las imágenes de comida?
17) How immersed were you in the feelings evoked by the food images?
6) While I was looking at the food images, I felt desire to eat them.

Stickiness: 4 5 10
10) While I was looking at the food pictures I was usually "dwelling in" my thoughts of these images.
Mientras veía las imágenes de comida, los pensamientos respecto a estas se mantenían "dando vueltas" en mi mente, 4) In the transition from one image to another, I usually remained thinking about the previous one.
Cuando pasaba de una imagen a otra, generalmente me quedaba pensando en la imagen anterior 5) In the transition from one image to another, the sensations triggered by the previous one remained with me Cuando pasaba de una imagen a otra, generalmente mis sensaciones sobre la primera imagen permanecían.

Meta-awareness: 3 8
3) During the food image presentation I could relate to my thoughts as mere mental events which came and went.
Durante la presentación de las imágenes pude relacionarme con mis pensamientos sobre estas, como meros eventos mentales que iban y venían. 2) I tend to stay "spinning" in my thoughts. 10) When I think about a situation, I usually realize that this is "my vision" of the situation and that there may be more interpretations of that situation in other people's points of view.

12) My thoughts sometimes reassure me.
Mis pensamientos s veces me tranquilizan.

Mis pensamientos muchas veces me ponen eufórico.
14) My thoughts make me angry.  Instruction manipulation over RTs, saliva or self-reports, was not due to differences in any of those mood or hunger variables prior to condition block performance. Bars denote standard error.

2.1.2) Reaction Times by Group:
We analyzed the effects of Instruction (mindful attention vs. immersed), food type (attractive vs. neutral) and response type (approach vs. avoid) over RTs during the AAT Consistent with our first hypothesis, we found an Instruction condition x Food type x Response type, F (1, 24) =4.5, p<0.05, η 2 G = .0008). To further assess the three-way interaction, we examined the effects of Food type and Response type in the immersed and mindful attention conditions separately. In the immersed condition, there was a Response type effect (F (1, 23)= 8.9, p<0.05, η 2 G = .002) where approach RTs were faster than avoidance RTs (Table  1,). And we found a positive food attractiveness bias (FAB) in concordance with an interaction of food type and response type (F (1, 24)= 5.9, p<0.05, η 2 G = .001). This interaction was driven by faster approach responses than avoidance responses toward attractive food images (t-test, t(24)=3.3 p<0.01) that did not occur agreement with the lack of Food type X Response type interaction (F (1, 24)= .006, p=.6 , η 2 G = .00008). (Right) Meditator group: in the 2 Instruction x 2 Food type x 2 Response type repeated-measures ANOVA , we found an Instruction x Food type x Response type, F (1, 24) =8.9, p<0.05, η 2 G = .0006). As before to further assess the last three-way interaction, we examined separately the effects of Food type and Response type in Immersed and Mindful attention conditions. In the Immersed condition, the 2 way interaction of Food type and Response type (F (1, 24)= 4, p=0.057, η 2 G = .0009) was present only as a trend.

2.1.3) Order Effect of Instruction on Behavior:
To assess whether there was an order effect of the instructions on the behavioral performance, we performed a repeated measure ANOVA over RT with Group, Start- . In all graphs, bars denote standard error. Significances *p < 0.05, **p < 0.01, ***p < 0.001.

2.1.5) Alpha-Amylase results:
As Alpha-amylase is secreted in the presence of food as part of the CPRs, we first hypothesized that an increase in Alpha-amylase activity would have taken place in response to the food cue exposure. Secondly, we also hypothesized that this effect would be more pronounced in the Immersed condition compared to the Mindful attention condition. while Alpha-amylase activity was sensitive to the Passive vs. Active manipulation of our task, it was not sensitive to our Instruction manipulation (Mindful attention vs. Immersed).
The current finding clearly ruled out Hypothesis 2: the degree of subjective realism does not impact alpha-amylase activity. However, the findings remain inconclusive regarding Hypothesis 1. More specifically, in line with Hypothesis 1, the increase from T0 to T1, could be viewed as reflecting food intake preparation. This interpretation is coherent with the effects of real food intake on Alpha-Amylase 6 . Alpha-Amylase was shown to increase in response to a standardized lunch in stimulated saliva of n = 18 healthy human subjects 6 and to be positively associated with self-ratings of satiety and fullness, and inversely associated with hunger and desire to eat 7 . Importantly, Alpha-amylase was also found to increase within 15 min after eating a standardized meal or ''sham-eating'' (chewing and expectorating) (n = 12 healthy men and women), even if the increase was significantly lower in the sham-eating condition 8 .
Against this interpretation the enhanced alpha-amylase activity from T0 to T1 could merely reflect enhanced sympathetic system activation (see for instance 9 -10 ). This alternative hypothesis is plausible at the light of the enhanced alpha-amylase activity from T2 to T1, which could be also explained along this line. Such increase would be predicted if one assumes that passively viewing food images during the passive exposure phase engages less the sympathetic system than actively doing a behavioral akin to the AAT. Future studies, will be required to explore these two alternative interpretations, for instance by comparing changes in alpha-amylase activity while immersing in non-food images compared to food images.

2.3.1) Trait Integrative Analysis:
As food craving as a trait correlated with AAT performances, we tested whether food craving scores were related to other trait characteristics of Instruction types (Dereification, Cognitive Fusion and BMI). As expected, food craving as a trait was positively correlated with BMI (Pearson r = 0. 2; p<0.05) (Fig. Supp 7a) and was also positively correlated with Cognitive Fusion Questionnaire scores (Pearson r = 0.5, p<0.001). Participants who reported greater Cognitive Fusion as a trait, also had higher scores of food cravings as a trait (Fig. Supp 7b). Moreover, food craving was negatively correlated with Dereification Questionnaire scores (Pearson r = -0.6; p<0.001) (Fig. Supp.   7c). To summarize, the more a participant reported food craving as a trait, the higher their BMI was, accompanied by a low tendency to dereify perception and a high propensity for cognitive fusion.