Neuroimaging the interaction of mind and metabolism in humans

Abstract

Hormonal and metabolic signals interact with neural circuits orchestrating behavior to guide food intake. Neuroimaging techniques such as functional magnetic resonance imaging (fMRI) enable the identification of where in the brain particular mental processes like desire, satiety and pleasure occur. Once these neural circuits are described it then becomes possible to determine how metabolic and hormonal signals can alter brain response to influence psychological states and decision-making processes to guide intake. Here, we provide an overview of the contributions of functional neuroimaging to the understanding of how subjective and neural responses to food and food cues interact with metabolic/hormonal factors.

Introduction

In the mid 1990s, high resolution functional neuroimaging techniques like positron emission tomography (PET) and functional magnetic resonance imaging (fMRI) became available to study the human brain. These tools afforded, for the first time, the ability to non-invasively identify in humans where in the brain particular mental processes occur. Since subjective experiences like pleasure, desire to eat, and hunger contribute to decisions about whether or not to eat, it is important to understand how these psychological variables are represented in the brain. Moreover, once these neural circuits are described it is then possible to determine how metabolic and hormonal signals alter brain response to influence psychological states and decision-making processes to guide intake. Here we provide a brief overview of the contributions of functional neuroimaging to the understanding of how subjective and neural responses to food and food cues interact with metabolic/hormonal factors.

Neural correlates of subjective responses to food

Many of the early studies of ingestive behavior focused on determining how subjective mental processes, like desire and pleasure, influence brain response to food and food cues. Gordon and colleagues showed that viewing high calorie food pictures increases desire to eat and decreases temporo-insular response [1]. Karhunen and colleagues showed that obese women have enhanced responses to food cues in the parietal cortex and that these responses correlate with ratings of perceived hunger [2]. Distinct from hunger are experiences of craving, which refer to an intense desire for a specific food item. In a study where subjects were fed a monotonous diet for 3 day and subsequently exposed to appetitive food cues in the MR scanner, subjects reported experiencing cravings and these cravings were associated with brain response in the insular cortex and dorsal striatum [3].

In an early study of the neural correlates of food consumption, we showed that responses in key reward areas, like the midbrain, striatum, insula and medial orbitofrontal cortex, correlate positively with ratings of pleasure experienced while eating chocolate when subjects were motivated to eat. However, with continued chocolate consumption, prefrontal circuits became active with responses correlating with decreasing pleasure that resulted from eating the chocolate to beyond satiety [4] (Fig. 1). These results, which have since been replicated and refined [5,6], complimented findings from Tataranni and colleagues showing that resting responses in the hypothalamus, insula and other limbic regions are elevated during hunger, whereas response in prefrontal regions are elevated in the sated state [7]. Similar effects of feeding on brain response to food cues (odors, images and menus) have also been reported [8–12].

Fig. 1.

Cortical regions demonstrating significant rCBF correlations with affective rating for question (ii). Regression analyses were used to correlate rCBF from averaged PET data (Choc 1−Choc 7) with affective ratings taken immediately after these scans (see Methods). Correlations are shown as t statistic images superimposed on corresponding averaged MRI scans. The t statistic ranges for each set of images are coded by color bars, one in each box. Bar graphs represent normalized CBF in an 8 mm radius surrounding the peak. The y-axis corresponds to normalized activity and the bars along the x-axis represent scans. The three colours represent scan type and correspond to the coloured bars in Fig. 2. Each bar graph corresponds to activations indicated by a turquoise line. (A) Coronal section taken at y=1 showing the decrease in rCBF in the primary gustatory area (bilaterally in the anterior insula/frontal operculum and in the right ventral insula). (B) Coronal section taken at y=−26 showing decreases in rCBF in the left thalamus and medial midbrain (possibly corresponding to the ventral tegmental area). (C) Sagittal section taken at x=−1 showing decreases in rCBF in the subcallosal region, thalamus and midbrain. (D) Sagittal section taken at x=42 showing the increase in rCBF in the right caudolateral orbitofrontal cortex. Activation is also evident in the motor and premotor areas. (E) Sagittal section taken at x=8 showing an increase in rCBF in the posterior cingulate gyrus (peak at 8, −30, 45) in subtraction analysis Choc 1—water-post (see Table 3 and Results section). This was the only region where CBF was consistently greater in affective scans regardless of valence, compared with the neutral chocolate scan (Choc 4) and the two water baseline scans (water-pre and water-post). (F) Horizontal section at z=12 showing an increase in rCBF in the retrosplenial cortex (area 30) that correlated with affective rating (ii) but not the affective rating (i) when scan order was covaried out of the regression analysis. Taken with permission from Small et al. [4].

Together these findings established that subjective states can be associated with brain response to food and food cues. They also suggested that there are separate circuits underlying meal initiation and meal termination. In particular, the striatum, insula, and medial orbitofrontal cortex are frequently associated with subjective states that might promote meal initiation, such as desire, hunger, and experienced pleasantness, whereas response in the prefrontal cortex is often associated with subjective states that might promote meal termination, such as satiety. The majority of more recent work supports and extends this general framework (e.g., [13–15] though see [5]). For example, a number of studies have shown that food consumption interacts with individual factors like dietary restraint to influence brain responses to food in amygdala and striatum [13,15–17] and that internal state interacts with stimulus features such as caloric density to influence response [14].

Anticipatory versus consummatory responses

A second general theme to emerge from the neuroimaging literature has been that distinct circuits respond to food cues (e.g., pictures of foods) versus food receipt (e.g., tasting a milkshake). For example, the ventral striatum and amygdala tend to respond preferentially to cues that predict the arrival of foods compared to the actual receipt of the foods [18,19]. Critically, these distinct responses seem to be independent of the modality being stimulated (e.g., visual versus gustatory) and are rather more sensitive to learning associations between cues (sight, aroma) and reinforcer availability (e.g., milkshake). For example, the ventral striatum and amygdala respond to visual stimuli that predict food aromas [10], but also to food aromas that predict delivery of juice or milkshake [19]. Thus food aromas drive responses in these regions preferentially when they represent the cue compared to the outcome. Similarly, ventral striatal responses to juice depend upon whether or not juice delivery is predicted or unpredicted [20–22]. Thus expectation, prediction and outcome appear to be key determinants of these differential brain responses during anticipatory and consummatory phases of feeding. This point is important to keep in mind when evaluating the influences of metabolic and hormonal signals on brain response to food and food cues because the same variable might have opposing effects on response in a single region depending upon whether the response is to a food picture or to food receipt. For example, several studies have shown that body mass index (BMI) is inversely related to striatal response to milkshake receipt [23–25] but positively related to striatal response to visual food cues [26,27]. One parsimonious explanation of these results is that they reflect the differential influence of body mass index on anticipatory versus consummatory responses to food.

Cognitive influences on brain response to food: Neuromarketing

An intriguing focus of recent research has been on the influence of brand, price, beliefs and expectations on perceptual responses to food cues and food receipt [28–33]. For example, response in the medial orbitofrontal region during wine tasting increases with perceived pleasure, which is manipulated by misleading subjects about price [31]. More specifically, pleasantness ratings and medial orbitofrontal cortex responses are higher to the same wine believed to cost $45 vs. $5, or $90 vs. $10 (Fig. 2). Moreover, as illustrated in Fig. 2, the magnitude of the effect reflects the price differential! Brand knowledge can also modify response in this region [28], while healthy food choices appear to rely, in part, on the ability of prefrontal circuits to modify response in this medial orbital region to reflect information about health [32,33]. These studies indicate that information about price, brand, and health can be incorporated into the value signal computed in the medial orbitofrontal cortex. This emerging area of research is of potential importance to metabolic research because of the potential for beliefs to influence metabolism and vice versa. For example, a recent study showed that ghrelin levels decrease faster following milkshake consumption if individuals believe the milkshake to be “indulgent” vs. “sensible”[34].

Fig. 2.

The effect of price on each wine. (A) Wine 1: averaged time courses in the medial OFC voxels shown in B (error bars denote standard errors). (B) Wine 1: activity in the mOFC was higher for the high- ($45) than the low-price condition ($5). Activation maps are shown at a threshold of P<0.001 uncorrected and with an extend threshold of five voxels. (C) Wine 1: activity in the vmPFC was also selected by the same contrast. (D) Wine 2: averaged time courses in the medial OFC voxels shown in (E). (E) Wine 2: activity in the mOFC was higher for the high- ($90) than for the low-price condition ($10). (F) Wine 2: activity in the vmPFC was higher for the same contrast. Taken with permission from Plassmann et al. [31].

The influence of hormonal and metabolic factors on brain response to food

Of particular interest for metabolic research is determining how circuits representing the affective and incentive components of food interact with hormonal and metabolic factors. Early studies correlated hormonal and metabolic factors with brain response at rest. Changes in plasma concentrations of free fatty acids correlate with changes in resting dorsolateral prefrontal response, with larger free fatty acid changes post vs. pre meal associated with larger brain changes post vs. pre meal [7]. The opposite relationship is observed between changes in plasma insulin response and response in insular and orbitofrontal cortex. Using an innovative temporal clustering analysis, Liu and colleagues investigated the timing of brain responses following ingestion of glucose [35]. Two peaks were identified. The first occurred 1–3  min following ingestion and included the orbitofrontal cortex, supplementary motor cortex, somatosensory cortex, cerebellum and anterior cingulate cortex. The second occurred 7–13  min post-ingestion and included the hypothalamus. Here, response decreased and this decrease correlated with fasting plasma insulin concentration, suggesting a relationship between glucose ingestion and the delayed response in the hypothalamus. In contrast, responses identified in the first temporal cluster showed no relationship with insulin levels. These early studies provided clear evidence for metabolic modulation of postprandial brain responses in humans and suggested that these responses have specific targets that are likely to play distinct roles in regulating intake.

This hypothesis has been borne out in more recent studies where metabolic state or hormonal levels have been directly manipulated. Peptide YY (PYY) is a physiological gut-derived satiety signal. Batterham et al. [36] performed intravenous injection of PYY in subjects undergoing fMRI. They found that orbitofrontal cortex response predicted subsequent intake when PYY had been injected to mimic the fed state. In contrast, during hunger and low levels of PYY, hypothalamic responses predicted feeding (Fig. 3). Thus, intravenous PYY injection switched brain circuits governing feeding from homeostatic to hedonic [36].

Fig. 3.

Group mean time series of change in signal (%) extracted from posterior hypothalamic region (a) and left caudolateral OFC (b) on saline (green, open, inverted triangles) and PYY (black, open circles) study days and plasma PYY concentrations on saline (blue, filled circles) and PYY (red, filled squares) study days. Data are mean±s.e.m. (n=8 per group). ((c)–(f)), Correlation plots of signal change (%) from the peak voxel in the posterior hypothalamus ((c) and (e)) and left caudolateral OFC ((d) and (f)) and caloric intake for each subject on saline ((c) and (d)) inverted, open, green triangles) and PYY study days ((e) and (f)) open, black circles). Lines indicate linear regression. Corresponding r and P values are displayed on the plots. Taken with permission from Batterham et al. [36].

Leptin is a circulating hormone secreted by adipocytes that acts on receptors in the hypothalamus to inhibit appetite [37,38]. Farooqi and colleagues measured brain response to food versus nonfood images in the fed and fasted state in two individuals with congenital leptin deficiency while they were leptin-deficient and leptin-treated [39]. Before leptin treatment response in the ventral striatum to food versus nonfood images correlated positively with liking ratings of food images in both fed and fasted states. After treatment the correlation with liking ratings was only present in the fasted state. The authors therefore suggest that the findings indicate that leptin modulates feeding-related mesolimbic sensitivity to visual food cues.

Page and colleagues combined fMRI measurement of the blood oxygen level dependent signal (BOLD) with a stepped hyperinsulinemic–euglycemic–hypoglycemic clamp protocol to investigate the effect of circulating glucose on brain and perceptual response to food images [40]. A mild hypoglycemic state was associated with greater response in insula and striatum and the magnitude of response correlated with subjective ratings of food wanting of high calorie food images. In contrast, and in keeping with the role of prefrontal circuits in meal termination, the euglycemic state was associated with greater prefrontal response and less interest in food (Fig. 4). Thus, as might be expected, circulating glucose levels have the opposite effect on “meal initiation” and “meal termination” circuits.

Fig. 4.

Condition×task effects. (A) Axial slices with group averages (n=14), covaried for BMI, showing brain response to food (high-calorie and low-calorie) cues under euglycemia compared with mild hypoglycemia (threshold of P<0.05, 2-tailed, FWE whole brain corrected). (B) Wanting and liking ratings for food during euglycemia (gray bars) and mild hypoglycemia (black bars). *P=0.02. (C) Brain response specifically to high-calorie food images under euglycemia compared with mild hypoglycemia (threshold of P<0.05, 2-tailed, FWE whole brain corrected). (D) Wanting and liking ratings for high-calorie foods during euglycemia (gray bars) and mild hypoglycemia (black bars); **P=0.006. Red/orange areas show greater activity, and blue areas indicate more suppressed activity during euglycemia relative to hypoglycemia. MNI coordinates were used to define brain regions. Taken with permission from Page et al. [41].

Hormones may also act on brain circuits to alter subjective states. Malik and colleagues showed that intravenous ghrelin administration enhances responses in insula, striatum, amygdala and orbitofrontal cortex to food pictures, with response in the latter two regions correlating with perceived hunger [41]. Insulin levels have also been associated with increased brain glucose metabolism in the ventral striatum and prefrontal cortex and decreased metabolism in the amygdala and cerebellum [42]. Fasting plasma levels of insulin are positively associated with response in the hippocampus and negatively associated with response in the medial prefrontal cortex in response to viewing high-calorie food images [43].

To investigate how a standardized caloric intake alters blood plasma levels of glucose and insulin and how these changes then translate into the neural and behavioral response to pictures of palatable food, Kroemer and colleagues examined brain response to food images before and after a glucose challenge [44]. Food pictures increased appetite while glucose ingestion decreased appetite. Importantly, this effect was greater for appetite ratings following exposure to food vs. nonfood images. Moreover, increase in plasma insulin following glucose load correlated with mean appetite ratings and reductions in insula and cerebellar response following food but not nonfood image exposure (Fig. 5). This demonstrates that insulin response to glucose ingestion influences responses in circuits responsible for generating conscious perception of appetite in response to food pictures.

Fig. 5.

Negative interaction with increases in insulin after the administration of glucose for the contrast food pictures-control pictures (P<0.001, uncorrected). Scatter plots show the insulin increase plotted against the mean signal change extracted from anatomical masks of the fusiform gyrus, the orbitofrontal cortex, and the ventral striatum. Yellow circles indicate the clusters of activation inside those regions of interest that survived the threshold. Note that the numbers refer to the respective MNI coordinate. Taken with permission from Kroemer et al. [44].

It should be noted that psychological factors have also been shown to influence homeostatic responses. Smeets and colleagues examined hypothalamic BOLD response to an oral versus an intravenous glucose challenge [45]. Both challenges reduced hypothalamic response but these responses were not identical. Compared to the intravenous challenge, the oral challenge showed a delayed, but more robust and persistent response (Fig. 6). This result indicates that the perceptual correlates that accompany the consumption of glucose contribute to the effect of this nutrient on hypothalamic circuits. The finding provides a clear example of the importance of simultaneously considering metabolic and psychological factors.

Fig. 6.

Mean±SD fMRI signal changes per minute in the hypothalamus in response to oral and iv glucose administration. ■, oral glucose; ▴, iv glucose; □, oral water; △, iv saline; t=0 min, onset of treatment. *Significant difference between treatment and vehicle (Student’s t-tests with Bonferroni-corrected threshold of P=0.0013). Horizontal black bar indicates approximate duration of treatment: drinking took ∼2 min, and iv treatments took ∼3 min to complete. Taken with permission from Smeets et al. [45].

What is also fascinating is that the metabolic system appears to interact with networks beyond circuits underlying food reward. It is now established that obesity is a risk factor for dementia [46]. One of the most consistent cognitive changes observed in obesity is decreased working memory, which refers to the ability to hold information “online” to carry out complex tasks such as learning, reasoning and comprehension. During the performance of a working memory task, in which individuals indicate whether a displayed stimulus is the same or different as the stimulus displayed two trials previously, obese individuals show reduced responses in the right parietal cortex compared to healthy weight and overweight individuals [47]. Critically, this decreased response is mediated by variation in insulin sensitivity [47]. Given the dramatic rise in type II diabetes and the public health impact of dementia, this is a particularly important area for future research.

Satiety can also be signaled by gastric distension, which is thought to release satiety factors and act through vagal afferents to influence brain circuits [48]. Human imaging studies have used gastric balloon expansion techniques to measure brain response to mechanical distension of the stomach. These studies confirm that gastric distension alters brain response [48,49] and that these responses may be associated with eating behavior. For example, Wang and colleagues used the Transcent Implantable Gastric Stimulator, which generates electrical signals to induce the expansion of the fundus while brain metabolism was assessed with 2-deoy-d[¹⁸]fluoro-d-glucose with PET [50]. They found that gastric stimulation was associated with elevated response in the hippocampus, striatum, orbitofrontal cortex, cerebellum and striatum, with response in the hippocampus correlating with a self-report measure of a tendency for emotional eating.

The influence of obesity and diabetes on brain representation of food

There is now an extensive literature of studies using functional neuroimaging techniques to examine how obesity influences brain response to food. One of the most consistent findings is that there are blunted responses in the meal termination circuit. For example, obese individuals have decreased prefrontal responses at rest [51], after meal consumption [52] and attenuated postprandial deactivation of the hypothalamus [53,54]. There are also a number of reports of elevated responses to food and food cues and these responses are often interpreted as reflecting heightened incentive motivation for food [25,27,55–58]. Likewise, it has been shown that living with type 2 diabetes increases response to cues and blunts the satiety response [59,60] and that individuals with insulin resistance also show blunted brain response to insulin delivery [24].

Another consistent finding is that obese individuals have blunted responses to milkshake consumption in the dorsal striatum [23,61,62] with the magnitude of this decrease predicting weight-gain in individuals at genetic risk for obesity, as well as impulse control disorders [62]. Importantly, this differential response likely reflects a neural adaptation to adiposity or overeating since lean youth at risk for obesity display heightened, rather than decreased responses [63] and since weight gain leads to decreased responses in this area to milkshake [64]. This pattern of results highlights the dynamic relationship between brain and obesity.

Another example of this dynamic relationship comes from a recent neuroimaging investigation of the effects of diet-induced obesity on brain response in the mini-pig [65]. In this study, brain response was compared between diet-induced obese and lean mini-pigs using single photon computed tomography. Obese mini-pigs showed reduced responses in prefrontal cortex and the magnitude of this effect was correlated with body weight, suggesting that the consistently observed dampened prefrontal response in obese individuals represents a consequence rather than a cause of obesity. Response in the ventral tegmental area and striatum were also reduced while increased responses were found in the thalamus and middle temporal gyrus.

An important implication of this dynamic relationship between brain and body weight is that in order to isolate responses that represent risk for the development of obesity, it will be necessary to conduct large-scale prospective studies in initially healthy weight individuals. Another implication is that weight loss may reverse potential neural adaptations [66]. Very little is understood about the effects of weight loss on brain response or on how metabolic and hormonal factors influence these dynamic responses. However, a key point is that weight loss may also change brain response in such a way as to hinder weight-loss maintenance [67,68]. For example, Rosenbaum and colleagues have provided preliminary data showing that changes in brain response that occur in response to weight loss can be reversed by restoration of leptin levels [53]. This indicates that an important future research direction will be the investigation of how the central and peripheral consequences of obesity interact.

The influence of exposure to engineered food on brain response to food

While it is now established that adiposity and the adverse physiological consequences of obesity, such as altered glucose metabolism, can influence brain representation of food, an important new theme to emerge is that brain response to food is altered by exposure to certain types of foods. For example, the question of whether artificial sweeteners may be contributing to increased obesity has been raised and considerable evidence has been acquired to support this possibility in rodents [69–71]. In particular, Davidson and Swithers have proposed that experience with non-caloric sweet tastes can disrupt the predictive relationship between sweet taste and calories, which then leads to an erosion in the animal’s ability to regulate intake, resulting in over-consumption and eventually weight gain [70].

In support of the view that sweet taste disrupts the predictive relationship between sweet taste and calories, two neuroimaging studies have reported that artificial sweetener use is associated with altered brain response to sweet tastes [72,73].

In the first, a negative relationship between amygdala response to sucrose and self-reported artificial sweetener was identified. Since the amygdala plays a critical role in the process by which flavors are associated with their post-ingestive effects [74] the diminished response is consistent with the possibility that artificial sweetener use degrades the predictive signal of sweet taste generated in the amygdala.

In the second study, brain response to saccharin and sucrose was compared in people who did or did not consume diet sodas [72]. A number of differences were identified. Of particular interest, a negative relationship was observed between diet soda consumption and response in the dorsal striatum to saccharin. This finding is intriguing because response to food in this region is consistently blunted in obesity, decreases with weight gain and also appears to be sensitive to frequency of consumption of a high fat/sweet food (ice cream) [75].

Taken together, these results emphasize that our altered food environment affects brain representation of food. Foods that have higher or lower nutritive value than those foods typically found in nature appear to alter brain response. Given that the role of the taste is to provide information about caloric content of foods, it makes sense that reward circuits would be sensitive to variations in these associations. A fruitful area for future studies will be to examine how metabolic and hormonal signals influence this neural plasticity.

Beyond response to food and food cues

It is often convenient to employ simple paradigms to explore the influences of metabolic factors and disease states on brain representation of food. There are, however, a number of more sophisticated paradigms that have been developed to isolate reward learning, decision making, goal directed versus habitual responding, attention to and memory for foods [8,11,20,76–83]. An important future direction will be to implement some of these more sophisticated designs in special populations or in conjunction with hormonal and metabolic manipulations. Such experimental designs will allow for a more refined understanding of how metabolic processes influence complex behaviors to guide feeding.

One area in which studies have fruitfully moved beyond response to food and food cues is the assessment of brain representation of self-control or impulsivity in obesity. Hendrick and colleagues employed the Stop Signal Task, in which subjects must inhibit responding to cues as a means of comparing cognitive control, in lean and obese women [84]. No differences were observed in performance, but during trials requiring subjects to inhibit responding, there was a negative correlation between BMI and activation in areas implicated in cognitive control including the left inferior parietal cortex, supplementary motor area, bilateral insula, and bilateral cuneus. In a similar study, Batterink and colleagues used a go/no-go paradigm to test the hypothesis that overweight adolescent girls are poorer at inhibiting pre-potent responses to appetizing food stimuli as compared to normal weight controls [85]. It was found that BMI was positively correlated with impulsivity and decreased activation in frontal inhibitory regions. Furthermore, Kishinevsky and colleagues designed a longitudinal study to investigate whether brain activity during a delayed discounting task could predict weight gain in obese women [86]. Their results indicated that decreased activation in areas implicated in executive function during difficult vs. easy delayed discounting trials predicted weight gain in these subjects. Collectively, this work highlights the association between adiposity, prefrontal dysfunction, and impulsive behavior.

Summary and future directions

A central challenge for metabolic research is to determine how the reinforcing properties of food interact with peripheral hormonal and metabolic signals to influence intake. Neuroimaging investigations offer the only non-invasive tool to localize mental processes in humans. As such, human neuroimaging occupies an important niche in metabolic research by enabling the identification of complete circuits that underlie desire, pleasure, craving and choice and by allowing scientists to establish how these circuits influence, and are influenced by, metabolic and hormonal factors. The past two decades of human neuroimaging studies of feeding have laid the groundwork for the convergence of psychology, neuroscience and endocrinology. Critical questions for the future demand their full integration.

Unknown is how metabolic responses interact with neural circuits that underlie goal directed vs. habitual responding or incentive motivation and error signaling. These factors are key because they have been linked to brain dopamine systems[87–89], which has, in turn, been linked to hormonal and metabolic signaling in animal work [90–92], weight gain and obesity[24,93–95]. Also, unknown, but of potential importance given the ubiquitous presence of marketing is how cognitive influences, like beliefs and expectations, can alter metabolic responses, and if and how exposure to certain foods can impact brain representation by way of peripheral mechanisms to influence food intake. Another key gap in the literature is the missing link between peripheral and central mechanisms in driving the dynamic relationship between brain and obesity. Finally, the blood flow responses that are assessed with fMRI cannot be equated with neuropharmacological signaling even if they occur in nuclei associated with the neurotransmitter of interest. Unfortunately, techniques that do enable measurement of transmitter signaling have far poorer temporal resolution compared to blood flow techniques due to the long half-lives of the tracers. One promising avenue for future research may be the combination of PET and fMRI studies. Other interesting possibilities that have already been used with some success include examining the influence of polymorphisms associated with altered signaling of a transmitter of interest on feeding paradigms [24,96] or altering neurotransmitter levels with pharmacological manipulations such as tryptophan depletion [97] or pharmaceutical administration [98,99].

In summary, food intake is determined by a complicated interaction between mind and body and neuroimaging provides the key portal by which these interactions can be measured in humans.

Conflict of interest

None declared.