Modification of food preferences by posthypnotic suggestions: An event-related brain potential study

Abstract

The preference for high-over low-calorie food and difficulties in inhibiting the desire for high-calorie food are important factors involved in unhealthy food choices. Here, we explored posthypnotic suggestions (PHS), aiming to increase the desire for vegetables and fruits, as a possible new tool to induce a preference for low-calorie food. Following the termination of hypnosis, PHS was activated and deactivated in counterbalanced order, while event-related brain-potentials were recorded. Two tasks were administered, a food-face classification measuring implicit food preferences, where stimuli were categorized as showing food items or faces, and a Go-NoGo task measuring inhibition, where food items were selected as being appropriate for making a salad or not. In the food-face classification task without PHS, the early visual P1 component, a marker of stimulus reward-associations, was larger in response to high-than low-calorie food pictures; PHS eliminated this difference. PHS also yielded faster RTs and larger amplitudes of a late positive component in low-versus high-calorie items. Hence, PHS appeared to neutralize the positive perceptual bias toward high-calorie food items and enhance the effective processing of low-calorie items by increasing motivated attention. In the Go-NoGo task, PHS decreased the NoGo-N2; PHS increased the early Go- and NoGo-P3, possibly by turning low- and high-calorie items more pleasant and unpleasant, respectively, requiring more proactive control to inhibit task-irrelevant food-related emotions. Further, in the Go condition, PHS quickened the rejection of salad-inappropriate high-calorie items and increased the amplitude of late-P3, indicating facilitated classification of high-calorie items and increased response monitoring. Together, PHS effectively increased the preference for low-calorie food and the inhibition of impulses toward high-calorie food; therefore, PHS may be a promising tool for supporting healthy and sustainable food choices.

Introduction

Unhealthy food preferences and food choices contribute to the global burden of disease and environmental sustainability (Clark, Springmann, Hill, & Tilman, 2019; Forouzanfar et al., 2015; Haddad et al., 2016). For example, the number of obese and overweight individuals with a body mass index (BMI) > 25 has grown steadily in most countries (Rodgers, Woodward, Swinburn, & Dietz, 2018). Traditional measures have fallen short of stopping the obesity epidemic (OECD, 2017) and have yet to show their effectiveness to change diets towards healthier and more environmentally friendly patterns as suggested by the Lancet commission (Willett et al., 2019). Therefore, novel tools are needed to help individuals to shift their food preferences towards healthier choices and resist temptations by unhealthy options. The present study explores posthypnotic suggestions (PHS) as a possible way to modify food preferences. Of the plethora of factors determining food consumption (e.g., Stok et al., 2017), we targeted two psychological variables, food preferences and the inhibition function (Guerrieri, Nederkoorn, & Jansen, 2008; Nederkoorn, Guerrieri, Havermans, Roefs, & Jansen, 2009; Nederkoorn, Houben, Hofmann, Roefs, & Jansen, 2010.)

Preferences for fatty, sweet, or salty food, are central in determining the health outcomes of diets (e.g., Clark et al., 2019) and the global burden of disease (Haddad et al., 2016). For example, obese individuals tend to consume more high-calorie food, rich in sugar and fat in comparison to individuals with normal weight (Ebbeling et al., 2004; Schrauwen & Westerterp, 2000; Seidell, 1998). Food preferences are already influenced in utero and during breastfeeding by the mother's diet (Maier-Noth, 2019; Wilson, 2015) but are subject to alteration or entrenchment throughout life (Emond et al., 2019).

Unhealthy preferences are reinforced by the ubiquitous availability of these kinds of food in modern affluent societies. For some individuals overcoming the negative influences of unhealthy food preferences is more difficult than for others. Thus, food consumption is related to the inhibition (Dohle, Diel, & Hofmann, 2018), the executive function (EF) responsible for the suppression of prepotent but inappropriate responses (Diamond, 2013; Miyake et al., 2000). Corroborating this notion, Nederkoorn et al. (2009) observed that implicit preferences for high-calorie food are most detrimental in individuals with low inhibition abilities. In a meta-analysis Yang, Shields, Guo, and Liu (2018) reported significant deficits in inhibition and EFs in over-weight and obese individuals. Blocking EFs with transcranial magnetic stimulation of the left dorsolateral prefrontal cortex (dlPFC) increased consumption of snack food (Lowe, Hall, & Staines, 2014). In adolescents, poor EFs are a predictor of obesity (Tee, Gan, Tan, & Chin, 2018). Finally, individuals with higher BMI showed less efficient EFs (e.g., Prickett, Brennan, & Stolwyk, 2015; Smith, Hay, Campbell, & Trollor, 2011), especially in terms of inhibition (Bartholdy et al., 2017). Therefore, in addition to food preferences, inhibition of reflexive but inappropriate desires appears to be an important factor in food choice.

In the present study, we assessed the effects of PHS, tailored to increase the preference for healthy food and – by implied contrast – inhibit impulses toward high-calorie food. That means, even though we did not expect participants to reject or devalue high-calorie food items as our PHS targeted low-calorie food (fruits, vegetables), it seemed conceivable that participants inhibited their desire for high-calorie items. To understand the mechanisms of PHSs one should consider that hypnosis is a state of consciousness with three central properties, concentration on oneself, dissociation from the surroundings and increased suggestibility (Green, Barabasz, Barrett, & Montgomery, 2005). PHSs are presented during the hypnotic state but will be activated only after the termination of hypnosis by a specific cue such as an associated hand gesture. Multiple studies have shown that, rather than abandoning self-control, hypnosis directs and implements more effectively one's cognitive control repertoire (e.g., Iani, Ricci, Gherri, & Rubichi, 2006; Raz et al., 2003; Raz, Fan, & Posner, 2005; Sheehan, Donovan, & MacLeod, 1988; Zahedi, Stuermer, Abdel Rahman, & Sommer, 2019; Zahedi, Stuermer, Hatami, Rostami, & Sommer, 2017), and may work like a very efficient form of mental practice (Zahedi et al. in prep). In cognitive behavior therapy hypnosis is frequently used to change cognitive biases and preferences towards particular types of stimuli or mental contents (Hertel & Mathews, 2011; Kihlstrom, 2014; Kirsch, Montgomery, & Sapirstein, 1995; Milburn, 2010). Also, it is shown that hypnosis can affect perception; for instance, visual perception (Schmidt, Hecht, Naumann, & Miltner, 2017) and pain perception (Perri, Rossani, & Di Russo, 2019) have been modified successfully with hypnosis; interestingly, in both studies, the changes in perception were related to changes in brain activities, such as the P3 component. Although, to our knowledge, no previous study has applied PHSs to food preferences, in children they can be modified already by simple stories presented outside of any hypnotic context (Duncker, 1938) and also by food advertisements (Emond et al., 2019). It is important to consider that hypnotic-like experiences are defined as conditions where someone is concentrated on a special object and dissociated from other objects, such as in reading an engaging book or watching TV (Shor & Orne, 1962). In this sense, advertisements might be even more similar to hypnosis as they may provoke not only concentration and dissociation but also aim to induce increased suggestibility. Further, in an interesting study, Ludwig et al. (2014) used PHSs to induce disgust toward pictures containing different food categories, when they were superimposed on a background with a specific color, such as green or red. Notably, their PHS was not directed toward the food stimuli themselves but to the background color. The changes in the perception were correlated with decreased activity in the ventromedial prefrontal cortex (vmPFC), possibly showing that PHSs caused participants to devalue objects suggested to be disgusting. Therefore, we expected PHSs to be capable of changing food preferences.

In order to assess the neurocognitive mechanisms underlying the changes obtained by PHSs, we measured event-related brain potentials (ERPs) derived from the EEG elicited by pictures of food. ERPs are a valuable addition to behavioral measures because they provide insight into the cognitive processes mediating between stimuli and responses. Especially as it has been shown that different ERP component, such as P1, N1, P3, and late positivity complex (Allen, Iacono, Laravuso, & Dunn, 1995; Terhune, Cardena, & Lindgren, 2010; Zahedi et al., 2019) were modulated by PHSs in different tasks. Of special interest for the present study was the effect of different categories of depicted food on certain ERP components. The first components of interest were the early visual components P1 and N1, likely generated in extrastriata and inferotemporal cortex, respectively (e.g., Hickey, Chelazzi, & Theeuwes, 2010; Meule, Kubler, & Blechert, 2013; Toepel, Knebel, Hudry, le Coutre, & Murray, 2009). P1 amplitude has been reported to be larger to reward-associated in comparison to neutral or punishment-associated stimuli (Hickey et al., 2010; Schacht, Adler, Chen, Guo, & Sommer, 2012), and was positively correlated with craving for the presented stimuli in smokers (Donohue et al., 2016). Pictures of high-calorie/high-fat food has been reported to elicit a smaller N1 component (150–200 ms) than their counterparts in the studies of Meule et al. (2013) and Toepel et al. (2009).

Another component of interest is the late parietal positivity (LPP) which increases to affective relative to neutral stimuli, and is attributed to motivated attention directed at these items (Schupp, Flaisch, Stockburger, & Junghofer, 2006). LPPs to pictures of food have been reported to be larger than to non-food items, to increase as function of hunger (Nijs, Franken, & Muris, 2008; Nijs, Muris, Euser, & Franken, 2010; Stockburger, Schmalzle, Flaisch, Bublatzky, & Schupp, 2009; Stockburger, Weike, Hamm, & Schupp, 2008), of immediate or delayed consumption (Meule et al., 2013), and whether the depicted food was edible or rotten (Becker, Flaisch, Renner, & Schupp, 2016). Therefore, early and late ERP components appear to be suitable measures of the immediate significance (early components) of food or motivated attention directed at them (LPP).

Some food-related ERP studies tapped into self-regulation or EFs. Thinking about short- or long-term consequences of high- or low-calorie food pictures yielded a positive correlation between emotional eating and the LPP (Meule et al., 2013). Deliberately increasing or decreasing appetite for high-calorie food pictures affected long-latency but not earlier ERP components (Sarlo, Ubel, Leutgeb, & Schienle, 2013). Further, in Go-NoGo tasks, the amplitude of the N2 component to NoGo stimuli – taken as a sign of conflict (Enriquez-Geppert, Konrad, Pantev, & Huster, 2010; Liu, Xiao, & Shi, 2017) - was larger when food rather than non-food items served as NoGo stimuli (Watson & Garvey, 2013) and the N2 amplitude predicted the amount of food consumed after the experiment (Carbine et al., 2017).

For assessing food preferences and EFs, we utilized a face-food classification and a Go-NoGo task, respectively. In the former task, pictures of different food items were presented intermixed with pictures of faces, while participants should classify these two picture categories by choice-response button presses. This calorie-unrelated task of explicitly classifying stimuli into food or face categories (Fig. S.4 in supplementary materials), aimed to distract participants from developing hypotheses regarding the experimenters’ intentions and to measure implicit changes in food preference induced by the PHS. As the specific properties of food pictures, such as calorie content or tastiness of the depicted food items, are irrelevant for accomplishing this task, it can be considered to tap into implicit food preferences. As our PHS aimed to render low-calorie items more desirable and attractive, we expected faster responses to the low-calorie items during the PHS-active than in the PHS-inactive condition; no response-related PHS effects were anticipated for high-calorie food items. In the ERPs of the food-face classification task, we expected PHSs-related modulations of P1 and LPP components, indicating alterations in reward-associations and motivated attention, especially to low-calorie items, respectively.

In the Go-NoGo task, participants were instructed to press a button in response to frequent high-calorie food items (Go responses) and withhold their response to infrequent low-calorie items (NoGo condition). NoGo trials are considered to challenge the inhibition function (Jr & Pennington, 1996; Miyake et al., 2000; Roberts; Weisbrod, Kiefer, Marzinzik, & Spitzer, 2000), because during them a frequent and therefore prepotent response shall be withheld. The NoGo-P3 component has been consistently related to inhibition, but the NoGo-N2 was related to different factors, such as response inhibition, conflict monitoring, or emotion regulation (e.g., Albert, Lopez-Martin, & Carretie, 2010; Enriquez-Geppert et al., 2010; Gajewski & Falkenstein, 2013; Liu et al., 2017; Yang et al., 2014; Zhao, Lin, Xie, & Liu, 2019). Notably, emotional stimuli may disrupt the inhibition process (e.g., Rebetez, Rochat, Billieux, Gay, & Van der Linden, 2015; Schulz et al., 2007; Verbruggen & De Houwer, 2007) and affect both N2 and P3 components. Yang et al. (2014) reported, when NoGo stimuli were emotional facial expressions – either positive or negative – rather than neutral expressions, response time to Go stimuli and NoGo-N2 amplitudes decreased. In addition, P3 amplitudes to emotional facial expressions, when serving as Go or NoGo stimuli were bigger than to neutral expressions. Zhao et al. (2019) found that task-irrelevant emotionally positive background pictures reduced the NoGo-N2 amplitude relative to neutral pictures and increased the NoGo-P3, even after strictly controlling for arousal in the emotionally loaded stimuli.

Our PHS aimed to change low-calorie food items into the one and only desirable food category, hence inducing positive emotions toward these items and likely reducing the desirability of high-calorie food items, possibly inducing negative emotions. Therefore, we expected activated PHS to affect both Go and NoGo conditions. Specifically, without PHS we expected pronounced conflicts when the frequent button-pressing response was to be inhibited and, therefore, a salient NoGo-N2 and NoGo-P3 (Enriquez-Geppert et al., 2010; Gajewski & Falkenstein, 2013). However, in the PHS-active condition, as NoGo stimuli should become more desirable and pleasant we expected a smaller NoGo-N2 and an increased NoGo-P3 (Albert et al., 2010; Yang et al., 2014; Zhao et al., 2019). Simultaneously, as in PHS-active condition high-calorie items should become less desirable or even aversive, we expected an increased Go-P3.