HHS Public Access

Author manuscript
Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
Author Manuscript

Published in final edited form as:

Obesity (Silver Spring). 2016 April ; 24(4): 829–836. doi:10.1002/oby.21424.

Brain Imaging Demonstrates a Reduced Neural Impact of Eating

in Obesity
Nancy Puzziferri, MD, MSa,b,c, Jeffrey M. Zigman, MD, PhDd,f, Binu P. Thomas, PhDd,e, Perry
Mihalakos, BSd, Ryan Gallagher, BAa, Michael Lutter, MD, PhDg, Thomas Carmody, PhDc,d,
Hanzhang Lu, PhDd,e, and Carol A. Tamminga, MDd
aDepartment of Surgery, University of Texas Southwestern Medical Center, Dallas, TX
Author Manuscript

bDepartment of Surgery, Veterans Administration North Texas Health Care System, Dallas, TX
cDepartment of Clinical Sciences, University of Texas Southwestern Medical Center, Dallas, TX
dDepartment of Psychiatry, University of Texas Southwestern Medical Center, Dallas, TX
eAdvanced Imaging Research Center and the Department of Radiology, University of Texas
Southwestern Medical Center, Dallas, TX
fDepartment of Internal Medicine, University of Texas Southwestern Medical Center, Dallas, TX
gDepartment of Psychiatry, University of Iowa, Carver College of Medicine, Iowa City, IA

Abstract
Author Manuscript

Objective—We investigated functional brain response differences to food in women with either
BMI <25kg/m2 (lean) or >35kg/m2 (severe obesity).

Methods—Thirty women 18-65 years old from academic medical centers participated. Baseline
brain perfusion was measured with arterial spin labeling. Brain activity was measured via blood-
oxygen-level-dependent functional magnetic resonance imaging (fMRI) in response to food cues,
and appeal to cues rated. Subjective hunger/fullness was reported pre- and post-imaging. After a
standard meal, measures were repeated.

Results—When fasting, brain perfusion did not differ significantly between groups; and both
groups significantly increased activity in the neo- and limbic cortices and midbrain compared to
baseline (p<0.05, family-wise-error whole-brain corrected). Once fed, the lean group showed
significantly decreased activation in these areas, especially the limbic cortex, while the group with
Author Manuscript

severe obesity showed no such decreases (p<0.05, family-wise-error whole-brain corrected). After
eating, appeal ratings of food decreased only in lean women. Within groups, hunger decreased
(p<0.001) and fullness increased (p<0.001) fasted to fed.

Conclusion—While fasting, brain response to food cues in women did not differ significantly
despite BMI. After eating, brain activity quickly diminished in lean women but remained elevated
in women with severe obesity. These brain activation findings confirm previous studies.

Contact Information: Nancy.puzziferri@utsouthwestern.edu, Nancy Puzziferri, MD, MS, Department of Surgery, University of Texas
Southwestern Medical Center, 5323 Harry Hines Blvd.; Mail Code 9156, Dallas, TX 75390.
All authors claim no potential conflicts of interest.
 Puzziferri et al. Page 2

Keywords
Author Manuscript

women; brain; food; imaging; hunger

Introduction
Even if not hungry, seeing food or thinking about food can stimulate eating (see Smeets,
Erkner and de Graaf(1), and Berthoud(2) for recent and broader discussion). The
anticipatory or conditioned responses to food cues, rather than hunger, affects us each
differently depending on gender, energy state, menstrual status, and previous associations to
those foods or food cues, to name a few(3-5). Those with the insight and/or inhibitory
control of visual cue urges succeed in maintaining a healthy weight(6). In contrast,
individuals with obesity are highly vulnerable to external food-related cues(7, 8) and eat
Author Manuscript

differently than lean individuals in identical environments (9, 10).

Much current research on eating involves characterizing the role of the brain. Central
nervous system mechanisms of homeostatic (energy-based) circuits integrate with hedonic
(reward/motivation-based) circuits to influence eating(11). Given that homeostatic needs are
easily met, differences in eating for pleasure are likely central to overeating. Physiological
feedback to our brain upon eating — in the form of changed levels of satiety-related gut
hormones, adipokines and/or vagal stimuli — mediate and regulate our intake and activation
of key eating-related brain centers (12-14). In turn, these changes to brain region activation
further influence behaviors and processes involving eating. If food eaten is pleasurable, its
intake is reinforced. Interestingly, while taste-based food choices compel women of all body
mass indices (BMI) to eat, when full, lean women will either stop eating or just sample a
food they crave rather than eating a large volume as do women with obesity(15).
Author Manuscript

Functional magnetic resonance imaging (fMRI) has been used in numerous studies to assess
differences in regional brain response to visual food cues between individuals with and
without obesity. A common outcome has been differential activation of brain regions
mediating reward behaviors. Although space limitations prevent mention of all these
works, some key findings from this literature follow: In woman with obesity as compared to
those who are lean, the dorsal striatum, which mediates decision-making related to
producing reward attainment actions, shows greater activation in the fed state. (16), reward-
system-associated regions—the ventral tegmental area, nucleus accumbens, amygdala,
orbital frontal cortex (7) and left anterior cingulate(5) — exhibit greater brain activation in
the fasted state, and when anticipating ingesting (rather than viewing) a basic sweet taste,
brain response is down-regulated in a fed state(4). In both fasted and fed states, women and
Author Manuscript

men with obesity, exhibit significantly greater activity in the medial prefrontal cortex when
viewing food cues compared to lean individuals (8), which is important because the
prefrontal cortex mediates motivated behaviors aimed at obtaining a reward. Also,
individuals with proneness to obesity do not attenuate their neural response after a meal as
do women and men with obesity resistance(17). Identifying brain circuit activation pattern
differences between individuals with or without obesity may generate functional biomarkers
that facilitate the development of obesity treatments. Similarly, identifying brain circuit

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 3

activation differences predicted to occur with bariatric surgery-induced weight loss may
Author Manuscript

bring opportunities for developing efficacious non-surgical treatments.

We examined brain responses to visual food cues in women with severe obesity or leanness
before and after a standard meal. We contrasted alterations in functional brain activation
within regions that regulate eating behaviors—both homeostatic and hedonic. We
hypothesized that regions important in mediating reward behaviors (ventral tegmental area,
nucleus accumbens, hippocampus and prefrontal cortex) would be differentially activated by
viewing food cues in the fasted or fed states, contrasting groups with severe obesity or
leanness.

Methods
Participants
Author Manuscript

Fifteen women with severe obesity and 15 age-matched lean women (18-65 years old) were
recruited through two academic medical centers. The University of Texas Southwestern
Medical Center and the Veterans Administration North Texas Health Care System
Institutional Review Boards approved the protocol and all participants gave informed
consent. Participants with severe obesity were BMI 35-50 kg/m2 and pre-bariatric surgery.
Lean controls were BMI 18.5-24.9 kg/m2. Exclusion criteria included untreated Axis I
psychiatric diagnoses, previous serious head injury, left-handedness, contraindications to
MRI, prior bariatric surgery, or untreated severe medical illness.

Study Design
Demographic characteristics were collected (Table 1). Psychiatric diagnoses were
determined using the Structured Clinical Interview for the Diagnostic and Statistical
Author Manuscript

Manual-4 (18). Current depressive symptoms were determined using the 16-item Quick
Inventory of Depressive Symptomatology (QIDS-SR16) (19).

Participants arrived at 9 AM in a fasted state (no food after midnight). Subjective hunger or
fullness (visual analog scale) was rated prior to imaging using a scale from -50 (least
hunger/fullness ever experienced) to +50 (greatest hunger/fullness ever experienced).
Participants were then positioned in the MRI scanner with their heads comfortably
immobilized. Baseline brain perfusion, measured by Arterial Spin Labeling (ASL), was
acquired prior to fMRI scanning. Participants' brain activities were measured with fMRI
blood-oxygen-level-dependent (BOLD) response to seeing food cues interleaved with
directional arrows. An appeal rating (scale of 1-3, not appealing to very appealing) for each
food shown was recorded. After this imaging session, subjective hunger or fullness was
Author Manuscript

rated again. During the next hour, a standard meal of 337 kcals (52% carbohydrate, 30% fat
and 18% protein) was served. Calories consumed were quantified. The meal consisted of:
lean beef or chicken, with potato or rice, and green beans; water-packed canned peaches;
and iced tea +/-splenda or water. After eating, participants again rated hunger or fullness.
Then participants were re-scanned using an fMRI paradigm identical to the first session
except with different food cues, avoiding familiarity. Afterward, subjective hunger or
fullness was rated again.

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 4

fMRI Food Task

Author Manuscript

Visual food cues (pictures) were similarly sized and presented (Supporting Figure S1). Five
fMRI BOLD runs per scan session were acquired in an event-related design. In each run, 40
food pictures were presented for 3 seconds each in a pseudo-randomized sequence over 3.5
minutes. Null “arrow” events (each 1.5 seconds) were randomly interleaved between the
food pictures. Ten additional “arrow” events were placed at the run beginning and end to
establish a baseline BOLD signal. The pictures included 10 high-calorie savory foods, 10
high-calorie sweet foods, and 20 low-calorie foods. While viewing food pictures,
participants were asked to rank foods by appeal (“not appealing”, “appealing” or “very
appealing”) using button press. Details on MRI procedures, and parameters are reported in
the Supporting Methods.

MRI Data Analysis

Author Manuscript

Brain images were processed using Statistical Parametric Mapping 5 (20) and fMRI Brain
Software Library (FSL; Oxford University, UK) (21). fMRI images were realigned to correct
for motion artifacts; runs with motion exceeding one voxel size were excluded. The
MPRAGE image was then co-registered to the realigned fMRI images. The Brain Extraction
Tool (in FSL) was used to extract the brain from the skull and subcutaneous fat in MPRAGE
images, because excess subcutaneous fat was found to generate distortions in the normalized
images. Image normalization was done to transform the skull-stripped MPRAGE image into
Montreal Neurological Institute template space, and the transformation parameters were
used to normalize the fMRI images. Normalized fMRI images were re-sampled into 2 mm
cubic voxels and smoothed using an 8mm full-width half-maximum Gaussian kernel to
minimize inter- participant anatomic variability. Time and dispersion derivatives of the
hemodynamic response function were included to obtain a better model of the data. A 128
Author Manuscript

seconds high-pass filter removed low-frequency noise and slow signal drifts.

Cerebral blood flow (CBF) data were acquired in the “before-meal” session only. CBF
procedures, parameters and analysis are reported in the Supporting Methods.

fMRI Statistical Analysis

fMRI data were analyzed using a general linear model in which stimulus onsets were
modeled as events and specified as regressors. These onsets were convolved with the
hemodynamic response function to account for lag between event onset and the expected
BOLD signal response. To account for variance from head movement, realignment
parameters were included as regressors. Flexible factorial design was used for two-group
(BMI >35kg/m2, BMI <25kg/m2) and two-condition (fasted, fed) analysis. Predicted
Author Manuscript

activations were considered significant at p<0.05 after correcting for family-wise-error

(FWE) across voxels.

Descriptive Statistical Analysis

Continuous demographic and clinical characteristics data were described by means and
standard deviations and compared by t-test. Categorical variables were described by number
and proportion of participants, and analyzed by Chi-square or Fisher's exact test. A repeated-
measures analysis of variance was used to assess subjective hunger and fullness with effects

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 5

for within-group (fasted versus fed), between-group (lean versus obese) and a within-group
Author Manuscript

by between-group interaction. Appeal ratings of food cues were collapsed to a binary

outcome (“very appealing/appealing” versus “not appealing”) before analysis. The outcome
was the proportion of very appealing/appealing responses given for the 200 rated pictures.
The arc sine transformation was applied to the square root of the proportions to improve the
normality of the data, and then the repeated measures analysis of variance was applied. SAS
version 9.3 (SAS Institute Inc., Cary, North Carolina) was used for all descriptive statistical
analyses.

Results
Group Characteristics
Thirty women entered and completed the protocol. Of the demographic/clinical
characteristics and current/lifetime Axis I diagnoses, only weight and BMI differed
Author Manuscript

significantly between groups (p=0.001, Table 1). No participants met criteria for current
depression (QIDS-SR16, all <11).

Eating Behavior
There were no significant differences between groups in standard meal calories consumed
(BMI <25kg/m2: 301 kcals, (SD 51); BMI >35kg/m2: 302 kcals, (SD 66); p=0.95; table 2).
Subjective hunger was significantly greater in the fasted versus fed states for both groups
(p<0.001; see absolute hunger values table 2). The subjective hunger group effect was
significant (p=0.02), indicating that participants with severe obesity, whether fasted or fed,
showed quantitatively less hunger than lean controls.

Subjective fullness was significantly lower in the fasted versus fed states for both groups
Author Manuscript

(p<0.001; see absolute fullness values table 2). Subjective fullness did not differ
significantly between groups, at any time point, either before or after eating their standard
meal (p=0.40, group effect). The groups did not differ significantly in their fullness ratings
in response to the meal (p=0.45, group × fasted-fed effect), suggesting that each group was
satiated.

Participants with severe obesity and lean participants differed in how they rated the appeal of
food cues between the fasted and fed states. The lean group showed a drop of 15% (95%
Confidence Interval: 5.0 to 24.1) in their rating of appeal as they moved from fasted to fed
(65% versus 50% rated very appealing/appealing). In contrast, the group with severe obesity
reported a drop of 4% (95% Confidence Interval: -3.1 to 11.5) (68% versus 64%), which
indicates sustained appeal of food cues after eating. The fasted/fed-by-group interaction was
Author Manuscript

not significant (p=0.070).

Lean Group fMRI outcomes (whole brain, Supporting Figure S2)

In fasted lean participants (Figure 1a), diverse brain regions were activated by food cues.
Activation of visual/striate cortices was consistent with visual stimuli and the visual cortical
activations extended to the cerebellum and parietal cortex. Areas within the prefrontal cortex
(including the anterior cingulate, medial prefrontal cortex and dorsolateral prefrontal cortex)

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 6

were activated, as were regions of the basal ganglia (particularly the caudate nucleus). The
Author Manuscript

medial temporal cortex, including the hippocampus and midbrain, in the region associated
with the ventral tegmental area, was also activated. After eating, activation diminished in the
anterior cingulate, medial prefrontal cortex, dorsolateral prefrontal cortex, caudate nucleus,
and midbrain (Figure 1b). The lean [fasted minus fed] analysis (Figure 1c; Table 3) showed
significant activation reductions after eating (p<0.05, FWE corrected) in the prefrontal
cortex (anterior cingulate, medial prefrontal cortex and dorsolateral prefrontal cortex), basal
ganglia/caudate nucleus, medial temporal cortex and midbrain regions. Activation in the
visual cortex and cerebellum were also significantly diminished (p<0.05, FWE corrected),
while activations in the ventral cortex were sustained.

Group with severe obesity fMRI outcomes (whole brain, Supporting Figure S3)
Fasted participants with severe obesity showed widespread activations by food cues (p<0.05,
Author Manuscript

FWE corrected) as did fasted lean participants (Figure 2a). The severe obesity-fasted group
showed brisk activations in the ventral cortex, with extension into the parietal cortex and
cerebellum. Prefrontal cortex activations were prominent, including the anterior cingulate,
medial prefrontal cortex and dorsolateral prefrontal cortex, extending to the inferior
prefrontal cortex and insula. Also prominent were medial temporal cortex activations,
including the hippocampus, and brainstem regions, including the ventral tegmental area. In
the fed state, many areas remained significantly activated (p<0.05, FWE corrected), unlike
the lean group (Figure 2b). The prefrontal, medial temporal, parietal and ventral cortical
regions all remained active, making the ‘fed’ activations in the obese brain resemble the
‘fasted’ state. The group with severe obesity [fasted minus fed] analysis showed significantly
reduced activations [p<0.05, FWE corrected] only in the insula, inferior prefrontal cortex,
and inferior midbrain; with no significant activation diminution within the prefrontal, medial
Author Manuscript

temporal, parietal or ventral cortices (Figure 2c; Table 3).

The lean group deactivated brain regions when fed compared to fasted, in contrast to the
group with severe obesity (Figure 3; Table 4). Regions in the prefrontal cortex, particularly
the anterior cingulate, dorsolateral prefrontal cortex and posterior cingulate cortex,
significantly changed going from fasted to fed state in the lean, but not in the group with
severe obesity. Additionally, in the severe obesity-fasted group, the hypothalamus activation
did not differ statistically (p=0.07) to the lean-fasted group. Both groups showed no
meaningful within-group differentiated hypothalamus activation from the fasted to fed
states.

Baseline Perfusion
To confirm that brain activation findings were not the result of differences between the
Author Manuscript

groups in their fasted (baseline) state, we analyzed baseline CBF data. A mask was obtained
from the regions that showed differences between the groups for the [fasted minus fed] state
(color voxels, Figure 3) and baseline CBF was obtained for each participant from this mask.
Baseline CBFs during the fasted state were: lean group = 70.5±9.8 ml/100g/min, group with
severe obesity = 69.7±9.8 ml/100g/min, demonstrating that the groups were not different in
the fasted state (p=0.83).

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 7

Discussion
Author Manuscript

Significant activation of brain regions known to mediate hedonic behaviors occurred in

fasted participants, both lean and with severe obesity, in response to visual food cues: the
prefrontal cortex (especially the anterior cingulate), basal ganglia (especially the caudate
nucleus) and medial temporal cortex, as well as sensory perceptual areas and the parietal
cortex. The hypothalamus, viewed as a center for body weight homeostasis, showed no
significant change between groups or across feeding states. After eating, the group with
severe obesity showed sustained “hungry” activation despite no statistical difference in
subjective reports of satiation to the lean group who had diminished “non-hungry”
activation.

These brain activation findings confirm previously shown characteristics for lean or groups
with obesity studied: without comparators (22, 23), in a different demographic (adolescents,
Author Manuscript

males, BMI 25 - 35kg/m2) (24, 25) or in a single state (fasted or fed, but not both) (7, 26,
27). Few study designs show the clear fed-versus-fasted state differences simultaneously
between and within groups. Similar to findings by Cornier et al. in obesity-resistant versus
obesity-prone participants(17), this a priori-designed study demonstrates an ongoing brain
response to visual food cues in severe obesity—especially in the neo- and limbic cortices,
and midbrain—after eating a satisfying meal, distinct from lean controls. These findings
suggest participants with severe obesity fail to engage a brain-wide eating-related process
associated with satiation. It remains to be clarified whether this brain response to food cues
is a trait or state response induced by obesity. We are following the participants with severe
obesity after bariatric surgery to test this outcome.

In this study, key aspects of brain response to eating are reflected by objective behavioral
Author Manuscript

response (appeal) and subjective motivation (hunger, fullness)—a translation of brain

findings to eating behavior. The brain response when fasted is congruent with behavioral and
motivation experience in both groups; brain activation is high, food pictures are appealing
and participants are at their hungriest. The brain response when fed remains congruent with
behavior and motivation experience in lean participants but is divergent in participants with
severe obesity. Unlike leans, despite diminished hunger/peak fullness, the group with severe
obesity-fed maintain high brain activation to and appeal ratings of food. Participants with
severe obesity reported significantly less absolute hunger than lean participants whether
fasted or fed, but group differences between pre- and post-meal hunger/fullness were
identical. Taken together, the severe obesity-fed state depicts an ‘uncoupling’ of effective
brain response to eating and degree of hunger/fullness. These findings may explain why
participants with severe obesity report an underlying drive to eat continually despite not
Author Manuscript

feeling hungry (28), or eat differently, consuming more than lean controls (29, 30).

The lean group decreased regional brain activation to visual food cues after eating, which
suggests that the brain withdraws attention from food-related stimuli when satiated. The lack
of diminished brain activation in participants with severe obesity once fed, lends support to
an appetitive-conditioning psychopathology model (31). Despite satiation, a high brain
activation pattern identical to fasting in participants with severe obesity reflects a neuronal
pressure around eating which is absent in lean participants. These brain differences may

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 8

assist clinical weight loss treatment by utilizing brain-based therapies or validating patients'
Author Manuscript

feelings of ‘something being wrong’ with regard to eating. Similar to the variability of
genetically driven weight gain and loss (32, 33), clinicians can acknowledge a disparity in
response to eating and do so without undermining treatment success.

This study has limitations. The sample size, while sufficient for fMRI data stability, is not
large per usual clinical study size. Our study meal (337 kcals) was set at the recommended
lunch caloric level for a 1200 kcal diet(34) and is less than the average 626 kcals meal
consumed by U.S. women (35); brain activation differences between groups may change
after a larger meal. While our focus on women is essential as the experience of obesity in
women differs in many respects from that in men—including heightened perceived physical
impairments (36), weight-related social stigma and discrimination (37), risk of depression
(38) and distinct risk factors (39)—our results may not be generalizable to men. Though
age-matched, with equal numbers of postmenopausal women per group, we could not
Author Manuscript

control for menstrual cycles, oral contraceptives or hormone replacement therapy. Sex
hormones are known to modulate appetite(40) and may affect these results. Finally, neither
molecular markers nor gut hormones known to affect brain function were analyzed; such
correlations are forthcoming. This study's strengths include its use of a longitudinal within-
and between-group design of absolute and change in brain activation, coupled with behavior
measures, and post-bariatric surgery follow-up in the obese group. This study contributes to
the high priority of developing effective obesity targets and treatment strategies.

Conclusion
This study's comparison between lean and participants with severe obesity's brain responses
to food cues demonstrates how brain activation patterns vary in a fasted versus a fed state. In
Author Manuscript

response to food cues, the brains of both lean and participants with severe obesity when
fasted show fMRI BOLD activation patterns widely distributed in the neo- and limbic
cortices. Once lean participants are fed, their brain BOLD activations to food pictures are
broadly reduced. In contrast, after participants with severe obesity are fed, they fail to show
remarkably altered patterns of brain activation to food cues. In particular, after eating,
participants with severe obesity maintain activation in the midbrain, one of the most potent
reward centers. Thus, once satiated after eating, participants with severe obesity continue to
perceive food as appealing and their brains continue to be activated by visual food cues as
though they were hungry. Future experiments will determine whether the observed
participants with severe obesity's brain activation patterns will change following bariatric
surgery, correlate with changes in body weight-related gut hormones, or differ between high-
versus low-caloric density visual food cues.
Author Manuscript

Supplementary Material
Refer to Web version on PubMed Central for supplementary material.

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 9

Acknowledgments
Author Manuscript

We would like to acknowledge the editorial support of Jon Kilner, MS, MA (Pittsburgh, PA), the dietary-recall
assessments by Rosemary Son, PA-C, RD (Dallas, TX), and the preliminary fMRI BOLD analyses by Yan Fang,
PhD (Dallas, TX).

Sources of financial funding: 1. Department of Surgery, University of Texas Southwestern Medical Center, Dallas,
TX 75390 provided funding for all aspects of the research.

2. NIH/NCATS Grant Number UL1TR000451 provided pilot research funding.

3. Advanced Imaging Research Center, University of Texas Southwestern Medical Center, Dallas, TX 75390
provided pilot research funding.

References
1. Smeets PA, Erkner A, de Graaf C. Cephalic phase responses and appetite. Nutr Rev. 2010; 68(11):
643–55. [PubMed: 20961295]
Author Manuscript

2. Berthoud HR. The neurobiology of food intake in an obesogenic environment. Proc Nutr Soc. 2012;
71(4):478–87. [PubMed: 22800810]
3. Legget KT, Cornier MA, Rojas DC, Lawful B, Tregellas JR. Harnessing the power of disgust: a
randomized trial to reduce high-calorie food appeal through implicit priming. Am J Clin Nutr. 2015;
102(2):249–55. [PubMed: 26109580]
4. Cornier MA, Shott ME, Thomas EA, Bechtell JL, Bessesen DH, Tregellas JR, et al. The effects of
energy balance, obesity-proneness and sex on the neuronal response to sweet taste. Behav Brain
Res. 2015; 278:446–52. [PubMed: 25447301]
5. Martens MJ, Born JM, Lemmens SG, Karhunen L, Heinecke A, Goebel R, et al. Increased
sensitivity to food cues in the fasted state and decreased inhibitory control in the satiated state in the
overweight. Am J Clin Nutr. 2013; 97(3):471–9. [PubMed: 23364016]
6. Sweet LH, Hassenstab JJ, McCaffery JM, Raynor HA, Bond DS, Demos KE, et al. Brain response
to food stimulation in obese, normal weight, and successful weight loss maintainers. Obesity (Silver
Spring). 2012; 20(11):2220–5. [PubMed: 22569002]
Author Manuscript

7. Stoeckel LE, Weller RE, Cook EW 3rd, Twieg DB, Knowlton RC, Cox JE. Widespread reward-
system activation in obese women in response to pictures of high-calorie foods. Neuroimage. 2008;
41(2):636–47. [PubMed: 18413289]
8. Martin LE, Holsen LM, Chambers RJ, Bruce AS, Brooks WM, Zarcone JR, et al. Neural
mechanisms associated with food motivation in obese and healthy weight adults. Obesity (Silver
Spring). 2010; 18(2):254–60. [PubMed: 19629052]
9. Schachter S. Obesity and eating. Internal and external cues differentially affect the eating behavior
of obese and normal subjects. Science. 1968; 161(3843):751–6. [PubMed: 5663800]
10. Lawrence NS, Hinton EC, Parkinson JA, Lawrence AD. Nucleus accumbens response to food cues
predicts subsequent snack consumption in women and increased body mass index in those with
reduced self-control. Neuroimage. 2012; 63(1):415–22. [PubMed: 22776461]
11. Williams KW, Elmquist JK. From neuroanatomy to behavior: central integration of peripheral
signals regulating feeding behavior. Nat Neurosci. 2012; 15(10):1350–5. [PubMed: 23007190]
12. Batterham RL, ffytche DH, Rosenthal JM, Zelaya FO, Barker GJ, Withers DJ, et al. PYY
modulation of cortical and hypothalamic brain areas predicts feeding behaviour in humans. Nature.
Author Manuscript

2007; 450(7166):106–9. [PubMed: 17934448]

13. Farooqi IS, Bullmore E, Keogh J, Gillard J, O'Rahilly S, Fletcher PC. Leptin regulates striatal
regions and human eating behavior. Science. 2007; 317(5843):1355. [PubMed: 17690262]
14. Schwartz GJ. The role of gastrointestinal vagal afferents in the control of food intake: current
prospects. Nutrition. 2000; 16(10):866–73. [PubMed: 11054591]
15. Dressler H, Smith C. Food choice, eating behavior, and food liking differs between lean/normal
and overweight/obese, low-income women. Appetite. 2013; 65:145–52. [PubMed: 23428940]

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 10

16. Rothemund Y, Preuschhof C, Bohner G, Bauknecht HC, Klingebiel R, Flor H, et al. Differential
activation of the dorsal striatum by high-calorie visual food stimuli in obese individuals.
Author Manuscript

Neuroimage. 2007; 37(2):410–21. [PubMed: 17566768]

17. Cornier MA, McFadden KL, Thomas EA, Bechtell JL, Eichman LS, Bessesen DH, et al.
Differences in the neuronal response to food in obesity-resistant as compared to obesity-prone
individuals. Physiol Behav. 2013; 110-111:122–8. [PubMed: 23313402]
18. Spitzer RL, Williams JB, Gibbon M, First MB. The Structured Clinical Interview for DSM-III-R
(SCID). I: History, rationale, and description. Arch Gen Psychiatry. 1992; 49(8):624–9. [PubMed:
1637252]
19. Rush AJ, Trivedi MH, Ibrahim HM, Carmody TJ, Arnow B, Klein DN, et al. The 16-Item Quick
Inventory of Depressive Symptomatology (QIDS), clinician rating (QIDS-C), and self-report
(QIDS-SR): a psychometric evaluation in patients with chronic major depression. Biol Psychiatry.
2003; 54(5):573–83. [PubMed: 12946886]
20. Wise RA. Drug-activation of brain reward pathways. Drug Alcohol Depend. 1998; 51(1-2):13–22.
[PubMed: 9716927]
21.
Author Manuscript

http://www.fmrib.ox.ac.uk/fsl/
22. Ochner CN, Kwok Y, Conceicao E, Pantazatos SP, Puma LM, Carnell S, et al. Selective reduction
in neural responses to high calorie foods following gastric bypass surgery. Ann Surg. 2011; 253(3):
502–7. [PubMed: 21169809]
23. Frank S, Laharnar N, Kullmann S, Veit R, Canova C, Hegner YL, et al. Processing of food
pictures: influence of hunger, gender and calorie content. Brain Res. 2010; 1350:159–66.
[PubMed: 20423700]
24. Dimitropoulos A, Tkach J, Ho A, Kennedy J. Greater corticolimbic activation to high-calorie food
cues after eating in obese vs. normal-weight adults. Appetite. 2012; 58(1):303–12. [PubMed:
22063094]
25. Bruce AS, Holsen LM, Chambers RJ, Martin LE, Brooks WM, Zarcone JR, et al. Obese children
show hyperactivation to food pictures in brain networks linked to motivation, reward and cognitive
control. Int J Obes (Lond). 2010; 34(10):1494–500. [PubMed: 20440296]
26. Stoeckel LE, Kim J, Weller RE, Cox JE, Cook EW 3rd, Horwitz B. Effective connectivity of a
Author Manuscript

reward network in obese women. Brain Res Bull. 2009; 79(6):388–95. [PubMed: 19467298]
27. Cornier MA, Von Kaenel SS, Bessesen DH, Tregellas JR. Effects of overfeeding on the neuronal
response to visual food cues. Am J Clin Nutr. 2007; 86(4):965–71. [PubMed: 17921372]
28. Zunker C, Karr T, Saunders R, Mitchell JE. Eating behaviors post-bariatric surgery: a qualitative
study of grazing. Obes Surg. 2012; 22(8):1225–31. [PubMed: 22527594]
29. Ernst B, Thurnheer M, Wilms B, Schultes B. Differential changes in dietary habits after gastric
bypass versus gastric banding operations. Obes Surg. 2009; 19(3):274–80. [PubMed: 19034589]
30. Spiegel TA, Shrager EE, Stellar E. Responses of lean and obese subjects to preloads, deprivation,
and palatability. Appetite. 1989; 13(1):45–69. [PubMed: 2782866]
31. Martin-Soelch C, Linthicum J, Ernst M. Appetitive conditioning: neural bases and implications for
psychopathology. Neurosci Biobehav Rev. 2007; 31(3):426–40. [PubMed: 17210179]
32. Levine JA, Eberhardt NL, Jensen MD. Role of nonexercise activity thermogenesis in resistance to
fat gain in humans. Science. 1999; 283(5399):212–4. [PubMed: 9880251]
33. Stunkard AJ, Harris JR, Pedersen NL, McClearn GE. The body-mass index of twins who have
Author Manuscript

been reared apart. N Engl J Med. 1990; 322(21):1483–7. [PubMed: 2336075]

34. Naleid AM, Grace MK, Cummings DE, Levine AS. Ghrelin induces feeding in the mesolimbic
reward pathway between the ventral tegmental area and the nucleus accumbens. Peptides. 2005;
26(11):2274–9. [PubMed: 16137788]
35. Wright, JD.; Wang, CY.; Kennedy-Stephenson, J.; Ervin, RB. Advance data from vital and health
statistics. 334. Hyattsville, Maryland: National Center for Health Statistics; 2003. Dietary intake of
ten key nutrients for public health, United States: 1999-2000.

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 11

36. Larsson U, Karlsson J, Sullivan M. Impact of overweight and obesity on health-related quality of
life--a Swedish population study. Int J Obes Relat Metab Disord. 2002; 26(3):417–24. [PubMed:
Author Manuscript

11896499]
37. Carr D, Friedman MA. Is obesity stigmatizing? Body weight, perceived discrimination, and
psychological well-being in the United States. J Health Soc Behav. 2005; 46(3):244–59. [PubMed:
16259147]
38. Onyike CU, Crum RM, Lee HB, Lyketsos CG, Eaton WW. Is obesity associated with major
depression? Results from the Third National Health and Nutrition Examination Survey. Am J
Epidemiol. 2003; 158(12):1139–47. [PubMed: 14652298]
39. Azarbad L, Gonder-Frederick L. Obesity in women. Psychiatr Clin North Am. 2010; 33(2):423–40.
[PubMed: 20385345]
40. Xu Y, Nedungadi TP, Zhu L, Sobhani N, Irani BG, Davis KE, et al. Distinct hypothalamic neurons
mediate estrogenic effects on energy homeostasis and reproduction. Cell Metab. 2011; 14(4):453–
65. [PubMed: 21982706]
Author Manuscript
Author Manuscript
Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 12

Study Importance Questions

Author Manuscript

What is already known about this subject?

• Differences in brain activity response to food and eating exist between people
with versus without obesity.

• Studies showing differences in brain activity response to food, generally focus

on one state (fasted or fed), and rarely include a group with BMI >35m/kg2
(severe obesity).

What does our study add?

• Our study uncovers a critical difference in the brain response to eating between
women with or without severe obesity. We uniquely establish no significant
difference in baseline brain perfusion, a surrogate of neural activity, between
Author Manuscript

groups at study onset.

• The difference, a failure in obesity of brain reward center activity to diminish

after a meal, uniquely augments previous findings by studying brain activity
across states (fasted and fed), providing within- and between-group
comparisons, and focusing on women with severe obesity.
Author Manuscript
Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 13
Author Manuscript
Author Manuscript
Author Manuscript

Figure 1. Lean Controls

1a. The group average image from lean controls in a fasted state shows areas of significant
activation during the imaging task (viewing food cues/pictures), contrasting activation
during food cues with directional arrows. These regions are significant at p<0.05, family-
wise-error corrected. Illustrative coronal slices show the anterior cingulate (ant cing),
dorsolateral frontal cortex (dlfc), basal ganglia (caudate and nucleus accumbens) and
hippocampal regions (hippo) as well as the ventral tegmental area in the midbrain, visual
cortex and cerebellar regions with extensive activations.
Author Manuscript

1b. After a meal and when rating themselves as satiated, the lean controls show highly
attenuated regions of brain activation to food cues compared to directional arrows. The same
illustrative coronal slices show limited regions of activation in the anterior cingulate,
dorsolateral frontal, and hippocampal cortex along with the visual cortex, relative to
directional arrows.
1c. Subtraction of the fed from the fasted group average images in lean controls (Lean,
fasted-fed) show regions in the lean controls where activations to food cues are greater in the
fasted state than in the fed state. Illustrative coronal slices show regions in the anterior

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 14

cingulate, dorsolateral frontal, and medial temporal cortex, striatum and nucleus accumbens
Author Manuscript

(nuc accum) as well as the ventral tegmental area/midbrain and cerebellum, with
significantly lower activation in the lean group while viewing food cues after eating than in a
fasted state. Cluster size and coordinates for all significant clusters are detailed in Table 2
and dense coronal slices from anterior to posterior brain are shown in Supporting Figure S2.
Author Manuscript
Author Manuscript
Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 15
Author Manuscript
Author Manuscript
Author Manuscript

Figure 2. Severe Obesity

2a. The group average image from the women with severe obesity in a fasted state shows
extensive areas of activation during the imaging task (viewing food cues/pictures) when
contrasting activation during food cues with directional arrows. The regions are significant at
p<0.05, family-wise-error corrected. The illustrative coronal sections show prominent
activation in the anterior cingulate, dorsolateral frontal, basal ganglia, and hippocampal
regions as well as the ventral tegmental area/midbrain, visual cortex and cerebellar regions.
The fasting activations in the group with severe obesity have the appearance of those in the
lean fasting scans.
2b. After a meal and when rating themselves as satiated, the women with severe obesity
Author Manuscript

show little attenuation of brain activation to food cues in the ‘hedonic’ regions of the brain,
including the nucleus accumbens, hippocampus, and ventral tegmental area/midbrain regions
where lean controls show significant activation attenuation.
2c. Subtraction of the fed from the fasting group with severe obesity average images (severe
obesity, fasted-fed) show regions in the group with severe obesity where activations to food
cues are greater in the fasted state than in the fed state. It is merely the insula, periventricular
regions (perivent) and midbrain regions, which show significantly lower activation in the
group with severe obesity after eating, and in a satiated state while viewing the food cues.

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 16

Cluster size and coordinates for all significant regions are detailed in Table 2; dense coronal
Author Manuscript

slices from anterior to posterior are available in Supporting Figure S3.

Author Manuscript
Author Manuscript
Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 17
Author Manuscript

Figure 3. Lean-Severe obesity (fasted-fed)

In the [fasted-fed] state × [lean-with severe obesity] group interaction (double subtraction),
we see the most conservative activation of regions where Lean controls deactivate after
eating but women with severe obesity maintain activation to the food cues even when no
Author Manuscript

longer hungry. Here it is the anterior cingulate and the dorsolateral frontal cortex as well as
the posterior cingulate cortex, precuneus (prec) and regions in the cerebellum that show
significant differences. Cluster size and coordinates for all significant regions distinguishing
fasted/fed states between the Lean/Severe obesity women are detailed in Table 4.
Author Manuscript
Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 18

Table 1
Group Characteristics
Author Manuscript

Lean Severe Obesity

Characteristic
(n=15) (n=15) p-valuea

Age in Years

Mean (SD) 45.1 (9.9) 40.6 (12.0) 0.27

Range 30-58 24-61

Height in Inches

Mean (SD) 65.5 (2.9) 64.5 (2.7) 0.34

Range 61-71 61-69.5

Weight in Pounds

Mean (SD) 136.3 (16.6) 254.2 (36.2) 0.001

Range 111-176.5 202-316

Author Manuscript

BMI in kg/m2

Mean (SD) 22.3 (2.0) 42.7 (4.8) 0.001

Range 18.7-24.6 35.7-50.2

Ethnicity, n (%) 0.39

Non-Hispanic 10 (67) 13 (87)

Hispanic 5 (33) 2 (13)

Race, n (%) 1.00

African-American 3 (20) 4 (27)

Caucasian 11 (73) 11 (73)

Asian 1 (7) 0 (0)

Axis I Diagnosisb n (%)

Author Manuscript

Active 0 (0) 2 (14) 0.48

Lifetime 2 (14) 6 (43) 0.21

QIDS-SR16

Mean (SD) 3.9 (2.2) 5.8 (3.8) 0.13

Note: The means presented in this table are the arithmetic means.

Abbreviations: BMI, Body Mass Index; QIDS-SR16, 16-item Quick Inventory for Depressive Symptomatology – Self-Report; SD, Standard
Deviation.
a
Continuous variable means were compared by t-test. Categorical variable frequencies were analyzed by Chi-square or Fisher's exact test.
b
As determined by Structural Clinical Interview for DSM-IV Axis I Disorders; n=14 per group.
Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 19

Table 2
Eating Measures
Author Manuscript

Measure Lean(n=15) With Obesity(n=15) p-valuea

intake standard meal kcals

Mean (SD) 301 (51)) 302 (66) 0.95

Range 176-337 135-337

Hungerb (fasted)

Mean (SD) 16 (30) -3 (23)

Hunger (fed)

Mean (SD) -24 (25) -37 (17)

Hunger (fasted v. fed; within group) lean, with obesity < 0.001, < 0.001

Hunger (lean v. with obesity; between group) 0.02

Author Manuscript

Fullness (fasted)

Mean (SD) -41 (9) -39 (14)

Fullness (fed)

Mean (SD) 10 (24) 13 (26)

Fullness (fasted v. fed; within group) lean, with obesity < 0.001, < 0.001

Fullness (lean v. with obesity; between group) 0.40

Within-group × between group interaction 0.45

Note: The means presented in this table are the arithmetic means.

Abbreviations: SD, Standard deviation, VAS, Visual analog scale

a
Continuous variable means were compared by t-test. Repeated-measures analysis of variance used to assess within group (fasted v. fed), between
group (lean v. with obesity), and within-group by between-group interaction.
Author Manuscript

b
Hunger/Fullness VAS scale -50 (least hunger/fullness ever experienced) to +50 (greatest hunger/fullness ever experienced).
Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 20

Table 3

Cluster size, t-values, peak coordinates and brain region labels for the lean group, and group with severe
Author Manuscript

obesity, (fasted-fed) contrast, significant at p = 0.05 (family-wise-error corrected).

Cluster size (voxels) T-value Coordinates: x,y,z Region

Lean Group

6361 4.97 8 -50 -18 Culmen, cerebellum anterior lobe

4.85 16 -88 -14 Lingual gyrus

4.85 10 -12 10 Thalamus

375 4.58 14 -36 32 Cingulate gyrus

3.02 14 -50 16 Precuneus

3.02 -6 -56 16 Posterior cingulate

6076 4.56 8 40 46 Medial frontal gyrus

4.27 -26 24 44 Middle frontal gyrus

Author Manuscript

4.23 -2 66 24 Superior medial frontal lobe

146 4.29 -64 -22 36 Postcentral gyrus

114 3.85 -10 -20 36 Cingulate gyrus

182 3.74 4 -34 -46 Medulla

3.44 2 -30 -38 Pons

Group with Severe Obesity

585 5.68 42 -20 4 Insula

4.02 46 -2 12 Insula

3.40 34 -32 2 Sub-lobar

735 5.23 -38 -2 -24 Sub-gyral Temporal lobe

4.35 -38 18 -26 Superior Temporal gyrus

Author Manuscript

4.34 -36 16 -34 Superior Temporal gyrus

2148 4.79 18 8 42 Sub-gyral Frontal lobe

4.64 12 14 56 Superior Frontal gyrus

4.58 26 -6 48 Middle Frontal gyrus

163 4.28 -4 -30 -36 Pons

3.42 14 -36 -38 Pons

2.57 0 -38 -42 Pons

233 3.71 48 28 -2 Inferior Frontal gyrus

3.09 52 16 -4 Inferior Frontal gyrus

156 3.48 -18 -36 -28 Culmen, Cerebellum anterior lobe

2.99 -16 -38 -16 Culmen, Cerebellum anterior lobe

Author Manuscript

2.44 -14 -44 -26 Culmen, Cerebellum anterior lobe

116 3.36 32 14 -24 Superior Temporal gyrus

3.12 34 26 -16 Inferior Frontal gyrus

127 3.11 -32 -4 64 Middle Frontal gyrus

3.03 -40 -14 48 Precentral gyrus

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 21

Cluster size (voxels) T-value Coordinates: x,y,z Region

2.84 -36 -10 54 Precentral gyrus
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.
 Puzziferri et al. Page 22

Table 4

Cluster size, t-values, peak coordinates and brain region labels for the lean group versus group with severe
Author Manuscript

obesity ([fasted-fed] × [lean-with obesity]) interaction, significant at p = 0.05 (family-wise-error corrected).

Cluster size (voxels) T-value Coordinates: x,y,z Region

526 5.03 8 -76 -34 Cerebellum posterior lobe

3.61 14 -76 -44 Cerebellum posterior lobe

3.05 8 -82 -22 Cerebellum posterior lobe

699 4.92 -32 46 8 Sub-gyral Frontal lobe

3.73 -20 50 14 Medial Frontal gyrus

3.15 -16 62 14 Superior Frontal gyrus

101 4.50 36 -6 38 Precentral gyrus

726 4.44 -28 10 68 Middle Frontal gyrus

4.14 -22 18 30 Sub-gyral Frontal lobe

Author Manuscript

3.54 -32 34 44 Middle Frontal gyrus

246 4.36 8 -48 -16 Cerebellum anterior lobe

3.17 -2 -50 -22 Cerebellum anterior lobe

405 4.20 10 32 14 Anterior Cingulate

3.38 6 44 12 Anterior Cingulate

3.36 12 28 36 Medial Frontal gyrus

201 3.74 40 -60 -46 Cerebellum posterior lobe

3.50 30 -52 -48 Cerebellum posterior lobe

776 3.62 14 -50 14 Sub-lobar

3.59 8 -54 18 Posterior Cingulate

3.34 -12 -58 18 Sub-lobar

Author Manuscript

432 3.60 26 32 48 Superior Frontal gyrus

3.55 26 18 42 Middle Frontal gyrus

114 3.55 6 40 50 Superior Frontal gyrus

2.69 -8 34 52 Superior Frontal gyrus

2.43 10 40 42 Medial Frontal gyrus

118 3.48 -8 -68 42 Precuneus

142 3.37 -2 -26 -4 Midbrain

3.32 6 -24 -2 Midbrain

175 3.04 -6 -36 48 Precuneus

2.92 -12 -50 38 Precuneus

2.74 -10 -48 30 Precuneus

Author Manuscript

Obesity (Silver Spring). Author manuscript; available in PMC 2017 April 01.