Devaluing Foods to Change Eating Behavior

Excessive eating of energy-dense foods and obesity are risk factors for a range of cancers. There are programs to reduce intake of these foods and weight loss, but the effects of the programs rarely last. This project tests whether altering the value of cancer-risk foods can create lasting change, and uses neuroimaging to compare the efficacy of two programs to engage the valuation system on a neural level. Results will establish the pathways through which the programs work and suggest specific treatments for individuals based on a personalized profile.

Obesity and intake of certain foods increase cancer risk, but the most common treatment (behavioral weight loss programs) rarely produces lasting weight loss and eating behavior change, apparently because caloric restriction increases the reward value of food and prompts energy-sparing adaptations. Interventions that reduce the implicit valuation of cancer-risk foods (e.g., red meats, refined sugar) may be more effective. Emerging data suggest that behavioral response training and cognitive reappraisal training reduce valuation of such foods, which leads to decrease intake of these foods and weight loss. Internalized incentive value is reflected in a ventromedial prefrontal cortex (vmPFC) / orbitofrontal cortex valuation system, which encodes the implicit reward value of food and is central to a reinforcement cycle that perpetuates unhealthy eating. Thus, the vmPFC valuation system is a promising target for intervention because changes to the system might disrupt the unhealthy reinforcement cycle. Interestingly, various interventions influence the vmPFC through distinct pathways. Behavioral training alters motor input to valuation regions, whereas cognitive training relies on lateral prefrontal "top-down" regions. The proposed translational neuroscience experiment will compare the efficacy with which two novel treatments cause lasting change in food valuation, and whether a composite of theory-based baseline individual differences in relevant processes (such as response tendencies and cognitive styles) moderate treatment effects. We will randomize 300 overweight/obese adults who are at risk for eating- and obesity-related cancers to behavioral response training toward healthy foods and away from cancer-risk foods, a cognitive reappraisal intervention focused on cancer-risk foods, or non-food inhibitory control training. Aim 1 compares the efficacy and mechanisms of action of these two interventions to reduce valuation of cancer-risk foods relative to the active control condition, using neural, behavioral, self-report, and physiological measures of the process and outcomes. Aim 2 is to establish the temporal pattern and durability of the effects across time; food intake and habits, body fat, BMI, and waist-to-hip ratio will be measured pre, post, and at 3-, 6-, and 12-month follow-up. Aim 3 uses machine learning to build and validate a low-cost, easy-to-administer composite that predicts whether and for how long an individual is likely to respond to intervention, and to which treatment. We hypothesize that self-report measures specifically related to valuation (e.g., willingness-to-pay) and to intervention-specific pathways to valuation (e.g., behavioral response tendencies, cognitive style) will predict differential response. Discovering these individual differences will provide a practical, low-cost tool to help interventionists "match" a given person to an effective treatment for that person. This project is very innovative because no study has directly compared the distinct and common effects of these treatments on valuation, used brain imaging to study the mechanism of effects, tested whether these interventions produce a lasting change in food valuation and body fat, or built and validated a composite that moderates response.

eating, overweight, obesity, cancer risk, translational neuroscience, cognitive reappraisal, functional magnetic resonance imaging (fMRI), vmPFC, inhibitory control, valuation

Quantify the degree to which cognitive and behavioral interventions alter the valuation of cancer-risk foods relative to an active control. We will recruit 300 overweight/obese adults who are at risk for eating- and obesity-related cancers and randomize them to a (a) behavioral response training toward low cancer-risk foods and away from high cancer-risk foods, (b) cognitive reappraisal intervention focused on cancer-risk foods (experimental arms), or (c) non-food inhibitory control training (active control arm). Valuation, our primary mediating process as implicated in the incentive sensitization model, will be measured using behavioral economics tasks and functional magnetic resonance imaging (fMRI) of the vmPFC at pre- and posttraining. Proximal, intervention-specific mediators will also be indexed with fMRI. A final analysis will compare the potency of the intervention-specific neural systems to alter valuation via connectivity to vmPFC.

In Arm 1 of Devaluing energy-dense foods for cancer-control, participants will complete computer delivered versions of the stop-signal, go/no-go, and dot-probe training tasks in 8 30-min biweekly visits to the lab, with breaks between training blocks in which participants sit with their eyes closed to allow consolidation of learning. Participants will also complete a weekly 15-min training task online from home. Total training time = 345 min. Training will involve 100 images of cancer risk foods that participants regularly eat, including red and processed meats; high-sugar foods; heavily salted, smoked, and pickled foods; fries, chips, and snacks with trans-fats, and 100 images of healthy foods that participants rate as palatable, including vegetables, fruits, nuts, and whole grains.

Arm 2 of the Devaluing energy-dense foods for cancer-control intervention will be delivered via computer-assisted in-person training. Between baseline and endpoint sessions, participants will practice reappraisal on a computer, under close supervision of a facilitator, in 8 30-min twice-weekly individual sessions. During sessions, participants will practice cognitive reappraisal to reduce the value of cancer risk foods. Participants will also practice reappraisal of cancer risk foods on a computer at home, twice weekly for 15 minutes, for a total intervention time of contact of 345 minutes. The facilitator will review homework completed by participants and offer corrective feedback. The home practice is intended to promote generalization of use of this skill in the natural environment.

In Arm 3 (active control) of the Devaluing energy-dense foods for cancer-control intervention will be identical in duration and contact time to the behavioral response training described above (345 min total), but will involve nonfood images (birds and flowers), as described in the pilot trial. Participants will be informed that this intervention is designed to improve response inhibition, which should lead to eating change and weight loss given that impulsivity increases the risk for overeating, ensuring the credibility of the control arm.

A 3-arm randomized controlled trial experiment study over 12 months. At baseline, participants will complete behavioral, neural, and self-report measures related to food, specifically measures of food valuation and of the proximal neural systems hypothesized to be linked to each of the 2 experimental arms. We will also measure food intake and body composition at baseline. Then participants will be randomized to one of 3 arms (2 experimental + 1 active control) for 8 30-min sessions to occur twice weekly at the University of Oregon for 30 days. At endpoint (~1 month following baseline), all behavioral, neural, and self-report measures will be reassessed, as will eating, habit, and body composition measures. Follow-ups at 3, 6, and 12 months will assess all measures except neuroimaging.

Assessed with the Automated Self-Administered 24-Hour (ASA24) Dietary Assessment Tool The National Cancer Institutes's standard self-assessment instrument to comprehensively measure food intake.

Barratt Impulsivity self-report questionnaire, measuring the construct of impulsivity. There are three subscales: Attentional impulsivity (8 items), motor impulsivity (10 items) non-planning impulsivity (12 items). Participants respond to each item on a 1-to-4 Likert scale and scores are averaged within subscales (yielding three 1-to-4 average scores) then averaged across the three subscales to yield one 1-to-4 overall score. Higher scores indicate higher impulsivity, which is a worse outcome.

Restraint Scale self-report questionnaire. This questionnaire measures the construct of dietary restraint. There are 2 subscales: concern for dieting and weight fluctuations. Participants answer 6 questions about concern for dieting (1-to-5) that are averaged to create a 1-to-5 score on dieting concern. Dieting concern is expected to be u-shaped in terms of better or worse, where no concern or extreme concern is worse and moderate concern is better. Participants answer 4 questions about weight fluctuations (1-to-5) that are averaged to create a 1-to-5 score for weight fluctuation. Great fluctuation is a worse outcome.

Need for Cognition self-report questionnaire, which measures the construct of cognitive engagement and enjoyment of thinking. Participants complete 18 items on a 9-point Likert scale (-4 to +4) and scores are averaged across all items to create a single score that ranges from -4 to +4. Higher scores indicate a better outcome, indicating more enjoyment of thinking processes.

Craving Regulation Scale self-report questionnaire, which measures the construct of self-regulation of food cravings. There are 24 items total, with 4 items within each of 6 subscales: avoidance of temptation, controlling temptations, distraction, suppression, goal/rule setting, and goal deliberation. Responses are on a 1-to-5 Likert scale and averaged within subscales to create 6 1-to-5 average ratings. Those six averages are also averaged to create an overall score. Greater scores indicate better self-regulation of craving, which is a desired outcome.

Food version of the Self-Report Habit Index self-report questionnaire. This measures the construct of habitual eating of healthy and unhealthy foods. The scale contains two subscales: healthy foods and unhealthy foods. Each subscale contains 12 items, and responses are on a 1-to-5 Likert scale. Responses are averaged within each subscale to create 1-to-5 average ratings for habitual eating of healthy and unhealthy foods, respectively. The subscales are reported separately and not combined. Greater numbers indicate more habitual eating, so lower averages on the unhealthy subscale and higher averages on the healthy subscale indicate a better outcome.

Change from Baseline Cancer Risk and Healthy Food Craving and Valuation at 1 month, Self-report Questionnaire 2

Food Craving Inventory self-report questionnaire measuring craving and valuation in dollars per serving of cancer risk and healthy foods. There are 28 items on each subscale (one for craving and one for valuation), and the items are averaged within each subscale. The range of the craving scale is 1-5 (i.e., average of 28 1-to-5 Likert ratings) and the range of the valuation scale is 1-4 (i.e., average of 28 1-to-4 Likert ratings). The subscales are reported separately and not combined. Greater numbers indicate more craving / value of the unhealthy foods, so lower numbers indicate a better outcome.

Change from Baseline Behavioral Response Biases Toward and Away from Cancer Risk and Healthy Foods at 1 month, Behavioral marker, Task 1

Change from Baseline Behavioral Response Biases Toward and Away from Cancer Risk and Healthy Foods at 1 month, Behavioral marker, Task 2

Change from Baseline Behavioral Response Biases Toward and Away from Cancer Risk and Healthy Foods at 1 month, Neural marker, Task 1

Premotor, basal ganglia, dorsal cingulate, and Thalamus Activity during standard inhibitory control task (Stop-Signal) with personal risk cues

Change from Baseline Behavioral Response Biases Toward and Away from Cancer Risk and Healthy Foods at 1 month, Neural marker, Task 2

Premotor, basal ganglia, dorsal cingulate, and Thalamus Activity during standard inhibitory control task (Go/No-Go) with personal risk cues

Dorsolateral Prefrontal Cortex and ventrolateral Prefrontal Cortex activity during Regulation of Craving Task for Food

Inclusion Criteria: - overweight to obese range (BMI 25-35) Exclusion Criteria: metal implants (e.g., braces, permanent retainers, pins) metal fragments, pacemakers or other electronic medical implants claustrophobia weight ˃ 550 lbs. Women who are pregnant or believe they might be pregnant people who have been diagnosed with past or current medical, psychiatric, neurological, eating disorders, or are taking psychotropic medications urine screen to exclude participants who are acutely intoxicated screen for handedness Beyond these criteria, participants will be recruited without exclusions based on gender, race, or ethnicity, so our sample will reflect the diversity in the local population (Lane County, Oregon) with regard to gender, race, and ethnicity.

Berkman ET, Burklund L, Lieberman MD. Inhibitory spillover: intentional motor inhibition produces incidental limbic inhibition via right inferior frontal cortex. Neuroimage. 2009 Aug 15;47(2):705-12. doi: 10.1016/j.neuroimage.2009.04.084. Epub 2009 May 6.

Berkman ET, Falk EB. Beyond Brain Mapping: Using Neural Measures to Predict Real-World Outcomes. Curr Dir Psychol Sci. 2013 Feb;22(1):45-50. doi: 10.1177/0963721412469394.

Berkman ET, Falk EB, Lieberman MD. In the trenches of real-world self-control: neural correlates of breaking the link between craving and smoking. Psychol Sci. 2011 Apr;22(4):498-506. doi: 10.1177/0956797611400918. Epub 2011 Mar 4.

Berkman ET, Kahn LE, Merchant JS. Training-induced changes in inhibitory control network activity. J Neurosci. 2014 Jan 1;34(1):149-57. doi: 10.1523/JNEUROSCI.3564-13.2014.

Giuliani NR, Calcott RD, Berkman ET. Piece of cake. Cognitive reappraisal of food craving. Appetite. 2013 May;64:56-61. doi: 10.1016/j.appet.2012.12.020. Epub 2013 Jan 9.

Giuliani NR, Mann T, Tomiyama AJ, Berkman ET. Neural systems underlying the reappraisal of personally craved foods. J Cogn Neurosci. 2014 Jul;26(7):1390-402. doi: 10.1162/jocn_a_00563. Epub 2014 Jan 6.

Giuliani NR, Tomiyama AJ, Mann T, Berkman ET. Prediction of daily food intake as a function of measurement modality and restriction status. Psychosom Med. 2015 Jun;77(5):583-90. doi: 10.1097/PSY.0000000000000187.

Stice E, Burger K, Yokum S. Caloric deprivation increases responsivity of attention and reward brain regions to intake, anticipated intake, and images of palatable foods. Neuroimage. 2013 Feb 15;67:322-30. doi: 10.1016/j.neuroimage.2012.11.028. Epub 2012 Nov 28.

Stice E, Lawrence NS, Kemps E, Veling H. Training motor responses to food: A novel treatment for obesity targeting implicit processes. Clin Psychol Rev. 2016 Nov;49:16-27. doi: 10.1016/j.cpr.2016.06.005. Epub 2016 Jul 21.

Stice E, Marti CN, Spoor S, Presnell K, Shaw H. Dissonance and healthy weight eating disorder prevention programs: long-term effects from a randomized efficacy trial. J Consult Clin Psychol. 2008 Apr;76(2):329-40. doi: 10.1037/0022-006X.76.2.329.

Stice E, Presnell K, Gau J, Shaw H. Testing mediators of intervention effects in randomized controlled trials: An evaluation of two eating disorder prevention programs. J Consult Clin Psychol. 2007 Feb;75(1):20-32. doi: 10.1037/0022-006X.75.1.20.

Stice E, Rohde P, Durant S, Shaw H. A preliminary trial of a prototype Internet dissonance-based eating disorder prevention program for young women with body image concerns. J Consult Clin Psychol. 2012 Oct;80(5):907-16. doi: 10.1037/a0028016. Epub 2012 Apr 16.

Stice E, Rohde P, Gau J, Shaw H. An effectiveness trial of a dissonance-based eating disorder prevention program for high-risk adolescent girls. J Consult Clin Psychol. 2009 Oct;77(5):825-34. doi: 10.1037/a0016132.

Stice E, Rohde P, Shaw H, Gau J. An effectiveness trial of a selected dissonance-based eating disorder prevention program for female high school students: Long-term effects. J Consult Clin Psychol. 2011 Aug;79(4):500-8. doi: 10.1037/a0024351.

Stice E, Yokum S, Burger K, Rohde P, Shaw H, Gau JM. A pilot randomized trial of a cognitive reappraisal obesity prevention program. Physiol Behav. 2015 Jan;138:124-32. doi: 10.1016/j.physbeh.2014.10.022. Epub 2014 Oct 30.

Stice E, Yokum S, Veling H, Kemps E, Lawrence NS. Pilot test of a novel food response and attention training treatment for obesity: Brain imaging data suggest actions shape valuation. Behav Res Ther. 2017 Jul;94:60-70. doi: 10.1016/j.brat.2017.04.007. Epub 2017 Apr 19.

Fisher PA, Berkman ET. Designing Interventions Informed by Scientific Knowledge About Effects of Early Adversity: A Translational Neuroscience Agenda for Next Generation Addictions Research. Curr Addict Rep. 2015 Dec 1;2(4):347-353. doi: 10.1007/s40429-015-0071-x. Epub 2015 Sep 28.