Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Jan 1.
Published in final edited form as: Nurs Res. 2023 Oct 31;73(2):91–100. doi: 10.1097/NNR.0000000000000702

Randomized Controlled Trial of Effects of Behavioral Weight Loss Treatment on Food Cue Reactivity

Ariana M Chao 1, Thomas A Wadden 2, Wen Cao 3, Yingjie Zhou 4, Delphina Maldonado 5, Michelle I Cardel 6, Gary D Foster 7, James Loughead 8
PMCID: PMC10922238  NIHMSID: NIHMS1938325  PMID: 37916843

Abstract

Background:

It is not known whether behavioral weight loss can attenuate blood oxygen level-dependent responses to food stimuli.

Objectives:

This randomized controlled trial assessed the effects of a commercially available behavioral weight loss program (WeightWatchers) compared to a wait-list control on blood-oxygen-level-dependent response to food cues.

Methods:

Females with obesity (N = 61) were randomized to behavioral weight loss or wait-list control. At baseline and follow-up, participants completed assessments that included functional magnetic resonance imaging scans to assess response to images of high-calorie foods (HCF) or low-calorie foods (LCF), and neutral objects.

Results:

There were no significant between-group differences in change from baseline to follow-up in any regions of the brain in response to viewing HCF or LCF. From baseline to follow-up, participants in behavioral weight loss, compared with wait-list control, reported significantly greater increases in desire for LCF. Changes in liking and palatability of LCF and liking, palatability, and desire for HCF did not differ between groups.

Discussion:

Behavioral weight loss was associated with increased desire for LCF without changes in neural reactivity to food cues. These results suggest that alteration of neurological processes underlying responsiveness to food is difficult to achieve through behavioral weight management alone.

Keywords: appetite, craving, magnetic resonance imaging, obesity

The modern obesogenic environment—filled with cues to consume highly palatable and energy-dense foods—is a major contributor to obesity. In part, the visual system drives food selection (Spence et al., 2016). The theory of Pavlovian conditioning suggests that cues, such as seeing food or food images, can become conditioned stimuli when repeatedly paired with food consumption (van den Akker et al., 2018). The conditioned stimuli can then elicit a wide range of psychological, physiological, and neurocognitive appetitive responses (Pursey et al., 2014; Yokum et al., 2011). These responses tend to be stronger in people with obesity relative to those without it (Kanoski & Boutelle, 2022; Meyer et al., 2015). The brain is a central organ responsible for body weight regulation and response to food stimuli (Hall et al., 2022). Individual differences in neural responses to visual food cues have been implicated as a contributor to overeating and weight status (Pursey et al., 2014).

Adults with obesity versus individuals without it demonstrate heightened food cue reactivity (Pursey et al., 2014). People with obesity, relative to those without it, have greater responsiveness to food images compared to nonfood stimuli in regions of the brain associated with reward and related functions including the insula (Rothemund et al., 2007; Scharmüller et al., 2012; Stoeckel et al., 2008), orbitofrontal cortex (Dimitropoulos et al., 2012; Scharmüller et al., 2012; Stoeckel et al., 2008), anterior cingulate (Martens et al., 2013; Martin et al., 2010; Stoeckel et al., 2008), nucleus accumbens (Stoeckel et al., 2008), and caudate (Rothemund et al., 2007; Stoeckel et al., 2008). Some, but not all studies (Dimitropoulos et al., 2012), have shown that in people with obesity, compared with those without it, high-calorie foods (HCF) versus low-calorie foods (LCF) elicit greater activation in brain regions such as the orbitofrontal cortex (Stoeckel et al., 2008), prefrontal cortex (Stoeckel et al., 2008), anterior cingulate (Stoeckel et al., 2008), insula (Stoeckel et al., 2008), nucleus accumbens (Stoeckel et al., 2008), caudate (Stoeckel et al., 2008), and putamen (Rothemund et al., 2007). An exaggerated salience of food information and associated hyperactivity in reward centers may override homeostatic control mechanisms in individuals with obesity. These differences help to explain why some individuals are at risk for weight gain when placed in similar environmental contexts.

Understanding the neural substrates involved in eating regulation is important to improve obesity treatment. Successful weight loss is typically characterized by decreased consumption of highly palatable, energy-dense foods (Chao et al., 2021). This contributes to reductions (and hopeful extinction) of conditioned appetitive responses to food cues (Jansen, 1998). After weight loss, individuals report decreased food cravings (Chao et al., 2019) and reward-based eating (O’Neil et al., 2012). However, it is not known whether behavioral weight loss can reverse reward system hyperactivation to food or HCF stimuli, and findings have been inconsistent. In a prospective study, 25 participants with overweight/obesity were enrolled in a 12-week psychosocial weight loss program that included educational, motivational, and behavioral components (Murdaugh et al., 2012). Participants lost 3.5% of initial body weight, and they demonstrated declines in blood oxygen level-dependent (BOLD) signals to HCF versus LCF images in the medial prefrontal cortex, inferior parietal lobe, precuneus, posterior cingulate cortex, premotor cortex, and angular gyrus. However, another nonrandomized study of 16 participants with obesity who were enrolled in a behavioral weight loss intervention (involving behavioral strategies, moderate calorie restriction, and titrated physical activity) found no changes in reward system activation to food versus nonfood images (Bruce et al., 2014). A pilot randomized study of adults with overweight/obesity showed that participating in a weight-loss intervention precipitated a shift in the reward system to favor LCF versus HCF. Participants (n = 8) in the behavioral weight management program, who lost 6.3 kg at 6 months, compared to wait-list controls (n = 5), who gained 2.1 kg, had increased activation for LCF images in the right ventral putamen, as well as decreased activation for HCF images in the left dorsal putamen (Deckersbach et al., 2014). These uncontrolled and pilot randomized trial findings suggest that reward-related neural activity can be altered with treatment. However, larger randomized studies are needed that address this question.

The objective of this randomized controlled trial (RCT) was to assess the effects of a widely available behavioral weight loss program (BWL; WeightWatchers [WW]) compared to a wait-list control (WLC) on neural response to food cues. Our primary aim was to compare differences between the BWL and WLC groups at week 16 in changes from baseline in BOLD fMRI response to food images. We hypothesized that, compared to the WLC, the BWL group would have significantly greater declines in BOLD response to LCF in several reward regions of the brain (insula, orbitofrontal cortex, anterior cingulate cortex, nucleus accumbens) and significantly greater increases in BOLD response to LCF in these reward regions. Our secondary aims were to compare differences between the BWL and WLC groups at week 16 in changes in self-reported measures of appetite and to test whether changes in these self-reported measures were associated with changes in neural responsivity to HCF and LCF. We hypothesized that the BWL group, compared to the WLC, would report significantly greater declines in reward-based eating, food cravings, and hunger. We also hypothesized that in the BWL group, greater declines in BOLD responses to HCF would be associated with decreased self-reported wanting and liking of these foods and that greater increases in BOLD responses to LCF would be associated with increased wanting and liking of LCF.

Methods

Participants

Participants for this RCT (NCT04202133) were recruited in three waves using advertisements and clinician referrals. Primary inclusion criteria were being female due to evidence suggesting differences in the processing of food stimuli by sex (Chao et al., 2017), 18–60 years of age, and having a BMI of 30 kg/m2 or more. Major exclusion criteria were: weight >158.8 kg (due to scanner weight restrictions); serious medical conditions; untreated thyroid disease or any changes in thyroid medication in the last 6 months; current psychiatric disorder that significantly interfered with daily living; active suicidal ideation; current substance use disorder or in remission < 1 year; presence or history of conditions that may interfere with magnetic resonance imaging; participation in a structured weight loss program in the prior 6 months; active WW member within the past 12 months; use of medications known to induce significant weight loss/gain; psychiatric hospitalization within the past 6 months; loss of ≥10 lb in the past 3 months; history or plans for bariatric surgery; impairment affecting task performance; epilepsy; and neurological trauma. This University of Pennsylvania’s Institutional Review Board reviewed and approved this study. Participants provided written informed consent.

Procedures

Following an initial telephone screening assessment, eligible participants attended an intake visit. The in-person interview was conducted by study staff who obtained informed consent and evaluated participants’ magnetic resonance imaging (MRI) and behavioral eligibility. This included screening with an MRI-eligibility checklist and assessment of the applicant’s eating and lifestyle behaviors with the Weight and Lifestyle Inventory (WALI) (Wadden & Foster, 2006) and of mood with the Beck Depression Inventory–II (Dozois et al., 1998). Individuals who remained interested and passed this assessment portion were asked about their medical history. Height, weight, blood pressure, and waist circumference were measured in duplicate. Blood pressure was measured using an automated monitor. Two readings were taken at 1-min intervals after participants had been seated for at least 5 min. Waist circumference was measured horizontally, halfway between the lowest rib and the top of the hipbone to the nearest 0.1 cm. Participants were next sent questionnaires to complete and attended a baseline study assessment visit that included an MRI. After participants completed baseline assessments, they were randomized in a 1:1 ratio to BWL or WLC with randomly varied block sizes (three, six, or nine) using sealed envelopes. Follow-up assessments occurred after the intervention or waiting period was complete. This was an open-label trial.

Study Interventions

BWL Group

Participants were treated in three cohorts of 10 to 11 participants. Cohort 1 started in January 2020, Cohort 2 in August 2020, and Cohort 3 in December 2020. We had originally planned to treat participants using 16-weekly sessions of in-person group treatment. Cohort 1 was treated in-person for nine visits until March 2020, when COVID-19 mitigation restrictions prevented in-person meetings. This cohort was converted to virtual, group-based sessions for an additional 16 weeks until we were permitted to resume MRI scanning. Cohorts 2 and 3 were each treated with 16 weeks of virtual, group-based treatment.

The WW program has been studied extensively and consistently demonstrated weight loss efficacy. Participants were provided with the WW program free of charge, including access to the app and weekly, group-based workshops. A trained coach and a guide from WW led the group workshops. The WW program provided evidence-based behavioral strategies to support members’ weight and wellness goals. The program followed a weekly curriculum accompanied by individualized behavioral goals across four main pillars (food, activity, sleep, mindset). The WW program was founded on national and international guidelines for a healthy eating pattern. The program included ZeroPoint™ foods (a list of foods that can be eaten in moderation without tracking), as well as a points system, where foods and beverages were given points value based on their caloric and nutritional characteristics (protein, fiber, fat, and sugar). Participants were given personalized dietary recommendations based on daily and weekly points, calculated based on age, height, weight, and sex. Members were encouraged to track their food and beverages, with the goal of staying within their daily and weekly points target. The WW digital app provided members with a self-guided, personalized weight loss and wellness plan that included a weekly check-in and progress report, as well as food, activity, water, sleep, and weight trackers, meal planning tools, recipes, a food barcode scanner, guided meditations and workouts, and support from peers via Connect (a members-only social community) and 24/7 access to chat with a WW-trained behavior change expert coach.

WLC Group

Participants in the WLC group were asked to stay weight stable, not to make changes in their eating and physical activity behaviors, and not seek treatment for weight or problematic eating during the waiting period. Adherence to these recommendations was assessed at the follow-up appointment, and no participants started a formal weight loss program during the waiting period. All WLC participants were offered the WW intervention after they completed the second fMRI scan.

Imaging Procedures

Study assessment visits were scheduled for the morning and lasted approximately 3 hr (60 min in the MRI scanner and 120 min to complete surveys, behavioral tasks, and physical measurements). Participants were asked to fast (including no caffeine or alcohol) for at least 8 hr before the appointment to increase the stimulus salience and create a more homogeneous hunger state across participants. Participants who required vision correction were instructed to wear contact lenses or were given MRsafe glasses in their prescription strength. Similar procedures, including an identical MRI scan, were repeated at follow-up. Participants were compensated $100 for each fMRI assessment visit.

Pre-scan

Upon arriving, participants were greeted by research staff who reviewed an MRI-eligibility checklist. Participants completed self-report measures, including the Food Cravings Questionnaire–State (FCQ–S; Cepeda-Benito et al., 2000) and visual analog scales for stress and hunger that ranged from 0 (not at all) to 100 (extremely).

Data Acquisition

We used a 3.0 Tesla Siemens Prisma scanner equipped with a 64-channel head coil. BOLD fMRI sequences included automatic higher-order shimming and both prospective and retrospective motion correction. Images were acquired using a whole-brain, single-shot gradient-echo (GE) echoplanar sequence with the following parameters: TR/TE = 2000/30 ms, FOV = 200 mm, matrix = 64 × 64, slice thickness/gap = 2.0/0 mm, 78 slices, and voxel size of 2.1 × 2.1 × 2 mm. Prior to BOLD fMRI, a 5-min magnetization-prepared, rapid acquisition gradient-echo T1-weighted image (MPRAGE, TR/TE= 2200/4.67 ms, FOV 240 mm, matrix 192 × 256, effective voxel resolution of 1 × 1 × 1 mm) was acquired for anatomic overlays of functional data and to aid spatial normalization to standard MNI space.

Food Image Task

The food image task was composed of five blocks of HCF (e.g., bacon, ice cream, cake, pizza), five blocks of LCF (e.g., apples, broccoli, banana), and five blocks of neutral images (i.e., office supplies). Each block was 30s in duration and presented in pseudorandom order. Each block contained five photographs of food presented for 5s and separated by 1-s interstimulus interval (fixation point). Food blocks were followed by a 30-s rest period. Participants were asked to imagine how much they “want the object right now.” Stimuli were digital photographs of food items depicted in the ready-to-eat state (e.g., all packaging was removed). Images were from public databases (Blechert et al., 2014; Foroni et al., 2013).

Post-scan

After the fMRI was completed, to assess the response to the tasks using the food images, participants were asked about liking and wanting each food right now, as well as palatability of the 50 presented foods. Ratings were made using visual analog scales that ranged from 0 (not at all) to 100 (very much).

Image Preprocessing

BOLD time series data were preprocessed using standard image analysis procedures using fMRI Expert Analysis Tool Neuropsychopharmacology (FEAT version 6.00) of FSL (FMRIB’s Software Library, Oxford, UK). Preprocessing included motion correction, slice time correction (interleaved), skull stripping, spatial smoothing (6 mm FWHM), and high pass filtering (cut off = 100s). The median functional volume was coregistered to the anatomical T1-weighted structural volume (bias-corrected and skull-stripped using Advanced Normalization Tools) and transformed into standard anatomical space (T1 MNI 1mm template). Transformation parameters were later applied to statistical maps. Time-series analysis was carried out using general linear models. High-calorie, low-calorie, and neutral blocks were modeled separately using a canonical hemodynamic response function with rest conditions treated as baseline. Confounding explanatory variables, such as temporal derivative and motion parameters (standard + extended), were also included in the model. Contrast maps for each condition were generated. A priori regions of interest (ROIs) included insula, orbitofrontal cortex, anterior cingulate cortex, and nucleus accumbens derived from AAL3 (Automated anatomical labelling atlas 3; Rolls et al., 2020). Mean percent signal change for a priori ROIs was exported for statistical analysis.

Questionnaires

Prior to the fMRI assessment, participants completed questionnaires that included the Reward-Based Eating Drive Scale (Mason et al., 2017), Food Craving Questionnaire-Trait Questionnaire (FCQ–T; Cepeda-Benito et al., 2000), and Eating Inventory (Stunkard & Messick, 1985). We used the 13-item Reward-Based Eating Drive Scale to assess reward-driven eating, including lack of satiety, preoccupation with food, and loss of control over eating (Mason et al., 2017). Items were assessed on a Likert scale from 0 (strongly disagree) to 4 (strongly agree). A score was computed by summing the items with scores ranging from 0 to 52. Higher scores reflect higher reward-based eating drive. At baseline, the Cronbach’s alpha was .87.

Food cravings were measured with the 39-item FCQ–T, which asks participants to indicate how frequently they experience food cravings on a scale ranging from 1 (never) to 6 (always). The FCQ–T assesses craving as a multifactorial concept and has nine subscales measuring food cravings in relation to intentions and plans to consume food, anticipation of positive reinforcement from eating, anticipation of relief from negative states due to eating, lack of control over eating, thoughts or preoccupation with food, hunger, emotions experienced before or during food cravings, cues that trigger cravings, and guilt from cravings or giving into them. The Cronbach’s alphas ranged from .71 to .94 for the subscales.

The 51-item Eating Inventory was used to assess restraint, disinhibition, and hunger (Stunkard & Messick, 1985). The restraint scale includes 21 items. Higher scores demonstrate more awareness of one’s eating and success in restricting dietary intake for weight control. The Cronbach’s alpha for the scale was .79. The disinhibition scale includes 16 items. Higher scores signify more overeating tendencies. The Cronbach’s alpha was .67. The hunger scale has 14 items and higher scores indicate more perceived hunger. In this sample, the Cronbach’s alpha for the scale was .79.

Power

Power for this study was based on current standards in neuroimaging studies. Previous fMRI studies that used groups of 20 subjects per condition were sufficient to detect significant differences in BOLD signal response to behavioral tasks (Simmons et al., 2011). Our sample size allowed us to detect regions with medium to large effect sizes (d > .5) in BOLD signal between groups with 80% power (Geuter et al., 2018).

Statistical Analyses

Descriptive statistics summarized baseline and demographic characteristics. Baseline differences in BOLD response between stimuli classes (e.g., food vs. nonfood images and HCF vs. LCF) were examined using paired t-tests. Baseline differences between treatment groups’ BOLD responses to image classes were examined using independent-sample t-tests. Statistical significance for baseline differences was p < .05.

We tested changes in each ROI using general linear mixed models with repeated effects with the time variables (baseline vs. follow-up), group (BWL vs. WLC), and their interaction. Trial contrasts were assessed using the subtraction method (i.e., HCF minus neutral stimuli; LCF minus neutral stimuli). We also performed analyses with the individual stimulus type (i.e., HCF alone, LCF alone). Changes in self-report and behavioral variables were examined using general linear mixed models with repeated effects. We examined differences among groups in brain-behavior relationships using correlations between BOLD activation changes in the ROIs and changes in self-report measures. For the primary analyses, p-values were adjusted for multiple testing of the ROIs using the Dubey/Armitage-Parmar Bonferroni method (Sankoh et al., 1997), as implemented in the Simple Interactive Statistical Analysis online calculator (https://www.quantitativeskills.com/sisa/calculations/bonfer.php?; Alpha = 0.05 & N = 08 & Corr = 0.40 & Df = 00). Based on this calculation, primary outcomes’ significance level was p = .01. Statistical significance was defined as p = .05 for secondary exploratory outcomes.

Results

Participants had a mean±SD age of 41.1±9.9 yr, weight of 104.7±19.6 kg, and BMI of 38.6±6.2 kg/m2 (Table 1). Most of the sample self-identified as Black (50.8%) or White (42.6%). Demographic characteristics did not differ significantly by group.

Table 1.

Demographic and clinical characteristics (Mean ± SD or N (%))

BWL (n=31) WLC (n=30) p-Value
Age, yr 42.0 ± 8.7 40.1 ± 11.1 .47
Gender, female 31 (100) 30 (100) >.99
Ethnicity .97
Hispanic/Latino 3 (9.7) 3 (10.0)
Race .26
American Indian 0 (0) 0 (0)
Asian 0 (0) 2 (6.7)
Black 16 (51.6) 15 (50.0)
White 13 (41.9) 13 (43.3)
Multiracial 2 (6.5) 0 (0)
Number of persons living in home 3.0±1.3 2.6±1.2 .30
BMI, kg/m2 38.6 ± 5.9 38.6 ± 6.6 .99
Weight, kg 107.6 ± 21.2 101.8 ± 17.5 .25
Waist circumference, cm 110.7 ± 11.6 108.9 ± 11.8 .56
Systolic blood pressure, mm Hg 126.5 ± 12.8 126.6 ± 14.7 .99
Diastolic blood pressure, mm Hg 72.4 ± 7.4 73.0 ± 9.0 .75
Beck Depression Inventory score 6.9 ± 7.4 6.8 ± 7.1 .95
First age overweight by 10 lbs or more, yr 18.1 ± 9.2 15.3 ± 7.2 .20
Weight loss goal, lbs 60.2 ± 31.0 53.8 ± 24.9 .38
Total alcohol consumption, drinks per week 1.6 ± 2.4 3.0 ± 3.8 .10
Self-reported physical activity, minutes per week 91.5 ± 100.7 117.6 ± 136.7 .40
Open in a new tab

Note. Yr=Year. Lbs=pounds. BWL=behavioral weight loss. WLC=wait-list control

Baseline Results

BOLD Response

Across participants, at baseline, BOLD activation in response to viewing food (combination of HCF and LCF) versus neutral images was significantly higher in the left insula (p = .01), right nucleus accumbens (p = .004), and left orbitofrontal cortex (p = .049). There were no significant differences in BOLD activation between viewing food versus neutral images for the left and right anterior cingulate, right insula, left nucleus accumbens, or right orbitofrontal cortex (p > .05). At baseline, no significant differences were observed between groups in BOLD response to food images in any of the brain regions examined (p > .05). The BWL group, compared to the WLC group had greater activation to neutral cues in the right nucleus accumbens (p = .04).

Across participants at baseline, no differences were observed in BOLD response in any ROI to viewing HCF versus LCF (p > .05). No significant differences were observed between groups in BOLD response to HCF and LCF in any ROI.

Liking, Wanting, and Palatability

At the baseline MRI scan, the mean±SD score for hunger was 43.0±29.0, and stress was 26.8±28.9 (out of 100). Total state food cravings were 35.5±12.7 (out of 75). Scores for liking, desire, and palatability of foods were 73.5±8.7, 52.2±16.1, and 67.0±11.2 (out of 100), respectively. The desire for LCF was 55.4±16.4 which was significantly higher than desire for HCF, 49.0±21.4 (p = .02). The mean score for liking HCF was 73.2±12.5, which was similar to scores for liking LCF which was 73.8±12.6 (p = .82). Palatability for LCF was 66.2±13.0 which was similar to palatability HCF 67.8±14.9 (p = .47). There were no significant differences between treatment groups on baseline scores for liking, desire, and palatability of LCF and HCF (p > .05).

Self-Reported Appetite

Baseline scores for self-reported appetite questionnaires are presented in Table S1. Scores did not differ significantly by group at baseline (p > .05).

End-of-Trial Results

Body Weight

Figure 1 shows the Consolidated Standards of Reporting Trials flow diagram. The study retention rate was 91.8%, with 83.6% of participants completing the follow-up MRI. The WLC group gained 0.2±3.5% of initial body weight, compared with a loss of 4.7±4.9% for the BWL group (p < .001). The average weight loss for Cohort 1 was 6.1±4.0% at week 25, for Cohort 2 was 6.7±4.9% at week 16, and for Cohort 3 was 0.3±3.5% at week 16. BWL participants attended an average of 97.6% of sessions. Changes in hunger, stress, and state food cravings from baseline to follow-up did not differ significantly by group (p = .30, .37, .99, respectively).

Figure 1.

Figure 1.

Open in a new tab

CONSORT Flow Diagram.

Notes: BWL=behavioral weight loss. WLC=wait-list control.

BWL and WLC did not differ significantly in the days (mean±SD) between the pre-test scan to starting treatment/wait-list (BWL: 13.9±15.9 vs WLC: 17.4±17.1 days; p = .42), and the groups did not differ significantly in the time between completing the treatment/wait-list and their post-test scan (BWL: 1.1±15.6 vs. WLC: 0.5±13.2 days; p=.88).

BOLD Response

There were no significant between-group differences in any of the ROIs for change from baseline to follow-up for the contrasts of HCF minus neutral objects or of LCF minus neutral objects (Table 2; p > .01). No ROIs demonstrated significant between-group differences in changes in activation to HCF alone or LCF alone from baseline to follow-up (p > .01; Table S2).

Table 2.

Mean percentage signal change from baseline to follow-up for the contrasts of HCF minus neutral objects and LCF minus neutral objects for each group (Mean ± SE)

HCF minus Neutral LCF minus Neutral
BWL WLC p-Valuea BWL WLC p-Valuea
L anterior cingulate cortex −0.05±0.05 −0.16±0.06* .15 0.00±0.05 0.00±0.06 .99
R anterior cingulate cortex −0.06±0.05 −0.01±0.06 .48 0.03±0.04 −0.03±0.05 .44
L insula 0.03±0.03 −0.07±0.03* .02 0.02±0.03 −0.00±0.03 .53
R insula 0.01±0.04 −0.03±0.04 .44 0.02±0.03 −0.07±0.04 .10
L nucleus accumbens −0.05±0.07 −0.23±0.08* .10 −0.15±0.06* −0.13±0.06* .85
R nucleus accumbens 0.02±0.07 −0.08±0.08 .37 −0.01±0.06 −0.11±0.06 .25
L orbitofrontal cortex −0.01±0.02 −0.02±0.02 .90 −0.01±0.01 −0.02±0.02 .50
R orbitofrontal cortex 0.03±0.02 −0.02±0.02 .06 −0.00±0.01 −0.02±0.01 .26
Open in a new tab

Note. HCF=high-calorie food images. LCF=low-calorie food images. L=left. R=right. BWL=behavioral weight loss. WLC=wait-list control.

a

Uncorrected p-value for between-group differences. Threshold for statistically significant multiple comparisons corrected p-value is p<.01.

*

p<.05 for within-group change from baseline to follow-up.

Changes in Appetite

From baseline to follow-up, participants in BWL, compared with WLC, reported significant increases in desire for LCF (p = .02; Figure 2). Change in liking and palatability of LCF did not differ between groups. Change in liking, palatability, and desire for HCF did not differ significantly between groups (p > .05). BWL, compared with WLC, reported greater improvements in cognitive restraint (p = .001; Table 3), control overeating (p = .04), and anticipation of positive reinforcement that may result from eating (p = .03). Changes in other subscales for food cravings, reward-based eating drive, hunger, and disinhibition did not differ between groups (p > .05). There were no significant correlations between the change in liking of, desire for, or palatability of food stimuli and change in ROIs in response to these stimuli (p > .05).

Figure 2.

Figure 2.

Open in a new tab

Mean ± SE change in liking, palatability, and desire for low-calorie and high-calorie foods.

Notes: p-values are for between-group differences in change from baseline to follow-up.

Table 3.

Mean±SE changes in appetite at follow-up, as measured from baseline

BWL WLC p-Value for Change
Reward-based eating drive −3.9 ± 1.4 −0.7 ± 1.5 .12
Cognitive restraint +3.5 ± 0.6 +0.3 ± 0.7 .001
Hunger −0.9 ± 0.4 −0.7 ± 0.5 .81
Disinhibition −1.5 ± 0.4 −0.4 ± 0.5 .10
Food cravings- Trait
Intentions and plans to consume food −1.5 ± 0.4 −1.0 ± 0.4 .41
Lack of control over eating −2.7 ± 0.8 −0.1 ± 0.9 .04
Thoughts and preoccupation with food −1.3 ± 0.8 0.0 ± 0.9 .27
Craving as a physiological state −1.5 ± 0.5 −0.8 ± 0.5 .28
Cues that may trigger food cravings −1.5 ± 0.6 −1.1 ± 0.7 .68
Emotions that may be experienced before or during food cravings or eating −2.3 ± 0.7 −0.4 ± 0.8 .07
Guilt from cravings and/or for giving into them −0.6 ± 0.5 +0.1 ± 0.6 .35
Anticipation of positive reinforcement that may result from eating −2.9 ± 0.7 −0.6 ± 0.8 .03
Anticipation of relief from negative states and feelings as a result of eating −1.3 ± 0.5 −0.6 ± 0.5 .32
Open in a new tab

Note. Values shown are mean ± SE. BWL=behavioral weight loss. WLC=wait-list control

Exploratory Analyses

Within the BWL group, baseline BOLD response to HCF alone and LCF alone was not significantly correlated with percent weight change (p > .05). Of the BWL participants who completed a follow-up fMRI scan, 51.7% achieved a ≥5% initial weight loss. Baseline BOLD response to HCF alone and LCF alone did not differ between participants who did and did not achieve this categorical loss (p > .05). Participants who did and did not achieve a ≥ 5% weight loss also did not differ significantly in change in response to the food stimuli in any ROI.

Discussion

Contrary to our hypothesis, participants randomized to BWL, relative to those in WLC, did not demonstrate more favorable changes in BOLD response to HCF or LCF in the insula, orbitofrontal cortex, anterior cingulate cortex, or nucleus accumbens. Our study is one of the first randomized controlled trials to examine changes in neural responses to food images with BWL. Our findings are similar to other nonrandomized studies that have found minimal changes in reward-processing regions of the brain after behavioral weight loss (Bruce et al., 2014; Murdaugh et al., 2012). However, our findings contradict a previous small pilot trial (N = 13) that randomized participants to a BWL or WLC (Deckersbach et al., 2014). In that study, BWL participants, compared with WLC, had increased mean activation in the right ventral putamen and decreased responsiveness for HCF in the left dorsal putamen (Deckersbach et al., 2014). Taken together, these results suggest that suppression of neurological processes underlying motivation for and responsiveness to HCF cues is difficult to achieve.

The long-term effectiveness of weight loss treatments might be improved by strengthening extinction learning to diminish automatic reward responses to HCF. Several interventions that target food reward response have been developed and tested (Stice et al., 2017, 2022). However, the efficacy of these interventions in mitigating responses to food stimuli in reward-related regions of the brain, as well as in inducing weight loss, has been mixed. For example, a pilot study was conducted on a food-response training intervention, which included five training tasks, such as a stop-signal and go/no-go training using food images. The food-response intervention was associated with significantly greater pre–post reduction in neural responsivity to HCF versus LCF in reward-related areas of the brain, including the insula (Stice et al., 2017). However, in a larger randomized controlled trial testing this intervention, no significant differences were observed between groups in body-fat loss and neural response to HCF (Stice et al., 2022). Additional studies are needed to examine the benefits of integrating interventions focused on bottom-up, implicit processes (Strack & Deutsch, 2004) to improve neural responses to food stimuli and weight loss.

We potentially observed minimal differences in BOLD response between groups because the study was powered to detect only moderately large effects. Larger samples may have resulted in different findings. There also may be a dose-response relationship between weight loss and changes in neural activation, similar to other weight-loss-related benefits such as changes in blood sugar and blood pressure (Ryan & Yockey, 2017). A larger caloric deficit could also lead to greater food reward declines compared with a smaller deficit. Trials have also found divergent pre- to post-treatment changes in neural responses to food cues with behavioral versus surgical weight loss (Bruce et al., 2014; Salem et al., 2021). Future studies should investigate variations in neural response with different obesity treatments, such as pharmacotherapy. The duration of this study was also short, and it could be hypothesized that short-term caloric restriction may enhance food reward, whereas longer-term restriction with changes in energy stores (e.g., body fat) may attenuate it. The shift in reward for low- and high-energy foods may occur as behaviors and cognitions become internalized, habitual, and more automatic. Further studies are necessary to examine long-term changes in neural response with longer-term weight loss.

Despite the lack of changes in BOLD response to LCF, we found that the BWL group, relative to the WLC, reported increased desire for LCF, as well as greater improvements in cognitive restraint and control overeating and less anticipation of positive reinforcement that may result from eating. Importantly, we found no increase in cravings or desire for HCF or in the liking or perceived palatability of these foods. These findings are consistent with previous research showing that food cravings (Kahathuduwa et al., 2017) and food reward tend to decrease with weight loss (Oustric et al., 2018).

Limitations

Strengths of this study include the high retention rate and treatment attendance in the BWL group. Limitations include that it was a single-site study with a small sample of females and limited power to detect small effect sizes. In addition, attrition was greater in the WLC than BWL. Thus, results should be generalized with caution. Similar to other literature (Zheng et al., 2022), we found no differences at baseline between neural response to HCF and LCF in our ROIs. Different food stimuli may have had differential effects. We examined participants fasting in the morning; results may have differed in a fed state. For the current study, BOLD response to HCF and LCF and liking, wanting, and palatability of these foods were not correlated. Participants rated foods outside the scanner, which may have accounted for the lack of correspondence between our neuroimaging and psychometric instruments.

Conclusion

Our findings suggest neural responsiveness to visual food cues does not change significantly after BWL. However, we did find improvements in appetite. Future investigations should seek to replicate these data in larger samples, assess interventions that may decrease reward-related responses to food cues, and identify longer-term neural changes that occur with weight loss.

Supplementary Material

Supplemental Data File (doc, pdf, etc.)_1
Supplemental Data File (doc, pdf, etc.)_2

Funding:

This study was funded by an investigator-initiated grant from WW International, Inc. AMC was supported, in part, by the National Institute of Nursing Research of the National Institutes of Health under Award Numbers K23NR017209 and R56NR020466. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Conflict of Interest:

AMC reports grants from WW International, Inc. She has also received grants from Eli Lilly and Co and consulted for Eli Lilly and Co, outside the submitted work. TAW reports serving on advisory boards for Novo Nordisk and WW International, Inc., and receiving grants from Novo Nordisk and Epitomee Medical Ltd. MIC and GDF are current employees and shareholders of WW International, Inc. The other authors have no relevant financial or non-financial interests to disclose.

Footnotes

Clinical Trial Registration: clinicaltrials.gov; https://clinicaltrials.gov/ct2/show/NCT04202133; NCT04202133; date first submitted: December 3, 2019

Contributor Information

Ariana M. Chao, University of Pennsylvania School of Nursing, Department of Biobehavioral Health Sciences, Philadelphia, PA, USA;.

Thomas A. Wadden, Perelman School of Medicine at the University of Pennsylvania, Department of Psychiatry, Philadelphia, PA, USA;.

Wen Cao, Perelman School of Medicine at the University of Pennsylvania, Department of Psychiatry, Philadelphia, PA, USA;.

Yingjie Zhou, University of Pennsylvania School of Nursing, Department of Biobehavioral Health Sciences, Philadelphia, PA, USA;.

Delphina Maldonado, University of Pennsylvania School of Nursing, Department of Biobehavioral Health Sciences, Philadelphia, PA, USA.

Michelle I. Cardel, WW International, Inc., New York, New York, USA; Adjunct Professor, Department of Health Outcomes and Biomedical Informatics, University of Florida College of Medicine, Gainesville, Florida, USA;.

Gary D. Foster, WW International, Inc., New York, New York, USA;.

James Loughead, Perelman School of Medicine at the University of Pennsylvania, Department of Psychiatry, Philadelphia, PA, USA;.

References

  1. Blechert J, Meule A, Busch NA, & Ohla K (2014). Food-pics: An image database for experimental research on eating and appetite. Frontiers Psychology, 5, 617. 10.3389/fpsyg.2014.00617 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Bruce AS, Bruce JM, Ness AR, Lepping RJ, Malley S, Hancock L, Powell J, Patrician TM, Breslin FJ, Martin LE, Donnelly JE, Brooks W,M, & Savage CR (2014). A comparison of functional brain changes associated with surgical versus behavioral weight loss. Obesity, 22, 337–343. 10.1002/oby.20630 [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Cepeda-Benito A, Gleaves DH, Williams TL, & Erath SA (2000). The development and validation of the state and trait food-cravings questionnaires. Behavior Therapy, 31, 151–173. 10.1016/S0005-7894(00)80009-X [DOI] [PubMed] [Google Scholar]
  4. Chao AM, Loughead J, Bakizada ZM, Hopkins CM, Geliebter A, Gur RC, & Wadden TA (2017). Sex/gender differences in neural correlates of food stimuli: A systematic review of functional neuroimaging studies. Obesity Reviews, 18, 687–699. 10.1111/obr.12527 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chao AM, Quigley KM, & Wadden TA (2021). Dietary interventions for obesity: Clinical and mechanistic findings. Journal of Clinical Investigation, 131. 10.1172/JCI140065 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Chao AM, Wadden TA, Walsh OA, Gruber KA, Alamuddin N, Berkowitz RI, & Tronieri JS (2019). Effects of liraglutide and behavioral weight loss on food cravings, eating behaviors, and eating disorder psychopathology. Obesity, 27, 2005–2010. 10.1002/oby.22653 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Deckersbach T, Das SK, Urban LE, Salinardi T, Batra P, Rodman AM, Arulpragasam AR, Dougherty DD, & Roberts SB (2014). Pilot randomized trial demonstrating reversal of obesity-related abnormalities in reward system responsivity to food cues with a behavioral intervention. Nutrition & Diabetes, 4, e129. 10.1038/nutd.2014.26 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Dimitropoulos A, Tkach J, Ho A, & Kennedy J (2012). Greater corticolimbic activation to high-calorie food cues after eating in obese vs. normal-weight adults. Appetite, 58, 303–312. 10.1016/j.appet.2011.10.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Dozois DJA, Dobson KS, & Ahnberg JL (1998). A psychometric evaluation of the Beck Depression Inventory–II. Psychological Assessment, 10, 83–89. 10.1037/1040-3590.10.2.83 [DOI] [Google Scholar]
  10. Foroni F, Pergola G, Argiris G, & Rumiati RI (2013). The FoodCast research image database (FRIDa). Frontiers in Human Neuroscience, 7, 51. 10.3389/fnhum.2013.00051 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Geuter S, Qi G, Welsh RC, Wager TD, & Lindquist MA (2018). Effect size and power in fMRI group analysis. Biorxiv, 295048. 10.1101/295048 [DOI] [Google Scholar]
  12. Hall KD, Farooqi IS, Friedman JM, Klein S, Loos RJF, Mangelsdorf DJ, O’Rahilly S, Ravussin E, Redman LM, Ryan DH, Speakman JR, & Tobias DK (2022). The energy balance model of obesity: Beyond calories in, calories out. American Journal of Clinical Nutrition, 115, 1243–1254. 10.1093/ajcn/nqac031 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Jansen A (1998). A learning model of binge eating: Cue reactivity and cue exposure. Behaviour Research and Therapy, 36, 257–272. 10.1016/s0005-7967(98)00055-2 [DOI] [PubMed] [Google Scholar]
  14. Kahathuduwa CN, Binks M, Martin CK, & Dawson JA (2017). Extended calorie restriction suppresses overall and specific food cravings: A systematic review and a meta‐analysis. Obesity Reviews, 18, 1122–1135. 10.1111/obr.12566 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Kanoski SE, & Boutelle KN (2022). Food cue reactivity: Neurobiological and behavioral underpinnings. Reviews in Endocrine and Metabolic Disorders, 23, 683–696. 10.1007/s11154-022-09724-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Martens MJI, Born JM, Lemmens SGT, Karhunen L, Heinecke A, Goebel R, Adam TC, & Westerterp-Plantenga MS (2013). Increased sensitivity to food cues in the fasted state and decreased inhibitory control in the satiated state in the overweight. American Journal of Clinical Nutrition, 97, 471–479. 10.3945/ajcn.112.044024 [DOI] [PubMed] [Google Scholar]
  17. Martin LE, Holsen LM, Chambers RJ, Bruce AS, Brooks WM, Zarcone JR, Butler MG, & Savage CR (2010). Neural mechanisms associated with food motivation in obese and healthy weight adults. Obesity, 18, 254–260. 10.1038/oby.2009.220 [DOI] [PubMed] [Google Scholar]
  18. Mason AE, Vainik U, Acree M, Tomiyama AJ, Dagher A, Epel ES, & Hecht FM (2017). Improving assessment of the spectrum of reward-related eating: The RED-13. Frontiers in Psychology, 8, 795. 10.3389/fpsyg.2017.00795 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Meyer MD, Risbrough VB, Liang J, & Boutelle KN (2015). Pavlovian conditioning to hedonic food cues in overweight and lean individuals. Appetite, 87, 56–61. https:// 10.1016/j.appet.2014.12.002 [DOI] [PubMed] [Google Scholar]
  20. Murdaugh DL, Cox JE, Cook EW III, & Weller RE (2012). fMRI reactivity to high-calorie food pictures predicts short- and long-term outcome in a weight-loss program. NeuroImage, 59, 2709–2721. 10.1016/j.neuroimage.2011.10.071 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. O’Neil PM, Theim KR, Boeka A, Johnson G, & Miller-Kovach K (2012). Changes in weight control behaviors and hedonic hunger during a 12-week commercial weight loss program. Eating Behaviors, 13, 354–360. 10.1016/j.eatbeh.2012.06.002 [DOI] [PubMed] [Google Scholar]
  22. Oustric P, Gibbons C, Beaulieu K, Blundell J, & Finlayson G (2018). Changes in food reward during weight management interventions–a systematic review. Obesity Reviews, 19, 1642–1658. 10.1111/obr.12754 [DOI] [PubMed] [Google Scholar]
  23. Pursey KM, Stanwell P, Callister RJ, Brain K, Collins CE, & Burrows TL (2014). Neural responses to visual food cues according to weight status: A systematic review of functional magnetic resonance imaging studies. Frontiers in Nutrition, 1, 7. 10.3389/fnut.2014.00007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Rolls ET, Huang C-C, Lin C-P, Feng J, & Joliot M (2020). Automated anatomical labelling atlas 3. NeuroImage, 206, 116189. 10.1016/j.neuroimage.2019.116189 [DOI] [PubMed] [Google Scholar]
  25. Rothemund Y, Preuschhof C, Bohner G, Bauknecht H-C, Klingebiel R, Flor H, & Klapp BF (2007). Differential activation of the dorsal striatum by high-calorie visual food stimuli in obese individuals. NeuroImage, 37, 410–421. 10.1016/j.neuroimage.2007.05.008 [DOI] [PubMed] [Google Scholar]
  26. Ryan DH, & Yockey SR (2017). Weight loss and improvement in comorbidity: Differences at 5%, 10%, 15%, and over. Current Obesity Reports, 6, 187–194. 10.1007/s13679-017-0262-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Salem V, Demetriou L, Behary P, Alexiadou K, Scholtz S, Tharakan G, Miras AD, Purkayastha S, Ahmed AR, Bloom SR, Wall MB, Dhillo WS, & Tan TM-M (2021). Weight loss by low-calorie diet versus gastric bypass surgery in people with diabetes results in divergent brain activation patterns: A functional MRI study. Diabetes Care, 44, 1842–1851. 10.2337/dc20-2641 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Sankoh AJ, Huque MF, & Dubey SD (1997). Some comments on frequently used multiple endpoint adjustment methods in clinical trials. Statistics in Medicine, 16, 2529–2542. 10.1002/(sici)1097-0258(19971130)16:22<2529::aid-sim692>3.0.co;2-j [DOI] [PubMed] [Google Scholar]
  29. Scharmüller W, Übel S, Ebner F, & Schienle A (2012). Appetite regulation during food cue exposure: A comparison of normal-weight and obese women. Neuroscience Letters, 518, 106–110. 10.1016/j.neulet.2012.04.063 [DOI] [PubMed] [Google Scholar]
  30. Simmons JP, Nelson LD, & Simonsohn U (2011). False-positive psychology: Undisclosed flexibility in data collection and analysis allows presenting anything as significant. Psychological Science, 22, 1359–1366. 10.1177/0956797611417632 [DOI] [PubMed] [Google Scholar]
  31. Spence C, Okajima K, Cheok AD, Petit O, & Michel C (2016). Eating with our eyes: From visual hunger to digital satiation. Brain and Cognition, 110, 53–63. 10.1016/j.bandc.2015.08.006 [DOI] [PubMed] [Google Scholar]
  32. Stice E, Yokum S, Gau J, Veling H, Lawrence N, & Kemps E (2022). Efficacy of a food response and attention training treatment for obesity: A randomized placebo controlled trial. Behaviour Research and Therapy, 158, 104183. 10.1016/j.brat.2022.104183 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Stice E, Yokum S, Veling H, Kemps E, & Lawrence NS (2017). Pilot test of a novel food response and attention training treatment for obesity: Brain imaging data suggest actions shape valuation. Behaviour Research and Therapy, 94, 60–70. 10.1016/j.brat.2017.04.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Stoeckel LE, Weller RE, Cook EW III, Twieg DB, Knowlton RC, & Cox JE (2008). Widespread reward-system activation in obese women in response to pictures of high-calorie foods. NeuroImage, 41, 636–647. 10.1016/j.neuroimage.2008.02.031 [DOI] [PubMed] [Google Scholar]
  35. Strack F, & Deutsch R (2004). Reflective and impulsive determinants of social behavior. Personality and Social Psychology Review, 8, 220–247. 10.1207/s15327957pspr0803_1 [DOI] [PubMed] [Google Scholar]
  36. Stunkard AJ, & Messick S (1985). The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29, 71–83. 10.1016/0022-3999(85)90010-8 [DOI] [PubMed] [Google Scholar]
  37. van den Akker K, Schyns G, & Jansen A (2018). Learned overeating: Applying principles of Pavlovian conditioning to explain and treat overeating. Current Addiction Reports, 5, 223–231. 10.1007/s40429-018-0207-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Wadden TA, & Foster GD (2006). Weight and lifestyle inventory (WALI). Obesity, 14, 99S–118S. 10.1038/oby.2006.289 [DOI] [PubMed] [Google Scholar]
  39. Yokum S, Ng J, & Stice E (2011). Attentional bias to food images associated with elevated weight and future weight gain: An fMRI study. Obesity, 19, 1775–1783. 10.1038/oby.2011.168 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Zheng L, Miao M, & Gan Y (2022). A systematic and meta-analytic review on the neural correlates of viewing high-and low-calorie foods among normal-weight adults. Neuroscience & Biobehavioral Reviews, 138, 104721. 10.1016/j.neubiorev.2022.104721 [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supplemental Data File (doc, pdf, etc.)_1
NIHMS1938325-supplement-Supplemental_Data_File__doc__pdf__etc___1.docx (15.2KB, docx)
Supplemental Data File (doc, pdf, etc.)_2
NIHMS1938325-supplement-Supplemental_Data_File__doc__pdf__etc___2.docx (15KB, docx)

RESOURCES