Brain Volume

Research on brain structure in EDs has been inconsistent,Reference Van den Eynde, Suda and Broadbent ^{8 –} Reference Van den Eynde, Suda and Broadbent ¹⁶ with early studies suggesting reduced total GM and WM volumes, and reduced or normal total brain tissue volumes after recovery.Reference Van den Eynde, Suda and Broadbent ⁸ For the study of regionally specific volume alterations, brain analysis methods have become available that allow automated whole brain comparison reducing bias.Reference Van den Eynde, Suda and Broadbent ⁸ We searched studies using the terms anorexia nervosa, bulimia nervosa, volume, and imaging. A recent systematic reviewReference Van den Eynde, Suda and Broadbent ⁸ suggested reduced GM volume in AN in insula, frontal operculum, occipital, medial temporal, or cingulate cortex, but concluded that there is a high level of heterogeneity and inconsistency across studies. One recent study found increased GM volume in the dorsolateral prefrontal cortex, but in contrast others reported for instance reduced dorsolateral prefrontal cortex volume in AN which predicted body mass index.Reference Joos, Kloppel and Hartmann ^{14 ,} Reference Brooks, Barker and OʼDaly ^{17 –} Reference Friederich, Walther and Bendszus ¹⁹ After short-term recovery, AN showed reduced GM in the insula, striatum, occipital, frontal, and parietal cortex,Reference Friederich, Walther and Bendszus ¹⁹ but brain tissue seems to increase with weight gain,Reference Roberto, Mayer and Brickman ²¹ and was normal after long-term recovery in a study in adultsReference Wagner, Greer and Bailer ²² and another in adolescents.Reference Lazaro, Andres and Calvo ²³ A very recent study in adolescents and young adults with AN suggested widespread cortical thinning in AN in the ill state, but that was not the case in a group after long-term recovery from AN.Reference King, Geisler and Ritschel ²⁰

Some studies suggested correlations between illness or environmental factors and brain volume. One study suggested that illness duration could be related to reduced cerebellum and mesencephalon volumes,Reference Fonville, Giampietro, Williams, Simmons and Tchanturia ²⁴ and another report associated obstetric complications with larger volume and functional connectivity in AN.Reference Favaro, Tenconi, Degortes, Manara and Santonastaso ²⁵ The mechanism regarding how such factors could cause altered brain structure is uncertain though.

A study in AN and BN indicated less gray matter in AN compared to controls and BN in the cerebellum, temporal, frontal, and occipital cortex, but reduced caudate volume in BN compared to AN and controls; furthermore, that study also found bilaterally increased somatosensory cortex volume in AN and BN groups.Reference Amianto, Caroppo and DʼAgata ²⁶ Another study in BN suggested reduced caudate volume in the dorsal striatum but normal nucleus accumbens volume.Reference Coutinho, Ramos and Maia ²⁹ The studies in BN otherwise suggested normal or increased localized GM volume in orbitofrontal cortex and striatum.Reference Van den Eynde, Suda and Broadbent ^{8 ,} Reference Schafer, Vaitl and Schienle ²⁷ A new study that investigated cortical thickness in BN found widespread brain surface volume reductions in frontal and temporo-parietal areas.Reference Marsh, Stefan, Bansal, Hao, Walsh and Peterson ²⁸

These variable results may reflect the heterogeneity of approaches, as only some studies corrected for age or overall brain volumes, some studies distinguished restricting from binge eating/purging AN while others did not, and the effects of comorbid diagnoses or medication were often not directly taken into account. Especially important for ED research are the acute effects of dehydration and starvation,Reference Streitburger, Moller, Tittgemeyer, Hund-Georgiadis, Schroeter and Mueller ³⁰ as well as excessive exercise.Reference Freund, Faust and Birklein ³¹ Those factors may contribute to brain volume changes that relate to the acute state of dehydration or nourishment, but may not be directly related an underlying specific ED pathophysiology. Especially the studies that show widespread alterations across more or less the entire cortex may be confounded by global effects of malnutrition, and may reveal less about brain pathology that actually drives eating disorder behavior. This hypothesis is supported by a study that found that the faster adolescents had lost weight and presumably the more they had restricted food and fluid intake, the smaller their brain gray matter volume was.Reference Bomba, Riva and Veggo ³²

In an effort to avoid the effects of acute starvation, malnutrition, and dehydration, we recently studied a sample of currently ill ED individuals who had been in highly supervised treatment including a fixed meal plan for between 1 and 2 weeks and thus were nutritionally highly controlled. In addition, we controlled for age, depression, anxiety, medication-use, and brain volume. We found that brain GM volume could identify shared abnormalities among ED groups but also distinguish AN from BN individuals.Reference Frank, Shott, Hagman and Mittal ³³ This sample of individuals with restricting-type currently ill (n=19) or recovered AN (n=24), ill BN (n=19), and healthy control women (n=24) showed increased GM volume of the medial orbitofrontal cortex gyrus rectus. In addition, ill and recovered AN had increased right insula GM volumes, while BN individuals had increased left insula GM volumes compared to controls. Furthermore, dorsal striatum volumes were reduced in BN and recovered AN, and predicted sensitivity to reward in all ED groups. We also studied adolescents with AN (n=19) and controls (n=22), with similar methods. AN adolescents showed increased left orbitofrontal and right insular volumes similarly to AN adults compared to controls (see Figure 1).Reference Frank, Shott, Hagman and Yang ³⁴ A recent study in obesity from our group suggests, in contrast, reduced orbitofrontal cortex gyrus rectus volume.Reference Shott, Cornier and Mittal ³⁵

Figure 1

Increased left orbitofrontal gyrus rectus and right insula volume in anorexia nervosa across states of illness and age groups.

The orbitofrontal cortex is important in food intake control, telling us when to stop a certain type of food, regulating so-called sensory specific satiety.Reference Rolls ³⁶ It is possible that larger orbitofrontal gyrus rectus in EDs is associated with stronger sensory experience of food and maybe also other salient stimuli,Reference Frank, Reynolds and Shott ³⁷ which could drive food avoidance. In fact, the medial orbitofrontal cortex has previously been associated with food avoidance,Reference Plassmann, OʼDoherty and Rangel ³⁸ and this region therefore could be a key structure in EDs. ED phenotypes differ with restriction in AN and episodic binge eating in bulimia on the contrary—behaviors that may be driven by insula and basal ganglia differences between the two disorders.Reference Frank, Reynolds and Shott ^{37 ,} Reference Frank, Reynolds, Shott and OʼReilly ³⁹ The left insula receives information on gastric distentionReference Craig ⁴⁰ and self-reported fullness.Reference Wang, Tomasi and Backus ⁴¹ Thus, altered insula size could interfere with normal interoception in bulimia, which may contribute to a reduced ability to sense “fullness” or satiation, and then guilt-experienced overeating may trigger the urge to purge after excessive food intake. In healthy controls, left sided anterior and posterior insula activation was associated with gastric distension, and this was mediated by body mass index (BMI)Reference Tomasi, Wang and Wang ⁴² —supporting the idea that alterations in the left insula could indeed contribute to inadequately feeling full and interfering with meal termination. The right anterior insula has been associated with self-recognition, the “abstract representation of oneself,”Reference Devue, Collette and Balteau ⁴³ and interoceptive awareness,Reference Critchley, Wiens, Rotshtein, Ohman and Dolan ⁴⁴ and so a fixed perception of being fat while severely underweight in ANReference Konstantakopoulos, Varsou and Dikeos ⁴⁵ could thus be related to right-sided, increased, abnormal anterior insula volume.

Importantly, these and other dataReference Bomba, Riva and Veggo ³² suggest that when studying currently ill ED individuals, it is imperative to control for nutritional state as well as comorbidity and medication use in order to be able to get more consistent results, as well as identify brain alterations that are important for relevant ED behavior, as opposed to results that are mostly related to quickly changing effects of starvation. Controlling for those effects is not only important for volumetric studies, but new research also indicates that this factor affects studies on brain activity and function.Reference Dukart and Bertolino ⁴⁶

White Matter Integrity

Brain WM axons physically connect cortical and subcortical brain structures, and thus could have critical impact on cognitive and emotional processing.Reference Dehaene and Changeux ⁴⁷ Brain research that targets WM function uses the magnetic resonance imaging (MRI) technique diffusion tensor imaging (DTI).Reference Filler ⁵ DTI measures water diffusion along WM tracts, which is expressed as fractional anisotropy (FA). Higher FA is thought to reflect better axonal coherence, density, and myelination.Reference Le Bihan ⁴⁸ Other measures, such as the apparent diffusion coefficient (ADC) or mean diffusion coefficient, provide information about the average diffusion-freedom water molecules have across cells and correlates with local cell breakdown.Reference Jiang, van Zijl, Kim, Pearlson and Mori ⁴⁹ In this section, we searched for white matter, anorexia, and bulimia nervosa as search terms.

The first study in this research area came from our group and showed in adult ill AN compared to controls reduced FA in the bilateral fimbria-fornix, fronto-occipital fasciculus, and posterior cingulum WM, and harm avoidance was predicted by bilateral fimbria-fornix FA in AN but not controls.Reference Kazlouski, Rollin and Tregellas ⁵⁰ Those findings were not due to WM volume deficits and suggested alterations in limbic and association pathways, which could contribute to disturbed feeding, emotion processing, and body perception in AN. The prediction of harm avoidance in AN by fimbria-fornix WM integrity suggested that this pathway may be mechanistically involved in high anxiety in AN. Altered fornix WM diffusivity in adults with AN was most recently supported by 2 additional studies.Reference Nagahara, Nakamae and Nishizawa ^{51 ,} Reference Via, Zalesky and Sanchez ⁵² One of those studiesReference Nagahara, Nakamae and Nishizawa ⁵¹ also found positive correlations between BMI and FA, suggesting that weight loss could be related to the changes found. A DTI study in a mixed sample of ill and recovered adult AN found FA reductions in the posterior thalamic radiation, which includes the optic radiation, and the left mediodorsal thalamus. Those alterations could be related to body image distortion, as those pathways connect the extrastriate visual cortex with other brain regions involved in body perception.Reference Frieling, Fischer and Wilhelm ⁵³ A small study in AN after recovery indicated no FA differences between AN and control subjects, although more severe illness history was associated with lower WM integrity, and low power may have precluded significant group differences.Reference Yau, Bischoff-Grethe and Theilmann ⁵⁴ WM integrity in adolescents with AN was lower in fornix, posterior frontal, and parietal areas, but higher in anterior frontal, orbitofrontal, and temporal lobes.Reference Frank, Shott, Hagman and Yang ³⁴ Those results of both increased and decreased WM integrity measures in adolescents are interesting, and could indicate an interaction of normal developmental and ED-specific changes of WM over time. That is, ED behavior could interfere with normal WM growth in adolescence, thereby leading to long-lasting or permanent WM alterations that will be part of the larger ED brain pathophysiology. In adults with BN, FA was decreased in the bilateral corona radiata extending into the posterior limb of the internal capsule, corpus callosum, right sub-insular white matter, and right fornix.Reference Mettler, Shott, Pryor, Yang and Frank ⁵⁵ The fornix alterations pointed to a possibly common alteration of WM across ED subtypes. Lesions to the corona radiata and internal capsule have been found in central taste disorders,Reference Onoda, Ikeda, Sekine and Ogawa ⁵⁶ and deep brain stimulation in the internal capsule leads to altered taste and smell perception, as well as anxiety, panic, and mood alterations.Reference Okun, Mann and Foote ⁵⁷ Thus lesions in the corona radiata in BN could be related to a variety of specific or associated behaviors found in BN, such as mood disturbance or anxiety.

In summary, the number of studies on WM integrity in EDs is very small, but so far the results suggest that reduced fornix FA is common across AN and BN groups. The fornix is part of the limbic system, which is involved in reward processingReference Salinas and White ⁵⁸ and emotion,Reference Dalgleish ⁵⁹ but also feeding regulation as well as behavior extinction.Reference Osborne, Silverhart, Markgraf and Seggie ⁶⁰ The clinical significance of those findings will need further study, and whether this is a premorbid condition or an effect from the illness.

Brain Function with Focus on Taste-Reward Processing

Food intake is driven by a complex interplay between cognitive, emotional, and energy homeostasis–maintaining mechanisms between brain and body.Reference Kringelbach ⁶¹ There is a cognitive or cephalic phase that involves desire or craving, as well as a consummatory phase involving the hedonic experience. These mechanisms were then further described as dopamine function–associated “wanting,” or the drive to approach a reward, and “liking,” or the hedonic experience during food consumption associated with opioid system activity.Reference Berridge ^{62 ,} Reference Kelley, Schiltz and Landry ⁶³ Those processes are regulated by the brain reward system, which integrates more basic metabolic hunger signals with higher order processing of taste and cognitive-emotional factors that drive whether we approach or do not approach food stimuli.Reference Lak, Stauffer and Schultz ⁶⁴ Important brain regions that regulate those processes are the insula, as the primary taste cortex and central gateway to the dopaminergic basal ganglia and midbrain, to higher order brain centers, including the prefrontal and cingulate cortex, which integrates cognition and emotions; the orbitofrontal cortex, which determines when to stop eating a type of food; and the amygdala, which associates stimuli with emotional experience and which is thought to modulate dopamine circuitry in the midbrain and striatum.Reference Kringelbach and Rolls ^{65 –} Reference OʼReilly ⁶⁷

Several studies have investigated brain taste-reward circuits in eating disorders, and in this section, we review functional imaging studies using the search terms anorexia, bulimia, imaging, and taste. The primary focus in this review is on simple sweet taste stimuli, such as sucrose. Individuals who were recovered from AN showed reduced functional brain response to repeated but increased response to randomly given sweet taste stimuli,Reference Cowdrey, Park, Harmer and McCabe ^{68 –} Reference Wagner, Aizenstein and Mazurkewicz ⁷⁰ and those results in opposite directions suggest that unpredictable and predictable stimulus presentation activate differently circuits or neurotransmitter systems, when studying AN. Whether there are stimulus impendent reward system alterations remains to be seen. Studies using monetary reward stimuli indicated that subjects who had recovered from AN or BN showed, with respect to brain activation, less of a distinction in brain response to gain versus loss,Reference Wagner, Aizenstein and Venkatraman ^{71 ,} Reference Wagner, Aizenstein and Venkatraman ⁷² suggesting that there could be distinct reward circuit alterations depending on the saliency of the stimulus. Another approach is to pair unconditioned taste stimuli with conditioned visual or auditory stimuli and then randomly omit an expected taste delivery or deliver a taste stimulus when none was expected. This leads to a discrepancy between reward anticipated or predicted and the reward actually received, the so called “prediction error,” which is reflected in dopamine neuronal response.Reference Schultz, Dayan and Montague ⁷³ We have previously applied a prediction error taste-reward task using the sugar solution and visual conditioned cues in AN, and compared this group with obese individuals; the rationale is that we might detect neurobiological alterations that lie on opposite ends,Reference Frank, Reynolds and Shott ³⁷ as basic research has suggested.Reference Avena, Rada and Hoebel ^{74 ,} Reference Johnson and Kenny ⁷⁵ We found that insula and ventral striatum prediction error response were greater or more sensitive in AN compared to controls, while obese individuals showed reduced response, supporting the notion that BMI and extremes of food intake may lead to adaptations of brain prediction error and thus dopamine brain response in humans. A similar study in BN found prediction error response reduced compared to controls,Reference Frank, Reynolds, Shott and OʼReilly ³⁹ but not as much as in OB (see Figure 2). BN subjects’ BMI typically lies between controls and OB subjects and BMI may therefore directly be predictive of dopamine related prediction error brain response. Those results of altered prediction error response across ED groups are potentially mechanistically important, as specific dopamine receptors have been implicated, such as the dopamine D1 and D2 receptor,Reference Maia and Frank ⁷⁶ making those receptors potential treatment targets. Otherwise, the literature on taste-reward in BN is small. After a small pilot study found lower anterior cingulate activation after sucrose stimulation in individuals who had bulimic symptoms in the past,Reference Frank, Wagner and Achenbach ⁷⁷ 2 studies in BN after recovery indicated increased brain response in that group compared to controls. One of those studies applied repeated fat and high viscous stimuli and reported increased activation in the anteroventral striatum.Reference Radeloff, Willmann and Otto ⁷⁸ The other study, which used a task with recurrent sweet taste solution delivery, reported heightened anterior insula activation.Reference Oberndorfer, Frank and Simmons ⁶⁹

Figure 2

Taste-reward prediction error response across groups with disordered eating patterns.

We know little about reward processing in youth with ED, and only one small study tested this system in adolescents with AN.Reference Bischoff-Grethe, McCurdy and Grenesko-Stevens ⁷⁹ In that study, AN adolescents exhibited an exaggerated response to losses compared to wins in posterior executive and sensorimotor striatal regions, which supports the notion that AN is associated with heightened sensitivity to salient stimuli.Reference Jappe, Frank and Shott ⁸⁰

In contrast to studying taste stimuli that are limited to basic nutrients such as sugars, others have investigated complex taste and food stimuli. This may have the benefit to reflect more real life situations. On the other hand, those stimuli may be more difficult to model with respect to the underlying neurobiology of taste pathways. The emotional or pleasantness value that ED subjects assign to a simple sugar solution tends to be similar to controls, while complex tastes such as highly palatable sugar/fat combinations could be associated with for instance higher fear in the ED groups, which might confound the more basic taste pathway signal. One study, for instance, applied chocolate milk and found greater activation in the right amygdala and left medial temporal gyrus in restricting-type AN compared to controls when hungry, contrasted against the satiety state.Reference Vocks, Herpertz, Rosenberger, Senf and Gizewski ⁸¹ This could be a sign of heightened vigilance and anxiety in that group, as having the chocolate milk indicates breaking the fasting and promoting weight gain by drinking the caloric beverage. Another study that applied a chocolate milkshake found that women with BN had a positive correlation between negative affect and activity in the putamen, caudate, and pallidum during milkshake anticipation.Reference Bohon and Stice ⁸² It was hypothesized that negative affect may increase the reward value of food in BN, but it may be more likely that negative affect became a conditioned response to palatable food, as it is associated with weight gain.

In summary, repeated and thus predictable application of basic sweet taste stimuli has been associated with reduced activation of taste-reward important regions, such as the insula and ventral striatum in AN, but increased activation in BN. The opposite, though, is the case for random application of those stimuli. This may suggest that there may be an interaction between conscious cognitive-emotional and more unconscious biological mechanisms that drive food approach and eating. A potential explanation here may be that during repeated taste application, the AN subjects “prepare” themselves and control brain response in order to avoid too high stimulation; however, during the random application, this control is not possible and an enhanced responsiveness in AN is coming to light. In BN, a reduced biological responsiveness to taste stimulation becomes obvious during random application, while the repeated taste stimulation could kindle and enhance the low baseline response, maybe through the repeated hedonic experience.

Networks

A variety of functionally connected brain networks has been identified that drive behavior, including the default mode network (DMN), salience network (SN), and executive networks for “higher-order” processing, as well as sensory and sensorimotor (SMN) networks for “lower-order” function.Reference Heine, Soddu and Gomez ⁸³ We used the search terms anorexia, bulimia nervosa, network, and connectivity, but few studies have investigated those networks in EDs. Importantly, the methods vary: some studies investigated connectivity during resting state, others during performance of a task; some studies adhere to more strict observation of distinctly described networks, while others investigated connectivity or synchronicity of various brain regions and then speculated on functional networks involved. One study found decreased activity within the ventral visual network in ill and recovered AN patients compared to controls, but also found increased somatosensory network activity in AN.Reference Favaro, Santonastaso and Manara ⁸⁴ This could be in line with a recent study that found altered connectivity in visual pathways when viewing human bodies,Reference Suchan, Bauser and Busch ⁸⁵ suggesting overall visual network disturbances that could be related to altered body/self perception. One study in recovered AN showed increased DMN activity,Reference Cowdrey, Filippini, Park, Smith and McCabe ⁸⁶ and in a study that contrasted AN, BN, and controls, patterns of connectivity between insula and frontal brain regions distinguished groups during a visual food cue task.Reference Kim, Ku, Lee, Lee and Jung ⁸⁷ A recent study found alterations in connectivity patterns within the cerebellum in AN and BN compared to controls,Reference Amianto, DʼAgata and Lavagnino ⁸⁸ but the meaning and/or functional implications are uncertain. Similarly, a study in a small sample of AN indicated reduced resting state connectivity in AN in the frontal cortex in regions that contribute to cognitive control.Reference Kullmann, Giel, Teufel, Thiel, Zipfel and Preissl ⁸⁹ That study further indicated an interesting distinction, that is, from the inferior frontal gyrus to the cingulum connectivity decreased, but to the orbitofrontal cortex connectivity increased in the AN group. A study from our group found reduced SN activity during taste stimuli delivery in ill and recovered AN compared to controls, which could be a trait-related biomarker or illness remnant altering the drive to approach food.Reference McFadden, Tregellas, Shott and Frank ⁹⁰ We also found reduced DMN and SMN activity but in ill AN only, suggesting state-dependent abnormalities, possibly related to altered interoception and body image in AN when underweight, but remitting following recovery.

Taken together, this literature is still very small, but it seems that visual and salience network alterations could be involved in the pathophysiology of EDs, as state-dependent factors or even as biological traits.

Neurotransmitters

Neurotransmitter receptor imaging studies assess the “functional availability” of neurotransmitter receptors in the brain. Several studies have advanced our knowledge over the past decade, although there has also been a limitation in available receptor ligands that can be used in humans, which thus limits our ability to characterize comprehensively neurotransmitter receptor systems. Serotonin (5-HT) 1A receptor binding was found to be elevated across most brain regions in a mixed group of symptomatic restricting and binge eating/purging–type AN subjects compared to controls, as well as in binge eating/purging type AN after recovery.Reference Bailer, Frank and Henry ⁹¹ In contrast, recovered restricting-type AN individuals showed normal brain 5-HT1A binding.Reference Bailer, Frank and Henry ⁹¹ In addition, reduced 5-HT2A binding in frontal, parietal, and occipital cortices in ill and recovered AN individuals was found.Reference Frank, Kaye and Meltzer ⁹² In summary, after recovery, 5-HT1A receptor binding seems to differentiate AN subtypes, whereas 5-HT2A receptor binding is reduced in both restricting and binge eating/purging AN in various brain regions. Since these disturbances occur after recovery, they may reflect either trait disturbances or scars from the illness.

Various studies could correlate 5HT receptor availability or the interaction between 5HT and DA receptors with harm avoidance, which is a behavioral correlate of anxiety,Reference Bailer, Frank and Price ⁹³ but the exact mechanisms or functional relationships need further study. In BN, 5-HT transporter binding was reduced in the thalamus and hypothalamus,Reference Tauscher, Pirker and Willeit ⁹⁴ but 5-HT1A receptor binding was increasedReference Tiihonen, Keski-Rahkonen and Lopponen ⁹⁵ in the medial prefrontal cortex, posterior cingulate, and parietal cortex. After recovery, BN had increased 5-HT1A binding compared to controls,Reference Bailer, Bloss and Frank ⁹⁶ and 5-HT1A binding in BN predicted inhibition. The dynamics between 5-HT receptor expression and synaptic 5-HT are not well understood, but reduced 5-HT2A binding in recovered BN subjects may be related to a higher level of endogenous 5-HT in the synaptic cleft, or a downregulation of the receptor.

A mixed group of recovered restricting-type and recovered binge-eating/purging–type AN women showed increased DA D2/D3 receptor binding in the antero-ventral striatum,Reference Frank, Bailer and Henry ⁹⁷ while decreased DA D2/D3 receptor binding was found in obesity.Reference Wang, Volkow, Thanos and Fowler ⁹⁸ Those findings support the possibility that D2/D3 receptor binding may be inversely related to weight and eating, with restricting-type AN on one end and obesity on the other end of the spectrum. It is possible that increased DA D2/D3 receptor binding in AN is part of underlying mechanisms that may explain why individuals with AN are able to resist eating. It is worth noting that food restriction sensitizes D2/D3 receptors, while excessive food intake downregulates DA D2 receptors in rats; similar mechanisms in AN or obesity could complicate recovery.Reference Johnson and Kenny ⁷⁵ A very recent study found that BN individuals had a trend to lower DA D2/3 receptor binding in the striatum as well as less DA release compared to controls in response to methylphenidate application.Reference Broft, Shingleton and Kaufman ⁹⁹ While overall this body of research is small, it suggests that food restriction in AN may increase and episodic binge eating in BN or chronic overeating in obesity could reduce DA receptor activity.

Disclosures

Guido Frank does not have anything to disclose.