Obesity has become a serious problem, threatening public health in the United States, Europe, and even some developing countries such as China. In the United States, overeating and obesity have been reported by the American Medical Association as the cause of 280 184 deaths per year. According to the Centers for Disease Control and Prevention, prevalence of obesity was 12.8% of the total population in 1960 and 22.8% of the population in 1994. Among adults aged 20 to 74, obesity (body mass index [BMT) greater than 30) rates have doubled from approximately 15% to 27%, over the past two decades. Obesity rivals the deaths attributed to alcphol and tobacco (including eawonmental tobacco) smoke. Researchers have come to a consenses that obesity is a disease but often debate whether it is related to depression, or other medical disease, personality disorders, or addictions. Almost a decade ago, we reported on the similarities of overeating and obesity to classical addictions.1,2 Others have reported on the comorbid disorders most commonly associated with obesity and addictions. Since that time, neuroimaging studies have supported the hypothesis that loss of control over eating and obesity produces changes in the brain that are similar to those produced by drugs of abuse. As eating for pleasure, rather than survival, becomes more prevalent, neuroimaging changes similar to drugs of abuse may be expected. In addition, newly discovered messengers such as galanin and cocaine- and amphetamine-reguiated transcript have effects in modulating eating behavior and may play roles in obesity, alcoholism and other drug dependencies.
A convergence of evidence from the bench in neuroscience to functional magnetic resonance imaging (fMRI) neuroimaging and clinical experience and data support the hypothesis that overeating and obesity can be a substance, in this case food, abuse disorder. Binge-eating disorder but not anorexia nervosa or bulimia may be considered in the spectrum of overeating and obesity. Applying new research methodologies to obesity and also addiction studies, therefore, may offer hope for understanding and the development of common treatments. In this context, the primary purpose of this article is to review a novel fMRI method for studying the neuro-hormonal mediation of hunger and satiety based on our previous work, focusing on the hypothalamus and its associated signaling pathways in the control of eating and obesity.3-4
Functional neuroimaging has provided a useful tool for the analysis of brain activity changes during a variety of cognitive and affective processes. The imaging techniques, such as positron emission tomography (PET) and fMRI, have allowed for the extensive in vivo characterization of the major components in the central nervous system and, most importantly, their interactive connections.5 But despite progress toward understanding the nervous system, functional neuroimaging studies have vastly overlooked influences of the endocrine system on the brain, and vice versa. One primary reason is that neuronal and hormonal processes may act on different time scales. With respect to the rapid changes (milliseconds to seconds) of neuronal activity, the responses of hormonal secretion mediated throughout the whole body are relatively slow (minutes to hours or longer). In neuroimaging studies, the slower time scale makes it difficult to trace down the time when these two systems interact.
As the first task of functional imaging is to localize where in the brain a neuronal event happens, we targeted the hypothalamus in our imaging study. It has been known for a long time that the hypothalamus plays a role in the control of eating behavior, and lesion and neural degeneration experiments implicate the hypothalamus as a regulating center of hunger and satiety.6"8 Research using animal models has demonstrated that damaging the ventromedial hypothalamus leads to a voracious appetite, while damaging the lateral hypothalamus results in a loss of appetite. However, the picture has grown in complexity in recent years. For example, the paraventricular nucleus in the lateral hypothalamus has been shown to have anatomical projections to the arcuate nucleus; the activities of these brain regions are modulated by neuropeptide Y, a central hormone playing a role in the regulation of appetite. The arcuate nucleus in turn sends projections to the ventromedial nucleus; these structures are both under the influence of sympathetic pathways and two peripheral hormonal signals: insulin and leptin.
In addition to the insulin and leptin signaling pathways (which may exert anorexic effects on modulating hypothalamic function in the control of food intake), animal studies have shown a role of the peripheral nervous system in stimulating or attenuating hypothalamic activity.9 The peptide cholecystokinin exemplifies the complex interplay between the endocrine and nervous systems in the regulation of hunger and satiety. Produced in the intestines following feeding, cholecystokinin binds to receptors in the stomach to prevent gastric emptying. The resulting stomach distension sends the feedback signals to the hypothalamus via the vagus nerve to promote satiety. Under such a complex schema, stationary brain mapping is not sufficient to depict changes in the brain due to the interactions with those signals.
The next task of functional imaging, therefore, is to detect when, as well as where, a brain response occurs, particularly if the timing of neural-endocrinal interactions is not known in advance. With the sympathetic and parasympathetic pathways well studied, a dynamic brain mapping method is important for directly analyzing how the autonomic pathways elicit changes in those physiologically functioning processes in a living brain. Although animal studies have provided much insight into the endocrine system, its imminent influences upon neural functions and how they give rise to the awareness (eg, the sense of hunger) of such autonomic changes have yet to be investigated in humans. Our goal is to provide new tools for fMRI applications in the study of neural correlates of hunger and satiety, regarding their both functional and structural organizations.
NEW METHODS AND FINDINGS
Positron emission tomography studies have established a correlation between hunger and increased blood flow in the vicinity of the hypothalamus and thalamus.10-12 The PET studies have likewise correlated satiety with the inferior parietal lobe and dorsolateral and ventromedial prefrontal cortices. However, PET technique does not provide a time course or dynamic depiction of brain activation and has the limitations of low spatial resolution. The fMRI technique can precisely distinguish between the hypothalamic and dialamic activity whereas PET cannot. The fMRI scan also offers greater temporal resolution (in a few seconds) and can be performed repeatedly on a single subject. Again, the PET scan cannot provide these features.
While fMRI is ideal for detecting changes in brain activity of a relatively short duration, an investigator usually needs to know the temporal pattern of such activity before attempts can be made to localize or analyze this activity. The ambiguity in precisely predicting when an endocrine change will induce a neural response or vice versa has been a significant barrier for in vivo studies of the interrelationships of these two systems with fMRI techniques.
A new fMRI technique, temporal clustering analysis, may overcome this methodological hurdle.13 In our experiment, the subjects underwent a lengthy continuous functional scan with simultaneous blood sampling (Figure 1 ). The question here is when and where there are brain responses after oral glucose ingestion. Based on a probability distribution of the overall brain voxels (a voxel is a 3-D computerized representation of brain volume analogous to a pixel in a 2-D picture) that reach the maximal signal change during and following stimulation, temporal clustering analysis converts a multiple-dimension fMRI data space into a simple relationship between the number of voxels reaching maximum and the time, to give the temporal maxima.3,14
Figure 1. Illustration of an fMRI experiment paradigm for study of neurobiocbemical interaction. An imaging protocol is shown for the time course of brain activity before (fasting), during, and after oral glucose intake. Blood was drawn simultaneously for measuring the plasma glucose and insulin concentrations.
After temporal clustering analysis has identified the onset and duration of activation, functional mapping techniques can be applied to further localize the active brain regions. In our study, a significant decrease of neural activity was found in the ventromedial hypothalamus, which occurred 10 minutes after glucose ingestion. Furthermore, the hypothalamic activity was significantly correlated to the fasting plasma insulin concentration. These results provide new insights into die dynamics of brain activation following oral glucose intake and may help us understand how our brain generates a signal of satiety, physiologically as well as psychologically.
Further fMRI analysis has indicated that the hypothalamic response to glucose ingestion is not only weaker but significantly delayed (?-9 min longer) in obese individuals.4 This finding raises the possibility that a delay or inability to reach satiety could perpetuate obesity and make treatment approaches difficult. Obese patients cannot remain abstinent from food. Eating until full becomes quite difficult if the time from initiation of eating until recognition is exaggerated and delayed. Behavioral interventions, such as eating when hungry, relearning how to eat slowly, learning to appreciate a smaller plate, and others may be helpful.
Animal studies have suggested that insulin may act as a modulator of appetite.13 While the secretion of insulin has been shown to increase immediately following a meal, the temporal relationship between insulin release and later (~10 min delayed) neural activity is not clear.16 We propose a moment-to-moment analysis model to better describe this relationship (Figure 2, page 130). In short, this model expresses the correlations between plasma insulin and brain response as a function of time, r(t). In doing so, it becomes possible to monitor second by second changes in the correlation of insulin levels to the brain activities (alter at different slopes). Our data indicate a dynamic brain-insulin interaction, implicating that insulin may have direct modulatory effect on the hypothalamic response. Furthermore, this modulatory effect may be diminished in people with obesity. With highdensity insulin measurement, the moment-to-moment model shows great promise to be applied to "real-time" imaging of brain-hormone interaction.
Figure 2. A dynamic model of neuro-hormonal interaction, as shown by moment-to-moment interrelationship of fMRI blood oxygenation level dependent (BOLD) signal and measured insulin secretion in blood samples.
The inlaid figure shows the temporal dependence of the correlation [r(t)J between the fMRI response at the hypothalamus and hormonal signal. At 0, 15, and 120 minutes after oral glucose intake there was significant correlation (P < 0.05) between the brain response and insulin levels in the lean subjects. There was no significant2 correlation in the obese subjects, even though there was a trend of correlation (r = 0.54; P= 0.1) 15 minutes after eating.
Knowing leptin is intimately related to insulin and insulin resistance in the regulation of energy balance, it would be important to consider the possible consequence of changes in leptin regulation during the control of food intake, which may lead to the altered neuroendocrine functions.7,9 While our fMRI study shows promising results for brain-insulin interaction, the correlation between circulating leptin levels and the hypothalamic response after eating is intriguing.4 As expected, the fasting plasma leptin concentration was significantly increased in obese compared to lean subjects. When the data from all the subjects (10 lean and 10 obese) were analyzed, no relationship between the plasma leptin concentrations and the fMRI response was observed (r = 0.26, NS). But when the data obtained from seven lean subjects (BMI < 25) was analyzed, the plasma leptin levels showed a significant correlation with the maximal fMRI signal intensity changes in the hypothalamus (r = 0.82, P < 0.02). These results indicated that the brain's response may be modulated by the baseline leptin, but this modulation could be impaired in obese subjects. These results also suggested that leptin may be a link between obesity, the insulin resistance syndrome (eg, hyperinsulinemia, hypertension, dyslipidemia, and dysfibrinolysis), and the increased risk for diabetes.
It is arguable that leptin alone may not be a factor contributing to those conditions but the central nervous system and its interaction with leptin could be involved. With such an integrated view, our fMRI study may open an avenue to investigate the central mechanisms relating obesity to leptin, and, importantly, the consequence of altered brain-leptin interaction in obese people. If it is true, this theory should be considered in light of the causes for obesity; for example, overeating and leptin resistance.17 The fMRI findings of abnormal hypothalamic function could account for eating problems leading to hyperglycemia, and thus open the door for testing new treatments of these illnesses. For example, insulin therapy may be feasible in overeating, high-risk individuals to better promote satiety. These treatments could potentially prevent the onset of type 2 diabetes in high-risk populations.
Limitations. Our study of eating and obesity has limitations that are due to the physical dimension of the MRI systems under use. The size of a magnet bore space allowed for a subject is not designed in a way for seriously obese people (those with BMI approximately 35 or greater). Therefore, setting a bodysize criterion for subject recruitment is not a trivial issue in the imaging study of obesity. Actually, any obese subject with a circumference of chest (including arms), waist, or thighs larger than 4.8 feet should be excluded from scanning. This restriction is necessary for two reasons: not only is the experimenter's accessibility to the subject's FV for blood sampling limited with obese subjects, but the experimenter's ability to manage the Oral ingestion of glucose solution or other substances becomes impaired. These problems are particularly pronounced in clinical populations when there are other complications, such as short limbs in patients with Prader-Willi Syndrome (PWS).
Besides those physical limitations, physiological aspects of the fMRI blood oxygen level dependent (BOLD) signaling mechanism are still under intensive study.18 Questions have been raised regarding the interpretation of the decreased fMRI signal (inhibited activity) at the hypothalamus after eating as compared to a fasting baseline.3,4 Subjects in the study were fasting overnight, so an enhancement of activity in the hypothalamic hunger center was expected during fasting; the ingestion of glucose may have reduced this baseline activity. An alternative explanation is that the glucose activated inhibitory pathways to the hypothalamus, causing the decrease in hypothalamic activity.
Anorexia Nervosa and Bulimia
One of our ongoing fMRI projects is to examine how emotional states affect eating. In general, we aim to understand the relationship between emotions and compulsive behaviors, including binge eating. One of the basic emotions relevant to eating and eating disorders is disgust.19 Disgust is thought to be associated with distaste, an adaptive mechanism to avoid eating potentially harmful foods.20 However, disgust also plays a profound role in the processes of communicating cultural and mortal values21, and in other disorders.21-24
A cardinal character of anorexia is the dread of body-weight gain, which may construct part of me diagnostic criteria. Other negative emotions such as sadness and anxiety, on the other hand, may provoke overeating. We are specifically interested in brain representation of the reward value of food and food-related stimuli and how hunger and satiation may modulate the brain response under emotional challenge (Figure 3). Patients with anorexia may be disgusted by fatness but also by fat people. Patients with bulimia nervosa may be disgusted by the fact that they just ate that pizza. Our preliminary data are consistent with literature regarding the functional neuroanatomy and neurocircuitry of satiety and emotional modulation.25 Activation was found in the insula, the orbitofrontal cortex, and the body of the right caudate nucleus during the hunger state. In addition, the insula and medial prefrontal cortex had increased activation during the sated state than the hunger state. Food stimuli induced brain response involving the amygdala and orbitofrontal cortex, which has been implicated in memory function.26 Change of satiety signaling has also been associated with activity in the orbitofrontal cortex.27,28 A single-photon computed tomography study showed abnormal activation in the prefrontal cortex in obese binge eating subjects during exposure to food.29
Figure 3. Visual stimulation in an fMRI experiment for human subjects fasting for 12 hours. The stimulation was presented before and after food intake in two fMRI scans. The photos are from a standard International Affective System* rated by emotional valence and arousal.
* Center for the Study of Emotion and Attention [CSEA-NIMH]. The International Affective Picture System: Digitized Photographs. Gainesville, FIa: The Center for Research in Psychophysiology, University of Florida; 2001.
Alteration in satiety signaling may be tied to binge eating. Another further fMRI application using the fast-eating paradigm is to identify functional neural correlates of hunger and satiation in PWS. Prader-Willi syndrome is a neurogenetic multi -system disorder characterized by infantile hypotonia, mental retardation, hypogonadism, obsessivecompulsive behaviors, self-injury, and pleasureless hyperphagia with high risk of obesity. Therefore, PWS is a good model for study of obesity. In addition, binge eating in patients with PWS can be assessed by imaging in vivo brain response during food intake or during food-related psychological challenge before and after treatment.30 We aim to resolve whether satiety signaling in PWS patients with binge eating and obesity is different in terms of neurohormonal interaction and the timing of such interaction. Our preliminary results demonstrated changes in the patterns of brain activity following oral glucose administration in the regions of the medial prefrontal cortex, insula, and hypothalamus. The PWS subjects were scanned two times, prior and post 8 weeks of treatment with topiramate. While topiramate was helpful for selfinjury, there was a small but not significant decrease in caloric intake by baseline and post-treatment hour-long caloric intake analyses. There is no significant difference in the magnitudes and onset of the brain activity changes between the first and second functional scans. However, a prominent latency of 24 minutes (compared to ~10 minutes in normals) from time of glucose administration to increased brain activity was found, which is consistent with the satiety dysfunction in individuals with PWS.
Now that functional neuroimaging investigators are able to determine when to look for endocrine related changes in the brain, new imaging methods may possibly tear down the barrier between nervous system and endocrine system research. With new functional imaging approaches, psychiatric disorders that are associated with eating, such as anorexia nervosa and bulimia nervosa, could be analyzed for better insight of the underlying brain mechanisms and for assessment of new therapeutic drugs. The new fMRI methods can also be applied toward neuroendocrine studies of, and well beyond, the realm of eating problems. Again these studies should be in light of the interaction between neuronal activities measured by functional imaging and the hormonal signais by blood sampling.
1. Jonas TM, Gold MS. Cocaine abuse and eating disorders. Lancet. 1986;1(8447):390-391.
2. Gold MS, Johnson CR, Stennie K. Eating Disorders. In: Lowinson JH, Ruiz P, Millman RB, Langrod, JG, eds. Substance Abuse: A Comprehensive Textbook. 3rd ed. Philadelphia, Pa: Lippincott, Williams and Wilkins; 1997:319-330.
3. Liu Y, Gao JH, Liu HL, Fox PT. The temporal response of the brain after eating revealed by functional MRl. Nature. 2000;405:1058-1062.
4. Matsuda M, Liu Y, Mahankali S, et al. Altered Hypothalamic response to oral glucose intake in obese humans. Diabetes. 1999;48:1801-1806.
5. Liu Y, Gao J-H, Liotti M, Pu Y, Fox PT. Temporal dissociation of parallel processing in the human subcortical outputs. Nature. 1999;400:364-367.
6. Grossman SP. Role of the hypothalamus in the regulation of food and water intake. Psychol Rev. 1975;82:200-224.
7. Schwartz MW, Woods SC, Porte D Jr, Seeley RJ, Baskin DG. Central nervous system control of food intake. Nature. 2000; 404:661-671.
8. Greenspan FS, Gardner DG, eds. Basic and Clinical Endocrinology. 6th edition. New York, NY: McGraw-Hill /Appleton & Lange; 2001.
9. Woods SC, Seeley RJ, Porte D Jr, Schwartz MW. Signals that regulate food intake and energy homeostasis. Science. 1998;280:1378-1383.
10. Tataranni PA, Gautier JF, Chen K, et al. Neuroanatomical correlates of hunger and satiation in humans using positron emission tomography. Proc Natl Acad Sci USA. 1999;96:4569-4574.
11. Gautier JF> Chen K, Uecker A, et al. Regions of the human brain affected during a liquid-meal taste perception in the fasting state: a positron emission tomography study. Am J CHn Nutr. 1999;70:806-810.
12. Gautier JF, Chen K, Salbe AD, et al. Differential brain responses to satiation in obese and lean men. Diabetes. 2000;49:838-846.
13. Liu Y, Fox PT, Liu H-L, Matsuda M, Mao J, Gao J-H. Temporal clustering analysis for tracking the maximal fMRI response in human brain. Proceedings of the International Society for Magnetic Resonance Medicine. 2000;8:238.
14. Liu Y, Gao J-H, Liu H-L, Matsuda M, Mao J, Fox PT. Temporal maxima in fMRI response. Neuroimage. 2000;11:S530.
15. Kalra SP. Circumventing leptin resistance for weight control. Proc Natl Acad Sci USA. 2001;98:4279-4281.
16. Bellisle F, Louis-Sylvestre J, Demozay F, Blazy D, Le Magnen J. Cephalic phase of insulin secretion and food stimulation in humans: a new perspective. Am J Physiol. 1985,249(6 Pt 1)£639-E645.
17. Kalra SP, Dube MG, Pu S, Xu B, Horvath TL, Kalra PS. Interacting appetite-regulating pathways in the hypothalamic regulation of body weight. Endocr Rev. 1999;20:68-100.
18. Logothetis NK, Pauls J, Augath M, Trinath T, Oeltermann A. Neurophysiological investigation of the basis of the fMRI signal. Nature. 2001;412:150-157.
19. Troop NA, Murphy F, Bramon E, Treasure JL. Disgust sensitivity in eating disorders: a preliminary investigation. lnt J Eat Disord. 2000;27:446-451.
20. Rozin P, Fallon A. A perspective on disgust. Psychol Rev. 1987,-94:23-41.
21. Power MJ, Dalgleish T, Power M, Dalgleish T. Cognition and Emotion: From Order to Disorder. Hove, UK: Psychology Press; 1997.
22. Gilbert P. Depression: The Evolution ofPoweriessness. Hove, UK: Erlbaum; 1992.
23. Phillips ML, Senior C, Fahy T, David AS. Disgust- the forgotten emotion of psychiatry. Br f Psychiatry. 1998;172:373-375.
24. Liu X Stein DJ, Shapira NA, Goodman WK. The Role of Disgust in ObsessiveCompulsive Disorder Assessed by Functional MRI !abstract]. American College of Neuropsychopharmacology; 2001:329.
25. James GA, He AG, Wagner-Miller A, Liu, Y. An fMRI study of hunger and insula during a fasting paradigm. Neuroimage Neurolmage. 2002;13:311S.
26. Morris JS, Dolan RJ. Involvement of , human amygdala and orbitofrontal cortex in hunger-enhanced memory for food stimuli. J Neurosci. 2001;21:5304-5310.
27. OTtoherty J, Rolls ET, Francis S, Bowtell R, McGlone F, Kobal G, Renner B, Ahne G. Sensory-specific satiety-related olfactory activation of the human orbitofrontal cortex. Neuroreport. 2000;11:893-897.
28. Small DM, Zatorre RJ, Dagher A, Evans AC, Jones-Gotman M. Changes in brain activity related to eating chocolate: from pleasure to aversion. Brain. 2001, -124(Pt 9):1720-1733.
29. Karhunen LJ, Vanninen EJ, Kuikka JT, Lappalainen RI, Tiihonen J, Uusitupa ML Regional cerebral blood flow during exposure to food in obese binge eating women. Psychiatry Res. 2000,-99:29-42.
30. Shapira NA Lessig MC Liu X He G, Driscoll D. Neuroanatomical Correlates of Satiation and Hunger in PraderWilli Syndrome: A Study Using fMRI [abstract]. Presented at: 17m Annual PWSA Scientific Conference; July 11, 2002; Salt Lake City, Utah.