Keywords

Addictions and some eating disorders (ED) share multiple characteristics (e.g., diminished control, elevated negative affect), but there are also substantial differences in proposed etiologies between the two classes of disorders. Further, there is significant variability within both ED (e.g., anorexia nervosa [AN] vs. binge-eating disorder [BED]) and addictive disorders (e.g., pathological gambling vs. opiate dependence), which increases the complexity of comparison. In the following book chapter, we will focus on the neuroimaging literature as a tool to potentially identify areas of overlap and distinction between eating and addictive disorders by examining neuroimaging studies that focus on functioning theoretically related to both groups of disorders: executive control, reward and motivation, emotional reactivity, memory/learning, and interoceptive awareness (see Table 4.1). Next, we will outline components that may be more unique to either disorder. Finally, we will outline areas of future research needed to further clarify the relationship between addictive and eating concerns in the hope that a greater understanding of the relationships between these disorders may lead to increased knowledge of etiologies and the development of novel, efficacious, and well-tolerated treatments.

Table 4.1 Similarities and dissimilarities between addiction and eating disorder neuroimaging literature
Full size table

1 Defining Eating and Addictive Disorders

The fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) has undergone significant revisions based on scientific inquiry into the etiology and treatment of psychological disorders, including new definitions of addictive and eating disorders (American Psychiatric Association, 2013). Importantly, DSM-5 includes a reorganization of addiction criteria, the classification of gambling and substance use disorders (SUD) together, and the inclusion of BED as a formal diagnostic entity. The reclassification of pathological gambling as a “behavioral addiction” reflects extensive work that uncovered similarities between pathological gambling and SUD (Leeman & Potenza, 2012; Potenza, 2006). Studies also suggest commonalities between addictions (both substance and “behavioral”) and certain ED, particularly BED (Gearhardt, White, & Potenza, 2011). Despite growing evidence of the potential clinical and social implications of a “food addiction” diagnostic category, there have been no food-related additions to the addiction classification of DSM-5 (Gearhardt et al., 2012), and the topic of food addiction remains debated (Avena, Gearhardt, Gold, Wang, & Potenza, 2012; Ziauddeen, Farooqi, & Fletcher, 2012).

Importantly, the inclusion of BED as a formal ED diagnosis allows for an extended investigation of its psychopathology and comparison to other ED. A transdiagnostic model of ED suggests that a core psychopathology is overevaluation of weight and shape concerns, induced by dieting and maintained by mood dysregulation (Fairburn, Cooper, Shafran, & Wilson, 2008). Both bulimia nervosa (BN) and BED are characterized by recurrent binge eating (eating a larger amount of food than intended during a discrete period of time) and a concurrent experience of a lack of control over eating. BN is differentiated from BED through recurrent inappropriate compensatory behavior in order to prevent weight gain and overevaluation of shape and weight concerns. In contrast, anorexia nervosa (AN) is defined by the restriction of food intake leading to a body weight less than normal, intense fear of (or behavior to avoid) gaining weight or becoming fat, and disturbance in the perception or evaluation of body weight or shape.

Elevated rates of comorbidity may suggest a core etiology involved in the development of eating and addictive disorders. BN appears to have a higher comorbidity with SUD than AN, which may be related to heightened impulsivity (Holderness, Brooks-Gunn, & Warren, 1994). Binging groups (BN and BED) are more likely to be high in impulsivity (demonstrating both high reward sensitivity and low inhibition) than AN groups, which are more likely to be high in obsessionality, perfectionism, and rigidity (Vitousek & Manke, 1994). Reward sensitivity has also been implicated in comorbid binge eating and substance abuse in women (Dawe & Loxton, 2004). These findings suggest that considering binge eating as an addictive process may be a fruitful way to understand failed regulation of reward processing across disorders.

Recently, the National Institute of Mental Health (NIMH) launched the Research Domain Criteria project (RDoC) to consider new ways of classifying psychopathology based on dimensions of observable behavior and neurobiological measures (Phillips, 2011). The intent is to translate rapid progress in research into an improved integrative understanding of psychopathology and the development of better treatments for mental disorders. The RDoC approach includes dimensional axes based on potential domains of impairment, including positive valence, negative valence, cognitive systems, social processes, and regulatory processes. In the case of addictive disorders and eating pathology, there are striking similarities between BN/BED/SUD in failed regulation of reward-related processes that may not appear in other ED (e.g., AN). This dimensional approach suggests that there may be similarities between binge eating and substance use that may extend traditional categorical models of psychopathology.

2 Executive Control

2.1 Eating Disorders

Neuroimaging studies of executive control (which is conceptualized as top-down processes implicated in self-control, goal-direction, and inhibition) in disordered eating have identified differences among AN and BN/BED participants. AN participants, relative to healthy control subjects, exhibit increased activation in regions implicated in executive control (e. g., the prefrontal cortex (PFC) and anterior cingulate cortex (ACC)) when exposed to food stimuli (Uher et al., 2003). During tasks of inhibitory control, participants with AN (restricting type) showed increased activation in the medial PFC (Oberndorfer, Kaye, Simmons, Strigo, & Matthews, 2011) and the posterior visual and inferior parietal cortical regions (Lock, Garrett, Beenhakker, & Reiss, 2011), which was interpreted as reflecting elevated self-control. Zastrow et al. (2009) found that AN patients (compared to healthy controls) were less able to switch behavioral response styles when changes were needed, and AN participants exhibited hypoactivation in the ventral anterior cingulate-striato-thalamic loop associated with motivation. In contrast, AN participants had hyperactivation in neural regions associated with executive control (e.g., frontal cortex) during task performance. Thus, individuals with AN might not only have greater cognitive control, but may use this executive control rigidly.

Unlike the elevated activation of control-related regions observed in individuals with AN, different activation patterns have been observed in BN and BED participants. BN patients displayed less activation in the PFC (Joos et al., 2011; Uher et al., 2004), ACC, temporal lobe (Joos et al., 2011), and lateral orbitofrontal cortex (OFC) relative to healthy control subjects when viewing food pictures (Uher et al., 2004). Patients with BN also exhibited hypoactivation in the PFC, inferior frontal gyrus (IFG), lenticular and caudate nuclei, and ACC during an executive-control task (Marsh et al., 2009, 2011). Lock et al. (2011) found that participants with binge-purge behavior (including both AN and BN diagnoses) had greater activation in the dorsolateral PFC compared to healthy controls despite performing similarly on an executive-control task. These findings were interpreted as possibly reflecting less efficient recruitment of executive-control regions in the binge-purge group. Similar to BN participants, obese women with BED exhibited hypoactivation in the IFG, PFC, and insula relative to obese women without BED and healthy control subjects during a cognitive-control task (Balodis, Molina, et al., 2013). Further, higher levels of attempted dietary restraint in BED were related to reduced activation in the IFG and insula (Balodis, Molina, et al., 2013). Participants with BED also exhibited hypoactivation in the PFC and insula during the processing of reward and loss (Balodis, Kober, et al., 2013). Finally, BED participants have demonstrated less activation in the lateral OFC relative to normal-weight control subjects when passively viewing food pictures (Weygandt, Schaefer, Schienle, & Haynes, 2012). These findings raise the possibility that the diminished control over food consumption associated with BN/BED may relate to neural differences in regions implicated in executive control and reward processing.

2.2 Addictive Disorders

Deficits in executive control associated with BN and BED are also implicated in addictive disorders. Structural abnormalities in drug users are related to self-control deficits. For example, decreased activation in prefrontal regions is related to reduced availability of striatal dopamine D2/D3 receptors in participants with addiction and decreases in baseline glucose metabolism in the OFC, ACC, and DLPFC (Volkow, Fowler, Wang, Swanson, & Telang, 2007; Volkow et al., 1997). Obesity is also related to reduced dopamine D2/D3 receptor availability (Wang et al., 2001) and genetic alleles that may be associated with reduced dopamine signaling (Stice, Spoor, Bohon, & Small, 2008). Further, animal models suggest that overconsumption of highly palatable foods may also reduce striatal D2 receptor levels (Johnson & Kenny, 2010), which may contribute to executive-control deficits associated with BN/BED. Yet, it is still unclear whether reduced D2/D3 receptor availability and hypofunctioning in control-related neural regions precede the development of addictive disorders/obesity or result from excessive substance/food consumption (or both).

Frontal cortical impairments in the OFC, ACC, and DLPFC have been described as a core feature of impulsivity and compulsivity in addiction (Fineberg et al., 2009; Goldstein & Volkow, 2011; Potenza, Sofuoglu, Carroll, & Rounsaville, 2011). Drug/alcohol users also have poorer performance on behavioral tasks reflecting impulse control, such as delay discounting and Go/No-Go (Bickel et al., 2007; Fu et al., 2008), which may relate to activation differences in the ACC and PFC observed in addicted individuals (Brewer, Worhunsky, Carroll, Rounsaville, & Potenza, 2008; Goldstein & Volkow, 2011; Leland, Arce, Miller, & Paulus, 2008). Thus, addiction-related executive-control findings may more closely relate to the undercontrol associated with BED/BN than the overcontrol associated with AN (see Table 4.1).

3 Motivation and Reward

3.1 Eating Disorders

Disordered eating is associated with motivation to seek out or avoid food and differences in hedonic reward in response to food consumption. As a central hallmark of AN is the avoidance of food consumption, one might predict that AN would be related to reduced activation in neural regions associated with motivation in response to food cues, yet the neuroimaging literature is not consistent. AN participants exhibit less activation in the occipital cortex to food pictures (Santel, Baving, Krauel, Münte, & Rotte, 2006; Uher et al., 2003, 2004), which may be related to a diminished salience of food in this disorder. Further, Holsen et al. (2012) found that participants with AN (active and recovered) relative to control subjects had hypoactivation in the in hypothalamus, amygdala, and anterior insula in response to high-calorie food images prior to eating, which may reflect diminished motivation for food. However, others have found that AN and BN patients exhibit elevated activation in the medial OFC and ACC in response to food pictures, which could be interpreted as either elevated craving or a general increase in emotional response to food (Uher et al., 2004). Functional connectivity analyses may shed some light on how activation in the same region may relate to different responses. Kim, Ku, Lee, Lee, and Jung (2012) found AN and BN participants had greater activation in the left anterior insula (part of the primary gustatory cortex) in response to food relative to nonfood pictures, but the connectivity between these disorders differed. Specifically, insula activation in AN participants was functionally connected to regions implicated in control (e.g., IFG), whereas in BN it was associated with a reward region (e.g., medial OFC) (Kim et al., 2012). Thus, gustatory cues may trigger reward responses in participants with BN, but signal control responses in patients with AN. Finally, Cowdrey, Park, Harmer, and McCabe (2011) found that individuals recovered from AN relative to control subjects had not only increased activation in the ventral striatum and occipital cortex to pleasant food stimuli but also greater activation in the insula, putamen, ACC, and caudate to aversive food stimuli. This pattern of results may reflect greater incentive salience of food (whether positive or negative in valence) for individuals with AN (Cowdrey et al., 2011).

Additional neuroimaging studies of BN have suggested elevated motivation and reward response to foods. Brooks et al. (2011) found that BN relative to AN participants had increased activation in reward and somatosensory regions (e.g., caudate, insula) in response to food images. Relative to control subjects, participants with BN had greater activation in the ACC (Schienle, Schäfer, Hermann, & Vaitl, 2009; Weygandt et al., 2012), insula, and ventral striatum (Weygandt et al., 2012) to food cues. Further, greater negative affect for BN participants is associated with elevated activation in the putamen, caudate, and pallidum while anticipating palatable food consumption (Bohon & Stice, 2011). Thus, negative emotional states may increase the motivational properties of foods for participants with BN. Both BN and BED patients have been found to have increased gray matter in the medial OFC (Schäfer, Vaitl, & Schienle, 2010) and increased activation in this region in response to food cues (Schienle et al., 2009), but there have also been differences identified between these two disorders. BN participants appear to have greater activation in the ACC, insula, and ventral striatum in response to food cues than do BED patients (Weygandt et al., 2012). Compared to control subjects, BED participants had elevated responses in the medial OFC (Weygandt et al., 2012) and increased dopamine release in the caudate and putamen when exposed to food stimuli (Wang et al., 2011), which might reflect increased motivation. Yet, BED relative to control participants has shown less activation in certain areas in response to food cues (e.g., ACC, ventral striatum) (Weygandt et al., 2012). Although the meaning of these results is not entirely clear, animal models of food consumption suggest that excess palatable food consumption may diminish functioning in reward-related neurocircuitry (Johnson & Kenny, 2010). Similarly, BED participants have been found to have less gray matter volume in the lateral OFC, medial OFC, and striatum (Schäfer et al., 2010). Thus, BED may be related to hypoactivation in certain reward regions due to overconsumption of food.

There are also differences among the ED during food receipt. AN is associated with a tendency to exhibit less activation in the ventral/dorsal striatum, insula, and medial OFC to taste (Vocks, Herpertz, Rosenberger, Senf, & Gizewski, 2011; Wagner et al., 2007), and (unlike in healthy control subjects) insula activation does not seem to differ based on the pleasantness rating. Thus, AN participants may be less sensitive to the hedonic nature of food. Although BN has been related to less activation in the ACC and cuneus in response to glucose (Frank et al., 2006), BN (relative to control and AN participants) has been associated with greater anterior ventral striatum activation to cream mixture. Thus, the association of BN with food-related reward activation is less clear. To our knowledge, no studies to date have examined the neural response to food receipt in BED participants.

3.2 Addictive Disorders

Motivation and reward are also key constructs in addictive behaviors. Motivation in addiction is related to dopaminergic response, especially in the NAc, ACC, OFC, DLPFC, amygdala, striatum, and ventral pallidum (Salamone, Correa, Farrar, & Mingote, 2007). Dopaminergic dysregulation is associated with drug-seeking behavior (Volkow & Li, 2005), and cues become powerful triggers of dopaminergic release and motivation in addicted individuals (Robinson & Berridge, 2001). Addicted participants exhibit increased craving and activity in many brain regions (e.g., ventral striatum, ACC, amygdala) during exposure to drug cues in fMRI studies (Chase, Eickhoff, Laird, & Hogarth, 2011; Shiffman et al., 2013). In contrast to the hyperactivation associated with drug cues, substance-dependent individuals versus healthy control subjects typically exhibit blunted dopaminergic release during actual drug consumption and report weaker hedonic responses (Martinez et al., 2007; Volkow, Wang et al., 2007; Volkow et al., 1997). It is still unclear whether the reduced neural response to consumption in participants with addiction is a preexisting risk factor related to a reward deficiency (Comings & Blum, 2000), a result of excessive substance use (Kalivas & O’Brien, 2007), or related to other factors. Yet, increased motivation in response to cues and a reduced response in reward-related regions during consumption appear to be important components of addictive behaviors. In comparison to disordered eating, the increased response in motivation regions to food cues in BED and BN are similar to the addiction literature, but the literature is not always consistent. The relative dearth of studies examining food receipt in ED may contribute to this inconsistency. Further examination of food reward is an important area of future study.

4 Memory and Learning

4.1 Eating Disorders

Differences in learning and memory may also contribute to ED. Participants with active AN (who are markedly underweight) appear to have deficits in cognitive processing, which may be related to malnourishment (Delvenne et al., 1997). Neuroimaging studies of memory in ill AN patients found that despite normal performance levels on a working memory task, AN participants compared to healthy control subjects exhibited greater activation in the temporal and parietal lobes (Castro-Fornieles et al., 2010). Further, underweight status is related to hypometabolism of glucose in these same regions (Delvenne et al., 1995). Following treatment and weight gain, these differences in neural response were no longer present for AN participants, which suggests the worse nutritional status may be related to less efficient memory-related neural processing (Castro-Fornieles et al., 2010).

Differences in reward learning exist in both AN and BN individuals. Wagner et al. (2007, 2010) found that both AN and BN participants relative to healthy controls had abnormal anterior ventral striatum response to wins and losses. More specifically activation in this reward-related region did not differ whether the participants had won or lost money. This pattern of findings may be related to difficulty discriminating between the valence (positive/negative) of salient stimuli (Wagner et al., 2007, 2010). Further, BN participants appear to have hypoactivation in the insula, ventral putamen, amygdala, and OFC relative to healthy controls during a food-learning task (Frank, Reynolds, Shott, & O’Reilly, 2011). Finally, obese BED participants relative to obese participants without BED exhibited during anticipatory reward/loss processing diminished activation bilaterally in the ventral striatum, as well as in the midbrain, thalamus, and amygdala (Balodis, Kober, et al., 2013). Thus, patterns of reduced neural activation during reward learning may relate importantly to disordered eating.

4.2 Addictive Disorders

Addiction is also related to differences in memory and learning. Substances of abuse appear to decrease activity in neural areas associated with short-term memory and attention (Lundqvist, 2005). Substance-related problems are associated with differential neural response to working memory tasks. For example, cannabis users compared to healthy control subjects perform the same on a task of working memory, but parietal cortical activation differs (Jager, Kahn, Van Den Brink, Van Ree, & Ramsey, 2006), and this pattern of results persists following 1 month of abstinence (Schweinsburg et al., 2008).Though opioid-dependent individuals have performed similarly to healthy controls on working memory tasks, they have concurrently shown heightened frontal, parietal, and cerebellar activation, which may be related to compensatory recruitment of frontal control regions and impaired working memory (Bach et al., 2012; Marvel, Faulkner, Strain, Mintzer, & Desmond, 2012; Yucel et al., 2007). Pretreatment deactivation in the thalamus during a working memory task predicts poorer addiction treatment response (Moeller et al., 2010). Further, chronic cocaine dependence is related to difficulties with adaptive learning that is related to increased connectivity between the ACC networks associated with mental processing (Camchong et al., 2011). Thus, like disordered eating, addictions appear associated with impairments in working memory and learning.

5 Emotional Reactivity

5.1 Eating Disorders

Increased negative affect has been related to the presence of disordered eating. AN is related to elevated anxiety and stress responsivity (Kaye et al., 2013). Kaye et al. (2013) suggest that AN is related to excess serotonin (related to harm prediction) and reduced dopamine (related to reward prediction), which may lead to increased tendencies towards aversive stimuli and over control. Elevated cortical serotonin 5-HT1A receptor binding is associated with AN (Galusca et al., 2008), and restriction of eating may rebalance serotonin and modulate negative affect or control anxiety for these patients (Kaye et al., 2013). Further, AN participants may find dopaminergic release anxiety provoking rather than hedonically pleasing (Kaye et al., 2013). Increased dorsal caudate/putamen dopamine D2/D3 receptor binding is associated with harm avoidance in AN (Bailer et al., 2007, 2012; Kaye et al., 2013), and dorsal caudate activation in response to negative and positive feedback is associated with trait anxiety in recovered AN (Wagner et al., 2007). Further research suggests that AN is associated with increased activation of neural regions implicated in emotional reactivity when exposed to food. AN patients relative to healthy control subjects had greater activation in the right amygdala when viewing food images (Joos et al., 2011) and greater activation in the amygdala when drinking a milk shake in a hungry state (Vocks et al., 2011). Thus, AN patients may have more intense negative emotional responses when exposed to food.

Binge-eating behavior in the context of ED is also associated with differences in emotion-related neural regions. Binge and purging behavior (in the context of both AN and BN) is related to elevated hypothalamus activation during a Go/No-Go task, which may reflect aberrant emotional responding to the need to inhibit responses (Lock et al., 2011). While some research has not found any differences between BN and healthy control subjects in response to affectively valenced stimuli (Schienle et al., 2004), other studies have found patterns of activation in BN that suggest an avoidant response to emotions. For example, BN patients relative to healthy controls had decreased neural response in the precuneus to facial expressions of anger and disgust and decreased amygdala response to angry faces, which may be related to emotional avoidance (Ashworth et al., 2011). Pringle, Ashworth, Harmer, Norbury, and Cooper (2011) identified that patients with BN versus healthy control subjects displayed hypoactivation in parietal, occipital, and limbic areas (including the amygdala) when responding to negative self-referential personality words (e.g., evil). This pattern of hypoactivation might reflect emotional blunting or habituation to negative self-thoughts in BN patients (Pringle et al., 2011). Little research has examined emotional reactivity in the context of BED, although this is an important area of future research.

5.2 Addictive Disorders

Increased difficulty with emotion regulation is also associated with addictive disorders. Li and Sinha (2008) suggest that deficiencies in prefrontal regions associated with SUD contribute to difficulties with emotion regulation (in addition to executive control). Substance dependence is related to decreased reward responsiveness to pleasant stimuli that are not drug related, which could be related to anhedonia (Volkow, Fowler, & Wang, 2002; Zijlstra, Veltman, Booij, van den Brink, & Franken, 2009). Emotional circuitry may also be altered in addiction. For example, chronic cannabis use is related to reduced activation in the ACC and the amygdala to masked affective faces (Gruber, Rogowska, & Yurgelun-Todd, 2009), and addicted individuals exhibit reduced activation in the amygdala to affective pictures (negative and positive) (Wang et al., 2010). Thus, addiction may be related to reduced emotional response to nondrug stimuli. This pattern of results appears similar to the general pattern of hyporesponsivity to emotional stimuli in BN. BN and SUD may be similarly related to the use of substances (drugs, food) as a way to regulate emotions, whereas AN appears to be linked with caloric restriction to manage mood states.

6 Interoceptive Awareness

6.1 Eating Disorders

Interoceptive awareness is defined as sensitivity to physiological stimuli originating from the body (Craig, 2002). The anterior insula (a key neural region involved in interoceptive awareness) has been found to be less active for AN participants when thinking about eating food, which suggests less interoceptive awareness (Brooks et al., 2012). In contrast, Gizewski et al. (2010) found that AN compared to control subjects exhibited greater activation in the anterior insula during exposure to high-calorie food pictures, which was interpreted as recall of previously negative eating experiences or elevated emotional arousal. Unlike healthy control subjects, AN participants did not exhibit activation in the insula when tasting sucrose, and subjective ratings of pleasantness were not associated with changes in insula activation (Wagner et al., 2007). In response to nonfood stimuli, AN relative to control participants had aberrant functioning in the anterior insula and dorsolateral PFC to pain (Strigo et al., 2013). Thus, AN individuals may exhibit difficulty in appropriately perceiving bodily signals, which may allow for greater ability to ignore signs of hunger (Strigo et al., 2013). Less research on interoceptive awareness has been conducted with binge-type ED, especially BED. Schienle et al. (2009) did find that participants with BN had greater activation in the insula to food cues, which suggests the disorder may be related to greater interoceptive awareness, although alternate explanations exist (e.g., greater emotional reactivity to food cues).

6.2 Addictive Disorders

Insula function has also been implicated in substance addictions. Insula activation has been related to craving for substances of abuse (Bonson et al., 2002; Wang et al., 2007). Elevated insula activation in response to cigarette cues predicts smoking relapse (Janes et al., 2010), and damage to the insula is associated with decreased craving and markedly increased success in abstaining from smoking (Naqvi & Bechara, 2009; Naqvi, Rudrauf, Damasio, & Bechara, 2007). Thus, differences in insula activation in AN may be related to reduced craving and increased ability to abstain from eating. In contrast, the limited literature in BN suggests that elevated insula response to food cues may be similar to hyperactivity of the insula during substance-related craving in addiction. Future research that examines neural responses to food craving in the context of ED is important.

7 Neuroimaging Studies of Comorbid Eating and Addictive Disorders

In addition to comparing neural response between the disorders, neuroimaging studies of patients with comorbid eating and SUD would be helpful in identifying overlapping/differing circuitry. For example, if patients with comorbid addiction and ED displayed similar patterns of neural activation to food and drug cues, this may provide evidence of shared underpinnings. Unfortunately, there is limited research on this topic. One neuroimaging study of participants with remitted bipolar disorder did examine the association of addiction and disordered eating-spectrum scores with brain activation during exposure to affective faces (Hassel et al., 2009). Substance use severity was related to reduced activation in the right PFC to happy faces and the right caudate to neutral faces, whereas elevated disordered eating was linked to elevated right ventral putamen activation to happy and neutral faces (Hassel et al., 2009). Yet, very few participants in the study had concurrent eating and addictive disorders. Further research on comorbidities will be important in understanding the relationship between disordered eating and SUD.

8 “Food Addiction”

“Food addiction” may be another factor potentially linking addiction and problematic patterns of eating. Recent studies have suggested that certain types of foods may be capable of triggering an addictive process in vulnerable individuals (Avena, Rada, & Hoebel, 2008; Johnson & Kenny, 2010). In a sample of young women who did not meet criteria for any ED, greater endorsement of addictive-like eating (e.g., tolerance, withdrawal, continued use despite consequences) was related to increased activation in the ACC, medial OFC, amygdala, DLPFC, and caudate in response to anticipated food receipt, but lower activation in the lateral OFC to food receipt (Gearhardt, Yokum, et al., 2011). This activation is similar to patterns of neural response associated with SUD, namely elevated reward/motivation-related activation to cues and diminished control-related activation during consumption. Thus, an addictive-like response to food may occur outside of the context of traditional ED (e.g., BED).

Clinical research found that approximately half of patients seeking treatment for BED met the threshold for “food addiction,” which was associated with more frequent binge eating, greater emotion dysregulation, and elevated pathology (Gearhardt, White, Masheb, & Grilo, 2013; Gearhardt et al., 2012). Future research examining the neural correlates of “food addiction” in ED populations would be useful in evaluating the possible contribution of an addictive process to disordered eating.

9 Important Differences Between Eating and Addictive Disorders

There are also important factors that are unique to eating and addictive disorders. For example, body and shape concerns are hypothesized to be important factors in causing and maintaining disordered eating (Fairburn et al., 2008). There have been some investigations that examine the neural correlates of body concerns in eating pathology. For instance, participants with AN exhibit greater amygdala response to body image words, images of their body, and pictures of their body morphed to be heavier, which suggests greater emotional response (Miyake et al., 2010; Seeger, Braus, Ruf, Goldberger, & Schmidt, 2002). In contrast, AN participants displayed greater activation in a reward-related region (i.e., the ventral striatum) to underweight images (Friederich et al., 2010) and also displayed greater activation in the insula and lateral PFC (Mohr et al., 2010). In contrast, participants with BN did not exhibit greater amygdala activation in response to images of their body morphed to look heavier (Miyake et al., 2010).

In addition to increased emotionality in response to body image, ED is also associated with avoidance of body image-related cues. For example, Vocks et al. (2011) found that AN and BN participants had less activation in the inferior parietal lobe (which is related to processing of emotions and sensory information) when viewing pictures of their own body, which the authors interpreted as avoidance. Patients with BN exhibited less activation than healthy control subjects in the middle frontal gyrus and lateral occipital cortex when viewing distorted images of their body (Mohr et al., 2011). Uher et al. (2005) also found that when AN and BN subjects rated body types, they demonstrated hypoactivity in the lateral fusiform gyrus and parietal cortex relative to healthy control subjects. This pattern of results has been interpreted as avoidance of the discomfort caused by activities associated with body image. Little neuroimaging research has been conducted on BED and body image.

Another important difference to consider is the role of the substance being consumed (or avoided) in eating and addictive disorders. Food is necessary for survival, and AN is related to significant malnutrition, which can impact the brain substantially (Delvenne et al., 1997). For example, patients with AN appear to have decreased gray matter in the hypothalamus, inferior parietal lobe, lentiform nucleus, and caudate (Titova, Hjorth, Schiöth, & Brooks, 2013), which may be related to starvation. In contrast to AN, addictive disorders are related to excess consumption of a substance to the point of intoxication, which may also impact the brain. For example, acute alcohol intoxication impacts the functioning of the cerebellum (Volkow et al., 1988) and the OFC, ACC, and primary motor cortex (Calhoun, Pekar, & Pearlson, 2004). Acute cocaine intoxication impacts the ventral tegmental area, substantia nigra, nucleus accumbens, basal forebrain, globus pallidus, amygdala, and subcallosal cortex (Breiter et al., 1997), and acute heroin administration alters cerebral blood flow in the amygdala (Guyer et al., 2007). Thus, food restriction and intoxication both have marked, but different, relationships to neural functioning.

10 Summary and Future Directions

In sum, the neuroimaging literature on BED/BN has many similarities with addiction neuroimaging findings, and AN appears to exhibit a more distinct pattern of results in multiple domains (see Table 4.1). These commonalities between BED/BN and addiction speak to the current debate about the potential role of an addictive process in certain types of eating problems (Avena et al., 2012; Ziauddeen et al., 2012). Yet, it is also important to consider the relatively small number of neuroimaging studies focusing on certain types of ED (i.e., BED) and in certain domains (e.g., motivation, reward, interoceptive awareness). More research is needed before strong conclusions can be drawn. An essential future direction is to evaluate the impact of binge consumption of highly processed, calorie-dense foods on neural systems. Neuroimaging studies of individuals with comorbid eating and addictive disorders would also contribute to our understanding of the relationship between these conditions. Additionally, exploring the association between “food addiction” characteristics and neural responses in patients with ED will be helpful in evaluating how addictive tendencies may contribute to disordered eating. Finally, a greater emphasis on longitudinal designs to parse out preexisting versus later developing (i.e., “causes vs. consequences”) events in neurobiology will be important for understanding both ED and addictions.