Introduction

As the average BMI increases globally, strategies to fight obesity have emerged one after another [1, 2]. Available treatments currently include behavioral modification (diet and exercise) [1, 3], psychological conditioning [4, 5], pharmacotherapy [6], as well as obesity surgery [7], gut microbiota therapy [8], and even Chinese traditional medicine or acupuncture therapy [9], all of which may be able to contribute to the success of battling the obesity epidemic. However, the effects of weight loss treatments vary significantly between individuals. In recent years, a growing amount of literature has highlighted brain processes involved in eating behavior, proposing that obese individuals exhibit brain functional abnormalities implicated in reward, attention, emotion, memory, homeostatic regulation of food intake, and executive function including inhibitory control of feeding behavior, appealing that treating obesity cannot remain “brainless” [10].

Behavioral and psychological interventions, the most common weight loss methods, often yield suboptimal results in long-term follow-up, resulting in minimal weight loss or high relapse rates [1], which could be partly attributable to the changes in brain activity [11]. Pharmacotherapies, most of which initially acts on or influence the brain [12], are also commonly used and have resulted in encouraging weight loss. However, concerns about brain function side effects have limited widespread use. Obesity surgery, the most effective treatment currently available for morbid obesity and diabetes, typically leads to 23%, 17%, 16%, and 18% changes in body weight for 2, 10, 15, and 20 years after surgery, respectively [7], with altered brain activity [13, 14]. In this narrative review, we present a brief overview of the brain regulation of eating, and then, we describe the crosstalk between brain and the three most commonly practiced weight loss treatments.

Central Nervous System Regulation of Eating

The Hypothalamus Regulation of Energy Homeostasis

Brain regulation of eating is complicated, and almost all of the neural systems are involved (Fig. 1). The hypothalamus is the primary region responsible for energy homeostasis. The regulation of this process depends on the precise orchestration of complex physiological responses such as food intake and energy expenditure by the hypothalamus. The key pathway to this regulation is the melanocortin system, which consists of two functionally antagonistic neuronal populations: one subset expresses the orexigenic neuropeptides agouti-related peptide (AgRP) and neuropeptide Y (NPY), while the other subset expresses the anorexigenic peptides proopiomelanocortin (POMC) and cocaine and amphetamine-regulated transcript (CART) [15]. External factors activated or deactivated different groups of neurons in the hypothalamus, influencing eating behaviors and energy balance. For example, circulating levels of insulin and leptin, proportionate to nutritional status and adipose tissue stores, inhibit AgRP neurons and activate POMC neurons, leading to decreased energy intake and increased energy expenditure [16,17,18]. Other key peripheral molecules impacting the hypothalamus and energy homeostasis include ghrelin [19], GLP-1 [20,21,22], adiponectin [23], irisin [24], and inflammatory factors (discussed in the later). In addition to the crosstalk between groups of neurons inside the hypothalamus, it also receives external signals and communicates directly with other brain areas such as reward, emotion, memory systems, as well as the cognitive control and other cortical areas. Thus, the dysregulation of the hypothalamus may also be subject to control/influence by higher systems in humans.

Fig. 1
figure 1

Brain regulation of eating. Brain regulation of eating involves almost all of the neural systems and is influenced by inflammation in the brain, hormones, and blood-brain barrier (BBB) functions. Strikes from the environmental or gene mutations disrupt the normal regulation of energy balance in the brain, and result in obesity. AgRP agouti-related peptide, CART cocaine and amphetamine-regulated transcript, D2 receptors dopamine 2 receptors, DLPFC dorsolateral prefrontal cortex, GLP-1 glucagon-like peptide-1, NPY neuropeptide Y, OFC orbitofrontal cortex, POMC proopiomelanocortin, pre-SMA pre-supplementary motor area, SN substantia nigra, VTA ventral tegmental area

Full size image

Higher Central Nerve System Regulation of Eating Behavior

The higher central nerve system in modification of energy balance is much more complicated than that in the hypothalamus. The regulation of eating behavior by activation or deactivation of certain regions in the brain is controversial in addition to their sophisticated cross-talks with each other. Conflicting results are reported by different research groups. Thus, we mainly discuss the mainstream ideas in the article.

The reward system, which consists of dopaminergic neurons originating in the ventral tegmental area (VTA) and substantia nigra (SN) in the midbrain project throughout the brain especially to key areas including the nucleus accumbens, striatum (especially the caudate), orbitofrontal cortex (OFC), and insula, has been extensively researched in relation to obesity in recent years. Numerous studies have indicated a respond of these areas to food cues during fMRI, especially in obese [25,26,27,28]. Some researchers have hypothesized that exposure to highly rewarding foods results in hyper-responsivity of the reward system to food cues, which leads individuals to seek foods more frequently and in greater quantity. This hyper-response theory is supported by fMRI studies in obese individuals with findings of increased activation in the nucleus accumbens, midbrain, and OFC to visual food cues [29, 30]. Other researchers are proposing a hypo-response theory that prolonged exposure to high-reward food results in lower availability of dopamine 2 (D2) receptors in the reward systems (especial the striatum), which are consistently reported in their PET studies [31, 32]. The decreased D2 receptors in reward system lead individuals to seek and consume more high-calorie foods to maintain reward and exacerbate obesity. These two seemingly paradoxical theories, the hyper-response and hypo-response theory, could in fact help to explain the occurrence and development of obesity, respectively.

The cognitive control system, sophisticated regulating of functions such as inhibitory control, food motivation, internal awareness, emotional processing, and impulse control, influences eating behavior in a very complicated and controversial way [33]. The prefrontal cortex comprises much of the cognitive control network, particularly the cingulate cortex, inferior frontal cortex, pre-supplementary motor area (pre-SMA), and dorsolateral prefrontal cortex (DLPFC). Several studies have demonstrated impaired inhibitory control in patients with obesity and a link between impaired control and future weight gain in normal-weight individuals [34,35,36]. Also, increased impulsivity has been found to be related to overweight and failing in attempts to lose weight [37].

The attention system, regulated by the parietal and visual cortices as well as some areas of the frontal cortex (e.g., anterior cingulate), has been repeatedly implicated in obesity. Obese individuals attend more to food cues under satiety and normal-weight individuals who pay more attention to food cues display patterns of overeating and weight gain [38].

The emotion system, which is primarily located at the amygdala, is well known to be a potent modulator of appetite. Joy and anger both increase appetite and create poorer dietary choices as compared to fear and sadness [39], and the effect is more pronounced in women than in men [39]. Indeed, food cues activate the amygdala [40], and activation of the amygdala could also predict the consumption of high fat or high-calorie foods [41].

Memory, primarily regulated by the hippocampus and parahippocampal formation, may also play a role in dysfunctional eating behaviors. The hippocampus and parahippocampal gyrus receive inputs regarding food cues from many other areas including the insula, orbitofrontal cortex, and arcuate nucleus of the hypothalamus [42]. It has been hypothesized that decreased functioning of the hippocampus leads to increased food intake and poorer dietary quality in turn leading to obesity [43,44,45] .

Generally, the increased activation of reward, attention, emotion, and impulsivity and motivation systems toward food, or decreased activation of the inhibitory control and impaired memory systems may play a vital role in the occurrence or exacerbation of obesity. However, no brain region acts in isolation in the regulation of eating, and there is also crosstalk with other contributors (e.g., hormones and inflammatory mediators), with intricate orchestration of complex signals through the body for regulation of eating behaviors.

Obesity Initiated by Inflammation in the Central Nerve System

Obesity commonly co-occurs with inflammation. Inflammation in the central nerve system happens earlier than in peripheral tissue and occurs preferentially in the hypothalamus since parts of it are unprotected by the blood-brain barrier (BBB) [46, 47]. It can induce hypothalamic dysfunction and lead to obesity in a two-phase process. During the early phase of inflammation, such as short exposure to a high-fat diet (HFD), the large amount of fat absorbed by the intestine causes activation of cytokines and inflammatory pathways in the hypothalamus [48]. Markers of hypothalamic inflammation increase significantly during the early days of HFD feeding, with reactive gliosis and neuronal injury manifesting during the first week, even before weight gain [46]. In parallel with the early occurrence of inflammation, 3 days of HFD feeding is sufficient to reduce hypothalamic insulin sensitivity significantly [49]. Importantly, these processes precede inflammatory events in peripheral tissues, such as the liver [47]. During the secondary inflammatory phase, prolonged inflammatory cascades lead to the activation of cellular stress mechanisms and inflammatory mediators released from non-neuronal cell types giving rise to long-lasting impaired metabolic control of the hypothalamus [50]. Proinflammatory cytokines, such as TNF-α, promote the onset of insulin and leptin resistance in the brain [51], and IKKβ/NF-κB, disrupts adult hypothalamic neural stem cells and mediates neurodegeneration [52]. Furthermore, long-term HFD feeding also alters synaptic plasticity, a change in synaptic strength in response to stimuli, in key hypothalamic neuronal systems [53].

In addition to the hypothalamus, the hippocampus is also vulnerable to inflammation despite full protection of the BBB [54, 55]. In a state of obesity, microglial action has been observed to impair hippocampal function by driving inflammation in the brain and resulting in impaired spatial recognition memory, depression and anxiety [54, 56]. Consequently, immune-mediated obesity may start with inflammation in the brain.

The BBB Dysfunction and Obesity

The BBB was first described for its ability to prevent the unregulated exchange of substances between the blood and central nerve system, including inflammatory factors described above. Over time, it has been characterized as an interface that enables regulated exchanges between the brain and peripheral cytokine and hormones. Hormones that regulate feeding are altered in obesity including insulin, leptin, adiponectin, and ghrelin that can cross the vascular BBB via specialized transport systems [57]. BBB transport is necessary for these proteins to exert their functions in the brain. Moreover, the BBB is also a secretory tissue which can secret some factors related to obesity either into the blood or the interstitial fluid of the brain [58].

Pathological changes to the BBB occur during obesity that may ultimately exacerbate disease and can lead to additional changes in the brain such as neuroinflammation and cognitive and memory impairment as described above. In addition to that, impaired BBB could result in the impaired transport of hormones [59] or the expression of other proteins at BBB [60], which leads to hormone resistance and altered cellular energy metabolism. However, the reversal of obesity can restore normal BBB functions. It has been reported that the reversal of obesity reduced FFA transport into the human brain by 17% [61]. Consequently, BBB, the mediator of peripheral communication to the brain, should not be neglected in the treatment for obesity.

The Changes of Behavioral and Psychological Interventions on Brain

Behavioral and Psychological Interventions: Conventional Weight Loss Therapy with Suboptimal Results

Behavioral and psychological interventions, the most commonly used weight loss method which modifies diet [1, 62, 63] and physical activity [1, 3, 64, 65], with or without the usage of psychosocial treatment (e.g., mindful eating, cognitive restructuring) for weight loss [4, 5, 66], also changes the brain activity. In an fMRI study of 19 postmenopausal obese women undergoing caloric restriction, Prehn et al. found improved memory score after 12-week negative energy balance for weight loss, paralleled by increased gray matter volume in the hippocampus (memory) [67].

While these interventions are simple and with low risk and economic burden, the short-term effects (weight loss) and longer-term effects (weight maintenance) of behavioral and psychological interventions are trivial, yielding minimal weight loss and high relapse rate [1, 68]. In the same study by Prehn et al., they also found that the brain restoring effects were transient only during the 12-week negative energy balance for bodyweight loss, and could not be detected after subsequent 4-week weight maintenance [67]. Moreover, subjects with more impulsivity and lower inhibition control have even worse outcomes, showing not only minimal weight loss and high relapse rates in the moment [37], but also greater risk for onset of binge eating, bulimic symptoms, and bulimia nervosa in the future [69, 70].

The Effects of Behavioral and Psychological Interventions on the Brain: an Explanation for Minimal Weight Loss

The suboptimal results of behavioral and psychological interventions could be attributable to the compensation response to the energy deficits during the interventions [71]. It resulted in the increased brain activation including regions implicated in motivation, attention, and reward valuation in response to food cues in subjects undergoing caloric restriction [72, 73], and biases brain reward systems toward high-calorie foods [74] (Fig. 2 and Table 1). By studying the fMRI of 196 dieting adolescents, Stice et al. found that fasting for hours correlated positively with activation in regions implicated in attention (anterior cingulate cortex), reward (putamen, OFC), and motivation (precentral gyrus) in response to food cues. Likewise, negative energy balance for 2 weeks (weight loss ≥ 1 kg) resulted in increased activation in attention (anterior cingulate cortex, ventral medial prefrontal cortex, superior visual cortex), and reward (caudate) regions in response to food, in addition to the restored memory system (hippocampus) [11].

Fig. 2
figure 2

The changes of behavioral and psychological interventions on brain. Behavioral and psychological interventions activates the food reward, attention, memory and motivation system, and always leads to minimal weight loss and high relapse rates. AgRP agouti-related peptide, CART cocaine and amphetamine-regulated transcript, D2 receptors dopamine 2 receptors, DLPFC dorsolateral prefrontal cortex, NPY neuropeptide Y, OFC orbitofrontal cortex, POMC proopiomelanocortin, pre-SMA pre-supplementary motor area, SN substantia nigra, VTA ventral tegmental area

Full size image
Table 1 Summary of weight-loss methods effects and regulation on CNS
Full size table

Moreover, brain activation toward food cues could also predict weight loss. By studying 25 obese individuals before and after a 12-week psychosocial weight-loss treatment and at 9-month follow up, Murdaugh et al. found that those obese individuals who were least successful in losing weight during the treatment showed greater pretreatment activation in response to high-calorie food vs. control pictures in brain regions implicated in the reward-system (nucleus accumbens and insula) and attention processes (anterior cingulate, superior occipital cortex, inferior and superior parietal lobule, and prefrontal cortex). Furthermore, less successful weight maintenance at 9-month follow-up was predicted by greater post-treatment activation in brain regions implicated in reward (insula, VTA, putamen), and attention (fusiform gyrus) [4].

However, in the Look AHEAD study with 10-year follow-up, by studying 232 patients with type 2 diabetes and overweight or obesity, McDermott et al. found a better way for weight loss, a specially designed long-term intensive lifestyle intervention program (participants were assigned calorie, fat gram, and physical activity goals designed to produce 10% weight loss). Compared with participants in regular diabetes support and education programs, patients in their lifestyle intervention program showed less activated reward system (left caudate) by high-calorie food cues, though with greater activated attention system (left angular and occipital cortex) which might diminish the weight loss effect. Still, those patients in their lifestyle intervention program showed a much more effective weight loss result than patients in regular diabetes support and education programs since the first year of follow-up. Yet, unfortunately, the differences between the two groups diminished year by year and lasted up to the 9-year follow-up [75]. Therefore, behavioral and psychological interventions programs must consider their influence on brain activity to achieve optimal results.

The Influence of Weight Loss Medicine on Brain

Pharmacotherapies: a Double-Edged Sword for Obesity

Pharmacotherapies, most of which initially act or have an influence on the brain, are also effective in weight loss options [12, 76]. By analyzing 50 publications comprising 43,443 obese individuals undergoing pharmacotherapy in a system review, Dong et al. found that the maximal mean weight loss relative to placebo for orlistat (120 mg), lorcaserin, naltrexone-bupropion, phentermine-topiramate (7.5/46 mg), and liraglutide was − 2.94, − 3.06, − 6.15, − 7.45, and − 5.5 kg, at weeks 60, 54, 67, 59, and 65, with mean rates of regain of 0.51 kg, 0.48 kg, 0.91 kg, 1.27 kg, and 0.43 kg per year, respectively [76]. The brain target of pharmacotherapies is dopamine, norepinephrine, or serotonin, the increased activation of which in the brain can stimulate hypophagia, weight loss, and in some cases, energy expenditure (Fig. 3 and Table 1). However, brain function safety concerns exist, and side effects such as suicidal ideation and depression can occur due to the broad range of targets in the brain of these medications. In the same meta-analysis, Dong et al. also described the 1-year dropout rates for orlistat, lorcaserin, naltrexone-bupropion, phentermine-topiramate, and liraglutide were as high as 29.0, 40.9, 49.1, 34.9, and 24.3%, respectively, mainly due to adverse effects [76].

Fig. 3
figure 3

The influence of weight loss medicine on the brain. Pharmacotherapy is an effective weight loss method and regulates brain activity. Nevertheless, concerns about its brain function safety problems limit its extensive use. AgRP agouti-related peptide, CART cocaine and amphetamine-regulated transcript, D2 receptors dopamine 2 receptors, DLPFC dorsolateral prefrontal cortex, GLP-1 glucagon-like peptide-1, NPY neuropeptide Y, OFC orbitofrontal cortex, POMC proopiomelanocortin, pre-SMA pre-supplementary motor area, SN substantia nigra, VTA ventral tegmental area

Full size image

Liraglutide

Liraglutide is an agonist of GLP-1 receptor, which presents in the human hypothalamus, medulla, and parietal cortex in addition to peripheral organs [77]. GLP-1 acts on the mesolimbic reward system in addition to the hypothalamus regulating homeostatic feeding. GLP-1 crosses the BBB at the area postrema and directly stimulates hypothalamic anorexigenic neurons (e.g., POMC) and deactivates the reward system involving the VTA and the nucleus accumbens [78]. Liraglutide has been reported to result in a weight loss of 4.7–6.1% during 56-week follow-up [79, 80]. By studying fMRI from 21 individuals with type 2 diabetes, Farr et al. found that liraglutide decreased the activation in the reward system (insula and putamen) as well as attention network (parietal cortex) in response to highly vs. less desirable food images. Meanwhile, participants taking liraglutide rated themselves as being fuller while fasting and trended toward feeling less pleasant to eat [77]. Nausea is the most common adverse effect, which is also related to the changes in activation in brain regions [77].

Naltrexone-Bupropion

Naltrexone-bupropion, approved for the treatment of obesity since 2014, stimulates dopamine and POMC neurons and blocks inhibitory feedback to mu-opioid receptors on POMC neurons, which results in decreased appetite and increased energy expenditure. Naltrexone-bupropion has been reported to result in 5.0–9.3% weight loss during 28–56 weeks observation [81,82,83,84]. As expected, fMRI study data from 40 obese women indicated that 4 weeks of naltrexone-bupropion treatment attenuated activation in the hypothalamus in response to food cues, as well as restored the activation of the memory system (hippocampal). Yet, researchers also found that the medicine influence the other higher central nerve systems, such as enhancing the activation of attention (anterior cingulate, superior parietal), reward (insula), and internal awareness systems (superior frontal) [85]. The wide range influence on the higher central nerve systems could enhance or diminish its effect on weight loss and cause side effects, which still need to be further investigated. Nausea is the most common adverse effect, and the FDA still carries the black box warning regarding suicidal ideation and actions.

Lorcaserin

Lorcaserin, a selective 5-hydroxytryptamine 2C receptor agonist, was approved in 2012. Evidence from plenty of studies suggests that lorcaserin has multiple psychological effects that contribute to weight loss, including elevation of satiety, reduction in craving, and impulsivity [86]. Clinical trials have reported resulting in a weight loss effect of 4.5–7.0% [87,88,89]. By studying 48 obese participants, Farr et al. found that lorcaserin exerts its weight-reducing effects by decreasing the attention (parietal and visual cortices), emotion (amygdala), and reward (insula) activity to food cues. Meanwhile, total caloric intake decreased about one quarter over 4 weeks in participants on lorcaserin. Moreover, baseline activation of the amygdala is associated with increased efficacy, suggesting that lorcaserin would be of particular benefit to emotional eaters [90]. Headache and dizziness are the most common adverse effects.

Currently, there are still plenty of centrally acting anti-obesity drugs, such as topiramate-phentermine that modulate norepinephrine release, GABA activity, voltage-gated ion channel modulation, inhibition of AMPA/kainite excitatory glutamate receptors, and inhibition of carbonic anhydrase; and phentermine, diethylpropion, benzphetamine, and phendimetrazine which are appetite-suppressant which stimulate norepinephrine release with minor dopamine release [12]. The wide range of targets in the brain, such as the energy homeostasis, reward, attention, and cognitive systems, allow pharmacotherapies with a fairly robust weight loss effect. However, the side effects, such as increased blood pressure and heart rate, insomnia, paresthesia, dry mouth, depression, anxiety, and constipation, which come from the unexpected targets of the medicines, limit their widespread use. In fact, concern about the safety of weight loss drugs has long been raised, with many weight loss drugs (such as sibutramine, rimonabant, caffeine, ephedra, and phenylpropanolamine among others) having been removed from the market [6, 12, 91].

Brain Regulation by Obesity Surgery

Obesity Surgery: the Most Effective Treatment for Obesity

Obesity surgery is currently the most effective long-term treatment for morbid obesity. The most commonly performed procedures are Roux-en-Y gastric bypass (RYGB), vertical sleeve gastrectomy (VSG), and adjustable gastric band (BAND). The mean changes in body weight for 2, 10, 15, and 20 years after surgery are 23%, 17%, 16%, and 18%, respectively [7, 14, 92]. In addition to the weight loss, our group and others have reported its ability to regulate glucolipid metabolism and insulin resistance, hormones balances (e.g., testosterone), inflammation, as well as decreasing the incidences of diabetes, myocardial infarction, stroke, cancer, and most importantly, overall mortality [7, 93,94,95,96,97]. Interestingly, patients also typically reported decreased hunger and lower caloric intake after surgery, in addition to decreased desire for core tastes and a shift in food preferences from high- to low-energy foods [98, 99]. This change in the motivation of caloric intake after obesity surgery is considered to be due to its effect on the brain, with plenty of fMRI studies showing decreased activation in the reward, attention, and motivation network and increase in the inhibition control system toward high vs. low energy food after bariatric surgeries [98, 100,101,102] (Fig. 4 and Table 1).

Fig. 4
figure 4

Brain regulation by bariatric surgery. Bariatric surgery is the most effective treatment currently available for morbid obesity and diabetes. It shows pronounced effects on brain activity, such as deactivation of regions implicated food reward, attention, motivation, memory, and emotion, and activation of regions implicated inhibition control toward food cues. AgRP agouti-related peptide, BAND adjustable gastric band, CART cocaine and amphetamine-regulated transcript, D2 receptors dopamine 2 receptors, DLPFC dorsolateral prefrontal cortex, NPY neuropeptide Y, OFC orbitofrontal cortex, POMC proopiomelanocortin, pre-SMA pre-supplementary motor area, RYGB Roux-en-Y gastric bypass, SN substantia nigra, VSG vertical sleeve gastrectomy, VTA ventral tegmental area

Full size image

Regulation of Brain Activity by Obesity Surgery

Gastric Bypass

Gastric bypass is the most efficacious obesity surgery, resulting in a weight loss of 32 ± 8% at 1–2 years and 27 ± 12% at 15 years after the surgery [7]. This excellent result on weight loss could not be achieved without its effect on the brain, resulting in decreased activation of reward and attention systems and increased activation of inhibitory control. By studying the fMRI of 16 patients 4 months after RYGB, Baboumian et al. observed decreased activation in attention systems (fusiform gyrus, inferior temporal gyrus, and right middle occipital gyrus) and increased activation in inhibition control (DLPFC, right medial prefrontal gyrus, and paracingulate) in response to high vs. low energy food cues, together with patients’ BMI decrease of 9.1 kg/m2 [13]. In addition, the authors also observed decreased activation in parahippocampal [13], region implicating in memory or reward-processing, which is still controversial. Likewise, postsurgical reduction in brain responsiveness within the attention network (superior parietal and precuneus) toward high vs. low-energy foods has also been observed by Zoon et al. in the fMRI of 19 RYGB patients, together with their reports of a shift in food preferences from high-energy foods to low-energy foods [103]. By studying the fMRI of 10 female patients 1 month post-RYGB, Ochner et al. also found significant reductions in brain activation within the reward pathway (ventral striatum, VTA, putamen, lentiform nucleus) toward high vs. low-calorie food cues, in accordance with patients’ changes in desire to eat from high-energy food to low-energy food [102]. However, the author also reported reduced activity in inhibitory control (DLPFC) and other prefrontal cognitive systems (ventrolateral prefrontal cortex and dorsomedial prefrontal cortex) toward high vs. low-calorie food cues in these patients [102], which was contradictory to the results in other studies [13]. The authors believe that the activity changes in the prefrontal cognitive system could imply reduced reward-processing which was also a vital function of the cognitive system [102]. Thus, further studies are needed to advance our understanding of the relationship between the prefrontal cortex and obesity surgery. Still, there were plenty of similar findings reported in other studies [104]. In fact, Frank et al. reported an fMRI study that even showed a restoration by RYGB in patterns of brain activation to food stimuli similar to those observed in normal-weight individuals after an average of 3.4 ± 0.8 years following RYGB, though patients’ self-rated disinhibition and hunger were only partly restored [105].

VSG

VSG is currently considered as the optimal obesity surgery for most patients with morbid obese. Although the weight loss (25 ± 9% in 1–2 years and 18 ± 11% in 15 years post-surgery) percentage is a little lower than RYGB [7], the non-weight loss outcomes are controversial between the two procedures [106, 107], and serious complications are least in VSG [107]. By studying the fMRI of 9 patients 4 months after VSG, Baboumian et al. found that similar to RYGB though less robustly, VSG also exert its weight loss effect by increased inhibition (DLPFC) and decreased attention (fusiform gyrus) and reward or memory (parahippocampus) activation in response to high- vs. low-energy food cues [13]. Moreover, by studying 22 patients 1 month after VSG, Li et al. reported that VSG also significantly decreased brain activation in the food motivation (right DLPFC) in response to high-caloric vs. low-caloric food cues. Accordingly, the patients also rated a significant reduction of craving for high-calorie food. The decrease in right DLPFC activation in response to high-caloric vs. low-caloric food cues after surgery was positively correlated with the reduction in craving for high-caloric vs. low-caloric food cues [108].

BAND

BAND has been reported with a weight loss of 20 ± 10% in 1–2 years and 13 ± 14% in 15 years after surgery. Changes in brain activation toward food cues have also been found in cognitive and reward systems. Bruce et al. found significantly less self-reported hunger, disinhibition, and increased cognitive restraint rated by ten patients 12 weeks after BAND surgery, accompanied by decreased brain activation to food vs. nonfood pictures in regions implicated in food motivation (medial prefrontal cortex, inferior frontal gyrus) and reward and memory (insula, parahippocampus), and increased activation in cognitive control and inhibition (anterior prefrontal cortex). Moreover, correlation analysis indicated that less activation to food vs. non-food cues at post-meal in the right inferior frontal gyrus was associated with self-rated increased cognitive restraint and reduced hunger from before to after surgery, and less activation to food vs. non-food cues at post-meal in the right middle frontal gyrus was associated with larger reduction in self-rated disinhibited eating from before to after surgery [109].

Different Procedures with Varied Effects on Brain Activity

Although all kinds of bariatric surgeries exert weight loss and affect the brain, brain activity changes vary depending on the different procedures. Scholtz et al. compared the two most common bariatric surgical procedures, RYGB and BAND, and found that RYGB caused a greater reduction in the activation of brain reward areas to high-energy food pictures compared with BAND. Accordingly, high-calorie foods were rated as less appealing and percentage energy intake derived from fat was lower in patients after RYGB than after BAND. Also, patients after RYGB has shown healthier eating behavior and less eating disorder psychopathology compared with the BAND [104]. By comparing fMRI data from patients after RYGB and VSG, Baboumian et al. reported more pronounced changes in brain activity in regions implicated inhibitory control (DLPFC) after RYGB [13]. Thus, RYGB appears to be more effective than the other two procedures in modulating brain activity such as reward and inhibitory control systems.

In addition to variation among the different procedures, different individuals also respond differently with varying effects on the brain and eating behavior. Compare with individuals without weight regain, those with weight regain post-surgery have shown a 2.2-fold higher rate of eating psychopathology [110]. Consistently, patients who are less successful at losing weight after surgery have shown to have a lower increase in activation of the areas involved in inhibition but no significant change in the reward areas compared with their more successful weight loss counterparts [111]. Thus, brain activity is a pivotal factor in the prediction of weight loss.

Bariatric Regulation of Resting-State Functional Connectivity of Brain Regions

Aside from its effects on the brain activation, obesity surgery also modifies resting-state functional connectivity of different brain regions [112, 113]. By studying resting-state fMRI of 17 patients after VSG, Li et al. found that the obesity surgery could recover the dysfunction of some brain regions, e.g., a decreased resting-state activities and increased functional connectivity in reward processing and cognitive control regions (orbitofrontal cortex, middle frontal gyrus, superior frontal gyrus, and gyrus rectus), which were associated with BMI in the correlation analysis [112]. Also, using the functional connectivity density mapping in 22 obese participants 1 month after VSG, Li et al. found that significantly reduced functional connection in cortical regions implicated in self-referential processing and interoceptive awareness along with strengthening of connectivity of these regions with cortical (DLPFC) and striatal (caudate) regions implicated in executive control/self-regulation in them [113].

The Potential Mechanism by Which Obesity Surgery Affects the Brain Activity

Correlation Between Changes of Hormone and Brain Activity After Obesity Surgery

There is still controversy regarding the changes in hormones after surgery among different procedures, which could be related to changes in brain activity. Li et al. reported that decreased brain activation after VSG was positively correlated with a reduction in ghrelin levels but not insulin and leptin levels [108]. However, Zoon et al. found no change in ghrelin levels after RYGB though an alteration in brain activation was also observed [103]. RYGB has been reported to have a stronger correlation with postprandial plasma PYY and GLP-1 [104, 114,115,116]. Hormone changes after BAND are less clear than the other two surgeries [104]. Therefore, more studies are needed to explore the regulatory mechanisms of obesity surgery on brain activity, especially its connection with hormone levels.

Obesity Surgery Modifies Inflammation in the Brain

Obesity surgery also modifies inflammation in the brain in patients with obesity, which partially accounts for its outstanding weight loss effect. Van de Sande-Lee et al. have studied 13 patients after RYGB surgery and found that the obesity surgery increased cerebrospinal fluid (CSF) concentrations of interleukin (IL)-10 and IL-6 levels, accompanied by a partial reversal of hypothalamic dysfunction on fMRI, massive loss of body mass, and dramatically decreased caloric and saturated fats intake [117]. Consistently, by studying the changes of myo-inositol concentration (a putative marker of neuroinflammation) in 23 patients with morbidly obese and intra-gastric balloon surgery during the 3-month follow-up, Gazdzinski et al. found that the obesity surgery suppressed brain inflammatory responses consistent with weight loss, preceding the remission of metabolic abnormalities [118]. Thus, regulation of inflammation seems to be another contributor to the restoration of brain activity after the obesity surgery.

Obesity Surgery Changes the Expression of Receptors in the Brain

Additionally, obesity surgery changes the expression of brain receptors. By analyzing the PET images of 5 women before and 7 weeks after RYGB or VSG, Dunn et al. found dopamine type 2 (DA D2) receptor availability decreased after obesity surgery. Regional decreases were caudate 10 ± 3%, putamen 9 ± 4%, ventral striatum 8 ± 4%, substantia nigra 10 ± 2% (all four regions from reward systems), amygdala 9 ± 3% (emotion system), hypothalamus 9 ± 3% (energy homeostasis), and medial thalamus 8 ± 2%, together with patients’ decreased rate of depression and binge eating [119]. The above changes in the brain receptors could also explain the changes of brain activity after the obesity surgery to some extent.

Brain Regulation Comparing Obesity Surgery and Behavioral and Psychological Interventions

From the discussion above, it is clear that the changes in the activation of brain regions after obesity surgery are completely opposite compared with that after behavioral and psychological interventions. Obesity surgery always leads to the deactivation of food motivation, reward and attention networks with patients’ reporting of decreased craving for food and shift preference from high to low energy food. Conversely, these brain regions are usually activated after behavioral and psychological interventions, and patients are at greater risk for future binge eating. Baboumian et al. have compared the fMRI of 25 patients 3–4 months after obesity surgery and 14 patients 3–4 months after weight loss behavior intervention and found totally opposite changes in brain activity between these two groups of patients. Patients after surgery showed increased activity in inhibitory control (DLPFC) and decreased activity in memory and attention (left parahippocampal gyrus and fusiform) in response to high-energy diet vs. low-energy diet. While patients after behavior intervention showed decreased activity in inhibitory control and increased activity in memory and attention [13]. Another study by Bruce et al. compared the fMRI data from 16 behavioral dieters and 15 patients after obesity surgery with similar weight loss of about 10% and found that the behavioral dieters showed increased responses to food cues in the medial prefrontal cortex (attention system) when compared to patients with obesity surgery [120]. The differences in brain activation may help to explain why weight loss diets typically do not produce pronounced and sustainable weight loss as obesity surgery.

Conclusion

In conclusion, brain regulation of eating is a complicated process that involves brain regions implicated in homeostatic regulation, reward, emotion, memory, attention, and executive function including inhibitory control of feeding behavior and food motivation. The effects and safety of weight loss methods are greatly dependent on brain regulation. Although behavioral and psychological interventions are pervasive and easy to implement, they have shown paltry weight loss effect both in short- and long-term observation. The activation of the food reward, attention, and motivation systems may be responsible for the minimal weight loss and high relapse rate. Pharmacological therapy, especially drug acting on the brain, is also an effective weight loss method. However, its safety concerns related to the brain, such as depression, suicide, and nausea, have limited its widespread implementation. Obesity surgery, which shows an excellent weight loss effect, maybe efficacious due to its effect on brain regulation and deactivation in reward, attention, and motivation as well as activation in the inhibitory control network.

Nevertheless, it still would be worth highlighting that conclusions based on fMRI studies are limited by the small sample size of studies (most of which included only dozens of patients) in addition to some controversial results in these studies. Moreover, correlations between weight loss and changes in fMRI do not necessarily prove that these methods lead to changes in brain activity that lead to weight loss. More larger clinical studies and mechanism researches are needed to further prove the relation between brain regulation and weight loss. Still, targeting the brain to improve efficacy and avoid side effect is probably a vital way forward toward achieving sustainable weight loss.