The effect of repeated episodic future thinking on the relative reinforcing value of snack food

Abstract

Obesity is related to delay discounting and relative reinforcing value of food. Episodic future thinking reduces delay discounting after one engagement. The effects of repeated engagement in episodic future thinking are unknown. We explored the effects of daily episodic future thinking on delay discounting, energy intake, and relative reinforcing value of food. Participants completed a delay discounting task, ad libitum buffet, and relative reinforcing value task following one engagement in episodic future thinking/episodic recent thinking and again after 1 week. One week of daily episodic future thinking did not reduce delay discounting compared to one engagement or episodic recent thinking. Engaging in episodic future thinking daily does not impact delay discounting, relative reinforcing value of snack food, or ad lib energy intake compared to one engagement in episodic future thinking.

Keywords

delay discounting energy intake episodic future thinking obesity reinforcing value of food

Introduction

Overweight and obesity prevalence in the United States has rapidly increased over the past 50 years, with recent data estimating that 39.8% of US adults had obesity in 2015–2016 (Hales et al., 2017). Obesity is a complex disease occurring due to chronic positive energy balance. Chronic positive energy balance results from overeating and poor food choices made over time (Carr et al., 2014), specifically a preference for highly palatable foods that provide immediate gratification rather than healthier, low-energy-dense foods that contribute to positive long-term health outcomes (Carr et al., 2014). Successful strategies for prevention and treatment of obesity in adults should focus on methods to reduce habitual overeating.

Overeating of high-energy-dense foods is related, in part, to a high relative reinforcing value (RRV) of these foods. The reinforcing value is an objective measure of motivation to obtain food and predicts ad libitum energy consumption (Epstein et al., 2004). People with obesity have a higher RRV of food compared to people with healthy weight (Carr et al., 2014). Methods that decrease the RRV of food may improve adherence to a nutrient-dense diet that has long-term health benefits.

People able to self-regulate behavior are more successful at adhering to a healthy diet, as they are able to delay gratification of highly reinforcing foods for greater long-term rewards such as improved health (Borghans and Golsteyn, 2006; Smith et al., 2005). Delay discounting (DD) is a construct of impulsivity defined as a preference for a smaller immediate reward over larger delayed rewards (Appelhans et al., 2012). Reducing DD may improve behavior change adherence and contribute to decision-making focused on long-term health goals instead of immediate gratification. Episodic future thinking (EFT), which involves vivid imagination of future events to lengthen time perspective, has been shown to reduce discounting (Mellis et al., 2019; Rung and Madden, 2018; Stein et al., 2016) and energy intake (Daniel et al., 2013b, 2015; Dassen et al., 2016; O’Neill et al., 2016) in adults.

Although it is well established that a single engagement in EFT reduces DD, few studies have investigated the effect of repeated engagement in EFT on DD or on eating-related behaviors. The current study tested the hypothesis that repeated, daily engagement in EFT for 1 week in people with overweight and obesity would reduce DD compared to an episodic recent thinking (ERT) control. We also investigated the effect of a single engagement and repeated EFT on the RRV of high-energy-dense snack food compared to low-energy-dense snack food and energy intake at an ad libitum snack food buffet compared to a control ERT condition.

Participants and methods

Participants

We recruited participants through a laboratory database, web-based advertisements, and flyers posted around the University at Buffalo campus. Men and women (n = 33) were included if they were between the ages of 18 and 50 years, had a score of <3.2 on the non-planning subscale of the Barratt Impulsiveness Scale (Patton et al., 1995), were weight stable (±5 lbs) for 3 months, had overweight or obesity (body mass index (BMI) ⩾ 25 kg/m²), reported moderate liking of at least half of the study foods, and owned a smartphone. We excluded people if they smoked cigarettes, were pregnant/breastfeeding, were taking medication that affected their appetite, had allergies to study foods, or were actively trying to lose weight. Using data from a study by Daniel et al. (2013b) that showed that a single session of EFT reduced DD compared with an ERT control, we found that a power of 0.8 and p < .05 could be achieved with a total sample size of 30 participants.

All participants were screened for impulsivity using the non-planning, self-control subscale of the Barratt Impulsiveness Scale, which is a validated self-report questionnaire designed to assess impulsiveness (Patton et al., 1995). The scale includes 30 items assessing three subtraits of impulsivity: attentional (attention and cognitive instability), motor (motor and perseverance), and non-planning (self-control and cognitive complexity; Patton et al., 1995). Non-planning impulsiveness is conceptualized as a lack of future thinking (Stanford et al., 2009). Previous research showed people with lower scores on the perseverance subscale of the Barratt Impulsiveness Scale were more responsive to EFT (Daniel et al., 2013a). We screened participants using the non-planning subscale to target a population that does not already plan for the future, to have room for improvement in terms of future orientation.

Procedures

We screened participants using a Survey Monkey questionnaire. The questionnaire inquired about age, preference for study foods, smoking status, weight status, and impulsivity. Research assistants reviewed screening surveys to determine if the individual met the screening criteria. If they did, we contacted participants and invited them to participate. A total of 326 people completed the screening survey, and 255 were excluded based on the exclusionary criteria, 33 declined participation, 5 dropped out, and 33 completed all procedures (Figure 1).

Figure 1.

Consort diagram showing the flow from participant recruitment to randomization and study completion.

Eligible participants were scheduled to visit the laboratory for four visits lasting 60–90 minutes each between the hours of 11:00 a.m.–5:30 p.m. Prior to scheduling, participants were randomized to either the EFT or ERT condition. Participants were not blind to their condition, but they were unaware of the alternative condition. We asked participants to refrain from eating and drinking anything other than water for 2 hours prior to each of their visits. The Institutional Review board at the University at Buffalo approved all procedures.

In the first session, all participants read and signed consent forms, completed a 24-hour and same-day dietary recall using the 5-pass method (Conway et al., 2004), developed and recorded their EFT/ERT vignettes. They then listened to one of the recordings they made prior to completing the DD task and again prior to participating in the ad libitum buffet. During the buffet, participants completed taste quality sheets for each of the study foods, a food preference form, and several questionnaires.

On the second and third visits to the laboratory, participants listened to one of their recordings, and then played a computer game (explained below) where they had the opportunity to earn both a high- and low-energy-dense snack food. Between visits two and three, participants listened to one of the recordings they made every day prior to three eating occasions (meals and snacks), so that they listened to all three each day. They listened to the recordings using the Mobile Audio Manager and Response Tracker (Sze et al., 2015). This website provided time-stamped documentation each time the participant listened to a recording, which allowed us to verify compliance to study procedures. After each listen, participants were asked to rate the vividness and positivity of the event. Vividness was described as the ability to visualize the event, and positivity as the positive emotion experienced when listening to the event. Positivity and vividness ratings were scored on a 5-point Likert-type scale ranging from 1 (not at all) to 5 (extremely) following each recording. We also asked them how the event made them feel (happy, sad, excited, okay, none of these) and to state the most vivid detail from the recording. Participants also completed a brief, five-trial adjusting delay task (Koffarnus and Bickel, 2014) after listening to each recording, which provided a measure of a discount rate for each participant, three times a day. During this 1-week period, participants also kept a dietary habit book, where they recorded everything they had to eat and drink throughout the week. A research assistant reviewed the habit book with the participant on their third visit to ensure all details were provided.

During the fourth visit, the participants again listened to their recordings prior to completing the DD task and completing the ad libitum eating task. At the end of this visit, participants also completed the Three-Factor Eating Questionnaire (Stunkard and Messick, 1985), the Binge Eating Scale Questionnaire (Gormally et al., 1982), and the Questionnaire of Weight and Eating Patterns (Spitzer et al., 1994) to assess dietary restraint, disinhibition, and to rule out potential eating disorders. No participant scored within the clinical range on these measures. At the end of the fourth visit, we compensated participants up to US$80 in a gift card of their choice for completing all study procedures.

EFT/ERT development

The EFT task required participants to generate positive future events that were going to occur in the next 1, 2, and 6 months. Each participant described two to three events at each time period and rated the positivity and significance of each event. Next, they described the most positive and significant events in vivid detail including who they would be with, what they would be doing, where they would be, and any sensory descriptors that would help them imagine themselves at the event. The ERT group followed an identical procedure; however, they described events that occurred within the past 1, 2, and 6 days. Participants recorded their EFT/ERT statements and then played them back using their smartphones. They listened to their recorded EFT/ERT vignettes before an eating occasion for 1 week between their second and third visits. Participants only had to listen to one recording prior to each eating occasion, three times a day using a smartphone-based web interface that recorded the times that participants listened to their recordings, asked four brief questions about the recordings, and had participants complete a five-trial DD task (Koffarnus and Bickel, 2014).

Primary measures

Assessment of impulsivity

Impulsivity was assessed during the first and fourth visits using a version of the DD task (Rollins et al., 2010) and the Barratt Impulsiveness Scale (Patton et al., 1995). The DD task required participants to make choices between an amount of money available immediately or US$50 available later. The magnitude of the immediate value was adjusted until it was subjectively equivalent to the later larger amount (starting at US$0.50 and progressively increasing to US$50). Subjective equivalence was obtained at seven delays (1 day, 2 days, 1 week, 2 weeks, 1 month, 2 months, and 6 months). Time delays were chosen to correspond with the delays for each event. Equivalence points were plotted for each delay and area under the curve (AUC) was computed. Greater AUC indicates less discounting, and a smaller AUC indicates greater discounting.

DD was also assessed daily using a brief, adjusting-amount version of the DD task described above (Koffarnus and Bickel, 2014). This task was delivered electronically immediately after the participant listened to each recording, which provided multiple daily measures of the discounting rate. The dependent measure analyzed from this task was the natural log of k (lnk) or the rate at which the participant discounted the future reward. A greater lnk value indicates greater discounting, while a smaller lnk indicates less discounting of the delayed reward. Participants also completed the Barratt Impulsiveness Scale on their first and fourth visits.

Ad libitum eating task

To assess food selection, ad libitum intake, and hedonics, we provided participants a buffet of 12 snack foods and asked them to taste and rate each of the foods. The buffet contained 100-g portions of six high-energy-dense (kcal/g > 4: candy-coated chocolate, chocolate chip cookies, fruit-flavored candies, ranch-flavored corn chips, plain potato chips, and cheese-coated crunchy snacks) and 300 g of six low-energy-dense (kcal/g ⩽ 1.0: grapes, canned peaches, mandarin oranges, canned pears, vanilla yogurt, and strawberry yogurt) foods. We instructed them to taste and rate each one for liking and taste qualities. After the participants completed taste quality and preference forms, they were given four questionnaires to complete (Self-Efficacy for Healthy Eating, Barratt Impulsiveness Scale, Food Security questionnaire, and the Positive and Negative Affect Schedule) and told that while they are completing them, they should feel free to help themselves to the buffet, as we would be discarding the food once they leave. The Self-Efficacy for Healthy Eating scale is a 9-item scale representing one’s belief that they can achieve a healthy diet (Bandura, 1986). The Barratt Impulsiveness Scale is a self-report instrument for measuring impulsivity (Patton et al., 1995). The Food Security questionnaire is an 8-item questionnaire from the United States Department of Agriculture (USDA) that is used to identify food-insecure households (Connell et al., 2004). The Positive and Negative Affect Schedule asks the participant to rate 20 feelings/emotions on a scale from 1 (very slightly or not at all) to 5 (extremely). These surveys were given as a distraction while participants were exposed to the snack food buffet. They were given the questionnaires and told that, while they were completing them, they could eat as much or as little of the snack food as they like. The purpose was to create a tempting food situation that may lead to overeating. Data from the Food Security questionnaire and Positive and Negative Affect Schedule were not analyzed. We analyzed total energy consumed in terms of kilocalories, as well as energy consumed from low- and high-energy-dense foods separately.

Assessment of RRV of food

Participants were instructed on how to use a computer-generated task to earn points toward a highly liked (rated ⩾60 mm on visual analog scale (VAS) scales) snack food. The task looks similar to a slot machine, with a box containing three different shapes that are in different colors. When the left button on the mouse is pressed, the shapes rotate and change color. When all the shapes match, the participant earns 1 point. After earning 5 points, the participant receives a portion of snack food (brought into the room by the experimenter). The reinforcement task was set at a progressive ratio schedule with the following fixed ratio response requirements: 20, 40, 80, 160, 320, 640, 1280, 2560, and 5120. The number of responses required to earn a portion of food doubled after one portion was earned. Identical games were presented on two different computers with one computer set to earn points for a low-energy-dense food and the other to earn points toward a high-energy-dense food.

The session ended when the participant no longer wished to earn points for access to food. Water was provided ad libitum. The dependent measure that was analyzed from this task was the breakpoint for high- and low-energy-dense food. Breakpoint refers to the highest schedule of reinforcement completed for a particular reward.

Weight, height, and BMI

Participants’ weight was measured on the first and fourth visits using a digital scale (SECA). Height was measured on the first visit using a SECA stadiometer. Height and weight were used to calculate participants’ BMI using the Centers for Disease Control and Prevention (CDC) Adult BMI Calculator (CDC, 2019) according to the following formula: BMI = kg/m².

Hunger and thirst ratings

Participants rated their hunger and thirst at the beginning of each visit and after participating in the ad libitum buffet and RRV game. Participants rated their degree of thirst, hunger, liking of the study foods, and desire to eat the study foods on a 100-mm VAS ranging from “not at all” to “extremely.” We instructed them to think of the line as a spectrum and to mark where they felt they fell on the spectrum at that time.

Dietary assessment

To determine usual energy intake during the week that participants routinely engaged in EFT or ERT, participants documented meals, snacks, and beverages. Participants also provided previous- and same-day dietary recalls at each laboratory visit. The dietary recalls were conducted using a multi-pass interview that is regarded as effective for eliciting the most accurate information from dietary recall (Johnson et al., 1996). Dietary intake was assessed using Nutritionist Pro software.

Analytic plan

Correlation analysis

Participant characteristics including BMI, age, and sex were summarized using descriptive statistics. Pearson product-moment correlation coefficients were generated to identify significant associations at baseline. Variables that were significantly correlated with the primary dependent measures (DD, RRV of low- and high-energy-dense food, and energy intake) were entered as covariates in subsequent analyses of these measures.

Differences between conditions (EFT/ERT) were analyzed using analysis of covariance (ANCOVA) models for DD (in terms of AUC) and energy intake during the ad lib buffet (in terms of total energy intake (kilocalories), energy from low- and high-energy-dense food), controlling for age. RRV was analyzed in terms of breakpoint for low- and high-energy-dense foods, described above. Differences in RRV between conditions were analyzed using analysis of variance (ANOVA).

To test differences in energy intake, DD, and RRV of low- and high-energy-dense snack food after 1 week of repeated EFT or ERT, repeated measures ANOVA was used to test the interaction between time and condition, with condition and sex as the between-subject factors and time as the within-subject factor. Pairwise comparisons with Bonferroni corrections were performed to explore significant interactions for each primary measure.

Post hoc time series analysis

Five-trial adjusting DD data, measured after daily listens, were analyzed post hoc using a linear mixed effect model, with the rate of discounting as the dependent measure. This analysis tested the hypothesis that DD would decrease the more the participant listened to their recordings in the EFT group. Fixed effects included the number of total engagements in EFT/ERT, baseline non-planning scores on the Barratt Impul-siveness Scale, condition, age, positivity rating for each listening period, and vividness for each listening period. Random identity effects were also included in the models. We explored candidate models in this analysis using Akaike’s Information Criteria by adding and removing main effects and the interaction with condition and time until the best-fitting model was found. In the final model, as fixed effects, we entered the number of total engagements in EFT/ERT, baseline non-planning Barratt Impulsiveness Scale scores, condition, the interaction of condition and number of listens, and the interaction of condition and baseline non-planning Barratt Impulsiveness Scale scores. As random effects, we included by-subject random slopes.

Data accessibility statement

All de-identified individual participant data accompanied by a data dictionary are available to anyone who wishes to access these data. This includes all relevant, de-identified participant data collected during the four laboratory visits and upon screening. The data will be made available immediately following publication with no end date. The data are available for analyses of any purpose. The data are exported from SPSS 26, and the data dictionary is a separate excel file that describes each variable along with a description of the code. These data can be found at https://doi.org/10.6084/m9.figshare.11897991.v1.

Results

Participant characteristics: There were no differences in participant characteristics (sex, BMI, age, impulsivity) between the two groups (all p > .05; Table 1).

Table 1.

Descriptive data of participant characteristics including sex, BMI, age, and scores on the BIS 11 non-planning, self-control subscale upon screening.

	Total	ERT	EFT	p value
N	33	15	18
Sex (M/F)	14/19	7/8	7/11	.461
BMI (kg/m²)	30.3 (24.1–47.7)^**	31.7	29.1	.221
Age	30 (19–48)	32.07	28.39	.343
BIS 11	2.9	2.80	2.87	.892

ERT: episodic recent thinking; EFT: episodic future thinking; BMI: body mass index; BIS: Barratt Impulsiveness Scale.

M(range).

DD

There was a positive association between age and DD at visit 4 (R(31) = .424, p = .014).

There was no effect of condition on DD on visit 1 when controlling for age (F(1,30) = 0.760, p = .390; Figure 2(a)) or following 1 week of daily EFT or ERT (F(1,30) = 0.314, p = .579).

Figure 2.

Mean (+SEM) (a) area under the curve (AUC) values for discounting of delayed rewards, (b) energy intake (kilocalories), and (c) breakpoint for responses for HED food as a function of time and condition. There were no group differences in AUC for delay discounting (a). For energy intake, there was a significant increase in energy intake (kilocalories) at the buffet from time 1 to time 2 in the EFT condition only (b). There were no main effects of time or group nor interactions between the two on RRV of HED food (c).

Energy intake

There was no main effect of time or condition (F(1, 27) = 0.046, p = .832) on energy (kilocalories) consumed at the ad lib snack food buffet (Figure 2(b); p > .05). There was a significant increase in laboratory energy intake between time 1 and time 2 in the EFT condition only (F(1, 27) = 6.71, p = .015; Figure 2(b)). There was no significant main effect of time (F(1,28) = 0.057, p = .814) or of condition (F(1,28) = 1.37, p = .252), or for their interaction (F(1,28) = 1.2, p = .283) for energy from high-energy-dense foods consumed at the snack food buffet. In addition, there was no significant main effect of time (F(1,28) = 0.053, p = .819) or of condition (F(1,28) = 1.29, p = .265), or for their interaction (F(1,28) = 1.25, p = .273) for low-energy-dense snack foods consumed at the buffet.

Snack food reinforcement

There was no effect of condition for RRV of high- or low-energy-dense food at baseline (F(1, 29) = 0.904, p = .350; F(1,29) = 1.83, p = .187, respectively). There was also no effect of condition on RRV of high-energy-dense food following 1 week of EFT or ERT (F(1,29) = 0.306, p = .584) (Figure 2(c)). There was a significant interaction for time and condition for RRV of low-energy-dense food (F(1, 29) = 5.89, p = .022). Pairwise comparisons with Bonferroni corrections showed RRV of low-energy-dense food significantly decreased following 1 week of daily engagement in ERT (F(1,29) = 4.62, p = .04).

Results from linear mixed effect model

The final model included an identity covariance structure for the random effects and an autoregressive moving average model (1,1) for the repeated effects covariance structure. Despite improving the fit of the model, none of these factors significantly predicted DD. However, there was a trend for EFT participants to increase their DD as they listened to their recordings more and more (β = .03, p = .059).

Discussion

The purpose of this study was to test the hypothesis that engaging in EFT daily for 1 week would reduce DD, energy intake at a buffet of snack foods, and RRV of high-energy-dense snack foods relative to an ERT control. Contrary to our original hypotheses, we found that repeated EFT did not reduce DD, RRV of high-energy-dense snack food, or energy intake during a snack food buffet compared to the control, or compared to a single EFT session. In addition, participants consumed more energy at a snack food buffet following 1 week of EFT compared to the ERT group that showed no change in energy intake. Finally, there was no evidence that repeated, daily EFT reduced DD when measured immediately after listening to the vignettes and, in fact, there was a trend for DD to increase over time in the EFT group. When taken together, these findings suggest that repeated exposure to EFT does not reduce DD, energy intake, or RRV of food.

In this study, participants in the EFT condition did not discount less than participants in the ERT condition on the initial visit. Previous studies have shown a significant reduction in DD following EFT (Daniel et al., 2013a, 2013b, 2015; Peters and Büchel, 2010). Since our study utilized a between-subject design and aimed to observe the effects of repeated EFT over the course of 1 week, we tested DD on the first visit only after participants had already developed their cues and engaged in EFT or ERT. Thus, there was no measure of a true baseline DD. In addition, although all participants in this study were screened for impulsivity using the non-planning, self-control subscale of the Barratt Impulsiveness Scale, we did not specifically recruit individuals who were high in DD. While previous works have identified a positive correlation between the Barratt Impulsiveness Scale and DD task discount rate (Cho et al., 2012), we may have been more likely to see changes in DD if we recruited a sample that was high in DD.

Following 1 week of daily engagement in EFT, our sample did not show a reduction in DD compared to a single engagement. This is contrary to a recent study by Mellis et al. which showed that repeated engagement in prospection reduced DD in a population of current and recent problem drinkers (Mellis et al., 2019). There are several major differences between our study and theirs which may explain the discrepancy. First, the population used in the Mellis et al. study was recent and current problem drinkers, who have higher DD to begin with (MacKillop et al., 2011; Moody et al., 2017). Our study used individuals with overweight and obesity but did not specifically recruit those who were high in DD. Another big difference is that in the Mellis et al. study, participants generated a new set of future cues on each visit, whereas our participants listened to the same cues that were generated on the first visit every day for a week. It is possible that our participants habituated to the imagery. Listening to a cue repeatedly may not engage the episodic memory as much as imagining details of the event to describe the event initially. Further, the Mellis et al. study had six exposures to the EFT cues spread out across 1–3 months, whereas participants in our study listened three times per day for 7 days, for a total of 21 listening sessions. Finally, the comparison condition in the Mellis et al. study was no cue and scarcity, whereas ours was a recent thinking control. A recent study reported that using ERT as a control introduces unnecessary variability into EFT interventions (Hollis-Hansen et al., 2019). People use past semantic and episodic memories to imagine their future (Hollis-Hansen et al., 2019); therefore, it is possible that our ERT group engaged in unintentional prospection. If this was the case, it is possible that the vividness of the imagery was stronger for those in the ERT group, as they were describing an event that had already occurred. Since vividness predicts the degree of reduction in DD (Peters and Büchel, 2010), this could explain why we did not show a difference between groups. Future studies should explore the effect of developing new EFT messages daily to enhance the reduction in DD and determine if this strategy reduces RRV of food or energy intake.

Another aim of this study was to determine if EFT had an effect on RRV of low-energy-dense snack foods relative to high-energy-dense snack foods. We hypothesized that participants engaging in EFT for 1 week would show a decrease in RRV of high-energy-dense foods. RRV was tested on visit 2 and again on visit 3 after participants had been engaging in EFT or ERT every day for 1 week. After the initial measurement of RRV, there was no difference in RRV of high- or low-energy-dense foods between groups. After 1 week, participants in the ERT group showed a significant reduction in RRV of low-energy-dense foods, whereas the EFT group had no change in RRV between the two visits. Previous literature shows that recalling recent food intake leads to reductions in subsequent food intake (Higgs, 2002; Higgs and Donohue, 2011; Vartanian et al., 2016). If participants chose recent events that were food related, this may explain why we did not see a difference between groups. After 1 week, participants in the ERT group showed a significant reduction in RRV of low-energy-dense foods, whereas the EFT group had no change in RRV between the two visits. The reduction in RRV over time for the ERT condition may be due to fatigue, habituation, or reduction in vividness of the event. The fact that the EFT group maintained the same level of RRV for low-energy-dense food may suggest that they are more motivated to consume healthy snacks or, at least, that the EFT prevented a reduction in the RRV of low-energy-dense foods over time.

Another main goal of this study was to investigate energy intake and snack food preference at an ad libitum buffet. The hypothesis was that those in the EFT group would consume less energy during a snack food buffet compared to those in the ERT group at baseline, and that those in the EFT group would consume less energy at a buffet following 1 week of EFT. However, the results did not show a difference in energy consumed between the EFT and ERT condition on the first visit. Further, participants in the EFT condition consumed significantly more energy (p = .015) at the snack food buffet following 1 week of EFT. This study did not screen participants for desire to control food intake, which has been done in previous studies that showed an effect of EFT (Daniel et al., 2013b). This may suggest that successful EFT is dependent on a pre-existing motivation to change eating behavior.

Our post hoc linear mixed model of the daily listening data showed a trend for participants in the EFT group to increase in DD the more they listened to their recordings. We suspect that this may be due to the initial effects of the EFT intervention wearing off over time. However, this did not reach statistical significance, and we cannot confirm this finding without a true baseline measurement of DD. It is possible that developing new vignettes daily would be the most effective method for daily thought training.

When interpreting the results from this study, it is important to consider the strengths and weaknesses. Strengths include randomization to a control or experimental group prior to the initial visit. Participants in the control group followed identical procedures to those in the experimental group, the only exception being the time frame of their episodic thinking. This study also included males and females, which allowed us to explore potential sex differences. We also included a broad age range (19–48 years), allowing greater generalizability of the results. Another strength of this study was the use of objective measures for DD, RRV, and energy intake measured in the laboratory and verifiable, smartphone-based implementation of the intervention. In addition, the experimental intervention used in this study was straightforward and easy to implement, reflecting practical use.

The weaknesses of this study must also be considered. One weakness of this study was the relatively small sample size (N = 33). While it is possible that we would have found significant differences with a larger sample size, the groups were similar in most of the outcome measures with no evidence of statistical trends. In addition, participants were all volunteers, introducing the potential for a self-selection bias. Participants were instructed to write short paragraphs describing their event in vivid detail, including any sensory descriptors that would assist in imagining themselves in the event. However, they were not instructed to frame their event in the present tense (i.e. “in one month, I am on the beach”), rather their statements were written in the future tense (i.e. “in one month, I will be on the beach”). In addition, the control (ERT) events were not screened for inadvertent prospection. We did not screen participants for desire to change their behavior, and active dieting was an exclusion criteria. EFT may be more effective in individuals who have trouble meeting health-related goals, rather than those who are impulsive but have not expressed desire to change their behavior. Similarly, the EFT may have had more of an effect on energy intake or RRV of food if the prospection was specifically designed to promote prospection about health, weight or eating-related goals. Future studies will need to examine this possibility.

Summary

The aims of this study were to investigate the effects of a practical, daily EFT intervention on DD, energy intake during an ad libitum buffet, and RRV of snack food. DD, energy intake, and RRV were tested at two time points, with 1 week of daily EFT or ERT in between. In our sample, EFT did not reduce DD, RRV of snack food, or ad libitum intake. Further, repeated, daily reflections did not enhance the effect. Our results do not support our hypothesis that repeated engagement in EFT reduces DD, RRV of high-energy-dense food, and energy intake. Future studies should explore the effect of developing new EFT messages more frequently to enhance the reduction in DD and reduce potential habituation to the vignettes. A recent study found that repeated generation of EFT cues reduced DD over time compared with the control condition (Mellis et al., 2019). Repeatedly creating new EFT messages could be a part of an intervention strategy to promote prospection and elongate time perspective by having people practice this skill on a regular basis.

Footnotes

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Tegan H Mansouri

References

Appelhans

Waring

Schneider

, et al. (2012) Delay discounting and intake of ready-to-eat and away-from-home foods in overweight and obese women. Appetite 59(2): 576–584.

Bandura

(1986) The explanatory and predictive scope of self-efficacy theory. Journal of Social and Clinical Psychology 4: 359–373.

Borghans

Golsteyn

BHH

(2006) Time discounting and the body mass index: Evidence from The Netherlands. Economics and Human Biology 4: 39–61.

Carr

Lin

Fletcher

, et al. (2014) Food reinforcement, dietary disinhibition and weight gain in nonobese adults. Obesity 22(1): 254–259.

Centers for Disease Control and Prevention (CDC) (2019) Adult BMI calculator. Available at: https://www.cdc.gov/healthyweight/assessing/bmi/adult_bmi/english_bmi_calculator/bmi_calculator.html

Cho

Pellecchia

Aminian

, et al. (2012) Morphometric correlation of impulsivity in medial prefrontal cortex. Brain Topography 26: 479–487.

Connell

Nord

Lofton

, et al. (2004) Food security of older children can be assessed using a standardized survey instrument. Journal of Nutrition 134(10): 2566–2572.

Conway

Ingwersen

Moshfegh

(2004) Accuracy of dietary recall using the USDA five-step multiple-pass method in men: An observational validation study. Journal of the American Dietetic Association 104(4): 595–603.

Daniel

Said

Stanton

, et al. (2015) Episodic future thinking reduces delay discounting and energy intake in children. Eating Behaviors 18: 20–24.

10.

Daniel

Stanton

Epstein

(2013a) The future is now: Comparing the effect of episodic future thinking on impulsivity in lean and obese individuals. Appetite 71: 120–125.

11.

Daniel

Stanton

Epstein

(2013b) The future is now: Reducing impulsivity and energy intake using episodic future thinking. Psychological Science 24(11): 2339–2342.

12.

Dassen

FCM

Jansen

Nederkoorn

, et al. (2016) Focus on the future: Episodic future thinking reduces discount rate and snacking. Appetite 96: 327–332.

13.

Epstein

Wright

Paluch

, et al. (2004) Food hedonics and reinforcement as determinants of laboratory food intake in smokers. Physiology & Behavior 81(3): 511–517.

14.

Gormally

Black

Daston

, et al. (1982) The assessment of binge eating among obese persons. Addictive Behavior 7(1): 47–55.

15.

Hales

Carroll

Fryar

, et al. (2017) Prevalence of obesity among adults and youth: United States, 2015–2016. NCHS Data Brief No. 288, October. Hyattsville, MD: National Center for Health Statistics.

16.

Higgs

Donohue

(2011) Focusing on food during lunch enhances lunch memory and decreases later snack intake. Appetite 57: 202–206.

17.

Higgs

(2002) Memory for recent eating and its influence on subsequent food intake. Appetite 39: 159–166.

18.

Hollis-Hansen

O’Donnell

Seidman

, et al. (2019) Improvements in episodic future thinking methodology: Establishing a standardized episodic thinking control. PLoS ONE 14(3): e0214397.

19.

Johnson

Driscoll

Goran

(1996) Comparison of multiple-pass 24-hour recall estimates of energy intake with total energy expenditure determined by the doubly labeled water method in young children. Journal of the American Dietetic Association 96(11): 1140–1144.

20.

Koffarnus

Bickel

(2014) A 5-trial adjusting delay discounting task: Accurate discount rates in less than 60 seconds. Experimental and Clinical Psychopharmacology 22(3): 222–228.

21.

MacKillop

Amlung

Few

, et al. (2011) Delayed reward discounting and addictive behavior: A meta-analysis. Psychopharmacology 216: 305–321.

22.

Mellis

Snider

Deshpande

, et al. (2019) Practicing prospection promotes patience: Repeated episodic future thinking cumulatively reduces delay discounting. Drug and Alcohol Dependence 204: 107507.

23.

Moody

Tegge

Bickel

(2017) Cross-commodity delay discounting of alcohol and money in alcohol users. Psychological Record 67: 285–292.

24.

O’Neill

Daniel

Epstein

(2016) Episodic future thinking reduces eating in a food court. Eating Behavior 20: 9–13.

25.

Patton

Stanford

Barratt

(1995) Factor structure of the Barratt Impulsiveness Scale. Journal of Clinical Psychology 6: 768–774.

26.

Peters

Büchel

(2010) Episodic future thinking reduces reward delay discounting through an enhancement of prefrontal-mediotemporal interactions. Neuron 66: 138–148.

27.

Rollins

Dearing

Epstein

(2010) Delay discounting moderates the effect of food reinforcement on energy intake among non-obese women. Appetite 55(3): 420–425.

28.

Rung

Madden

(2018b) Experimental reductions of delay discounting and impulsive choice: A systematic review and meta-analysis. Journal of Experimental Psychology: General 147(9): 1349–1381.

29.

Smith

Bogin

Bishai

(2005) Are time preference and body mass index associated? Evidence from the National Longitudinal Survey of Youth. Economics and Human Biology 3: 259–270.

30.

Spitzer

Williams

Kroenke

, et al. (1994) Utility of a new procedure for diagnosing mental disorders in primary car: The PRIME-MD 1000 study. Journal of American Medical Association 272(22): 1749–1756.

31.

Stanford

Mathias

Dougherty

, et al. (2009) Fifty years of the Barratt Impulsiveness Scale: An update and review. Personality and Individual Differences 47: 385–395.

32.

Stein

Wilson

Koffarnus

, et al. (2016) Unstuck in time: Episodic future thinking reduces delay discounting and cigarette smoking. Psychopharmacology 233(21–22): 3771–3778.

33.

Stunkard

Messick

(1985) The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research 291: 71–83.

34.

Sze

Daniel

Kilanowski

, et al. (2015) Web-based and mobile delivery of an episodic future thinking intervention for overweight and obese families: A feasibility study. JMIR Mhealth Uhealth 3(4): e97.

35.

Vartanian

Chen

Reily

, et al. (2016) The parallel impact of episodic memory and episodic future thinking on food intake. Appetite 101: 31–36.