Eating disorders are psychiatric conditions primarily characterized by disturbances in eating behaviors as well as in thoughts and emotions related to eating. Individuals suffering from these disorders experience disordered beliefs related to weight and body image, which can lead to severe psychological and physical harm [1]. The most common of these disorders, and the primary focus of this paper, are anorexia nervosa (AN), bulimia nervosa (BN), and binge eating disorder (BED).

AN is a disorder related to an extreme fear of weight gain and a distorted view of one’s own body, often leading an individual to take extreme measures to maintain or lose weight [2]. The pathologic behaviors exhibited by patients with AN include excessive physical activity, extreme dietary restriction, and purging [2]. Additionally, this disorder can impact an individual’s cognitive and emotional functioning and is often accompanied by medical and psychiatric comorbidities [2], such as bipolar, depressive, and anxiety disorders, as well as life-threatening conditions, including amenorrhea, vital sign abnormalities, malnutrition, the loss of bone mineral density, and abnormal lab findings [1]. AN has a lifetime prevalence of approximately 0.3% [3].

BN involves recurrent episodes of binge eating followed by subsequent inappropriate actions to avoid weight gain from the binge eating episode [4]. These compensatory methods can be harmful and may include self-induced vomiting, laxatives, or prolonged periods of starvation [5]. The psychiatrist Gerald Russell differentiated patients suffering from anorexia nervosa and bulimia nervosa by describing those with BN as normal weight or overweight, while individuals with anorexia nervosa are severely underweight [5]. Sequelae of the disorder due to purging can be fatal and include esophageal tears, gastric rupture, and cardiac arrhythmias [1]. Up to 3% of females and 1% of males will experience BN within their lifetimes [3].

BED is the most common eating disorder [6]. In contrast to bulimia nervosa, BED involves an individual experiencing recurrent episodes of binge eating without subsequent actions to compensate for overeating [4]. The disorder, naturally, is often associated with obesity and metabolic syndrome due to binging and the subsequent high caloric intake without the compensatory measures central to AN and BN [7]. The prevalence of BED worldwide for 2018–2020 is between 0.6 and 1.8% in adult women and 0.3 and0.7% in adult men [8].

2. Current Neuromodulatory Options and Their Target Networks/Nodes

2.1. Non-Invasive Brain Stimulation

Non-invasive brain stimulation refers to methods of modulating brain and network activity via non-surgical procedures. These methods commonly include repetitive transcranial magnetic stimulation (rTMS) and transcranial direct current stimulation (tDCS). In the context of eating disorders, both methods have been thoroughly studied in vivo, with rTMS representing a larger share of the studies. Currently, neuromodulatory techniques are neither considered first-line nor approved therapies for eating disorders ^[9][39]. This is true for both rTMS and tDCS, the two most commonly investigated non-invasive methods discussed in this research; however, there is substantial evidence to suggest these methods are both safe and beneficial in the treatment of AN, BN, and BED ^{[10][11][12][13][14][15][16][17][18]}[40,41,42,43,44,45,46,47,48]. In the articles discussed below, the application of rTMS (summarized in Table 1) and tDCS (summarized in Table 2) often followed a prolonged eating disorder course, many times refractory to pharmacotherapy and psychotherapy, which is suggestive of neuromodulation being indicated in cases of resistant eating disorders. As previously discussed, there are numerous aberrant brain networks involved in disordered eating. Most of the studies of non-invasive brain stimulation have targeted the dlPFC, and several have targeted the dmPFC ^[19][20][32,49]. Likewise, the dmPFC has been shown to have differential activity in AN, BN, and BED when compared to healthy controls ^[21][50]. As mentioned prior, the PFC and its partitions play a role in both executive inhibitory control and reward processing, the dysfunction of which is contributory to all of the eating disorders discussed here ^[22][51]. Naturally, the targeting of these regions aims to override the aberrancy of the various frontal–striatal networks. To date, no other targets have been investigated, but this is certainly a potential area of study in this nascent field. ^[56][85]. Smaller studies have largely supported the findings of the TIARA study and the use of TMS for the treatment of AN. McClelland and colleagues performed a randomized, double-blind, sham-controlled trial of rTMS to the left dlPFC in 49 patients ^[27][56]. At 24 h post-treatment, they found a tendency toward improvement in core symptoms, although it did not reach statistical significance. Examining changes in core symptoms, McClelland et al. ^[23][52] and Van den Eynde et al. ^[24][53] reported their experiences with rTMS to the left dlPFC in two and ten patients with refractory AN, respectively, both finding improvements in core symptoms. Case reports by Kamolz et al. ^[23][52] and Choudhary et al. ^[28][57] also demonstrated significant improvements in BMI, core symptomatology, and in the case of Kamolz et al.’s patient with comorbid MDD, depressive symptoms. Targeting the dmPFC bilaterally, Woodside and colleagues performed a case series of rTMS in 19 AN patients with comorbid MDD and, in some participants, PTSD ^[11][41]. The authors found significant improvements in AN symptomatology and mood symptoms, namely anxiety and depression. While one case report by Jaššová and colleagues ^[29][58] reported no change either in BMI or core symptomatology and Mclelland et al.’s ^[25][54] case series of two patients found no improvement in BMI, the existing body of evidence suggests rTMS over an extended timeframe, applied to either the dlPFC or dmPFC, is an effective treatment for AN.

rTMS for BN

Second to AN, BN is the most studied eating disorder with respect to rTMS. In regard to the binge–purge behavior characteristic of BN, RCTs exploring rTMS for the treatment of BN were more limited and have conflicting reports of efficacy. Van den Eynde et al. performed a randomized, sham-controlled, double-blind study with 38 patients to investigate rTMS to the left dlPFC ^[12][42], finding self-reported food cravings and the frequency of binging episodes to be significantly decreased. However, both Walpoth et al. ^[33][62] and Gay et al. ^[36][65] conducted RCTs with 14 and 47 participants, respectively. Neither of these two groups were able to find significant differences in binge–purge behaviors between the treatment and sham groups. When examining psychometrics, the results were equally variable. Walpoth et al. ^[33][62] were not able to find significant improvements in the measures of depression or obsessive-compulsive behavior between groups. Conversely, Guillaume et al. investigated the effects of the rTMS treatment on decision making and impulse control ^[57][86], finding a statistically significant improvement in impulse control, as well as decision making when assessed using the Iowa gambling task. There also exist studies of smaller sample sizes that explore rTMS in BN, finding more consistent therapeutic results. In another study by the Van den Eynde group—a case series of rTMS applied to the left dlPFC in seven left-handed patients ^[34][63]—the group found decreases in BN core symptomatology with variable changes in mood depending on hand dominance. Hausmann and colleagues’ case study presented a patient with BN and comorbid MDD who underwent left dlPFC rTMS ^[32][61], resulting in the complete remission of binging behavior and a 50% decrease in depressive symptoms. Targeting the bilateral dmPFC, Downar and colleagues reported the case of a patient with BN and comorbid MDD ^[35][64]. The patient experienced a complete remission of binging–purging behaviors up until their 2-month follow up. In a subsequent study by the same group, Dunlop and colleagues reported a case series involving 28 patients ^[26][55], in which 16 of the 28 patients experienced a >50% reduction in weekly binges at their 4-week follow-ups. fMRI imaging was obtained before and after treatment, showing lower baseline functional connectivity between the dmPFC, and the structures of the lateral OFC and right posterior insula. Responders additionally exhibited lower baseline functional connectivity between the dACC, the right posterior insula, and the right hippocampus, which is consistent with the aforementioned dysfunctional reward pathways. Dunlop and colleagues also identified lower baseline functional connectivity between the dACC and the ventral striatum as well as the anterior insula in responders and found that such connectivity increased with treatment. Conversely, non-responders were found to have high baseline frontostriatal functional connectivity, which was decreased by rTMS and correlated to a worsening of symptoms ^[26][55]. Despite promising case studies and case series, multiple sham-controlled, double-blind RCTs failed to observe any difference between rTMS and a sham for patients with BN ^[33][36][62,65]. As such, further RCTs would prove beneficial in characterizing the target, the parameters, and the patient selection that are optimal for rTMS to be effective in BN studies.

rTMS for BED

There are few studies addressing rTMS for BED. There is a clear need for more studies to define the role of non-invasive brain stimulation in the treatment of this disorder. Sciortino and colleagues published one of the few studies on the topic, presenting two patients with BED and comorbid treatment-resistant bipolar disorder type II who were treated with iTBS to the left dlPFC ^[13][43]. Both patients experienced complete remission in their binge eating symptoms that lasted until their 12-week follow-up visits with only a minor improvement in their depressive symptoms; manic symptoms were absent throughout.

3.1.2. Transcranial Direct Current Stimulation (tDCS)

tDCS is another form of non-invasive brain stimulation that has been studied in a variety of psychiatric and neurologic disorders, including MDD ^[58][87], schizophrenia ^[59][88], substance use disorder ^[60][89], OCD ^[61][90], GAD ^[62][91], and eating disorders. Although there are currently no FDA-approved protocols for tDCS, it has shown promise in numerous clinical studies ^[63][92]. In tDCS, electrodes are placed on the scalp, and a weak current is passed through the brain between the two electrodes. Anodal stimulation is thought to be excitatory, while cathodal stimulation is thought to be inhibitory ^[64][93]. Most tDCS protocols deliver a current of 1–2 mA in 10–20 min treatment sessions, and patients undergo 10–20 treatments ^[63][92].

tDCS for AN

There have been several studies investigating the effects of tDCS on AN, as well as a few ongoing RCTs that are not yet concluded. The largest of the completed studies is a single-blind trial consisting of 23 patients by Costanzo and colleagues. They applied tDCS to the left dlPFC and compared the effects to those of standard therapy ^[15][45]. They identified increases in BMI in the tDCS group only at a one-month follow-up. Khedr and colleagues performed an open-label, single-arm study consisting of seven patients who received tDCS to the left dlPFC ^[37][66]. In the study, the group found a statistically significant improvement in the core AN and depressive symptomatology at the 1-month follow-up visits. It is worth mentioning a case report published by Mares et al. ^[39][68] whereby a patient with comorbid PTSD was discovered to have type I diabetes mellitus (DM) during tDCS to the left dlPFC. It is unclear whether the onset of the patient’s DM was a direct consequence of stimulation. The currently ongoing studies include a randomized, sham-controlled clinical trial of tDCS to the left inferior parietal lobule (IPL) and is being conducted by Phillipou and colleagues ^[38][67] and Ursumando et al. ^[40][69]. Phillipou et al. chose the IPL given the decreased functional connectivity between midbrain structures and the IPL that they previously observed in patients with AN ^[65][94]. Ursumando et al. ^[40][69] are conducting a randomized, sham-controlled clinical trial of tDCS to the left dlPFC in the pediatric population.

tDCS for BN

The literature surrounding the tDCS of BN is relatively sparse, with only one primary research article having been identified. That study, conducted by Kekic and colleagues, consisted of a double-blind, sham-controlled proof-of-principle trial utilizing tDCS applied to the right and left dlPFC for 39 patients with BN ^[16][46]. In their study, patients received three sessions of tDCS—anode right/cathode left, cathode right/anode right, and sham—in a counterbalanced, randomized order. A variety of binge eating and psychologic tests were performed after each session and at 24 h. Binge eating symptoms and the ability to value delayed rewards improved with both the right and left anodal montages. Interestingly, mood symptoms improved only with the right anodal montage. Such results signify the importance of further exploration of tDCS’s role in the treatment of BN ^[16][46].

tDCS for BED

The literature regarding tDCS for the treatment of BED is more numerous. Overall, the results have been consistently promising. In a case series involving 30 patients with BED, Burgess and colleagues performed 2 mA tDCS to the dlPFC or a sham in a counterbalanced study paradigm ^[41][70]. They found tDCS to be associated with decreased food cravings and intake, as well as binge eating desire (with a larger effect size in the male cohort). In a double-blind RCT, Max et al. investigated the effect of tDCS to the dlPFC in a counterbalanced order in 31 patients ^[43][72]. They performed a food-modified antisaccade learning task and found that the 2 mA tDCS protocol decreased latencies and binge eating episodes, whereas the 1 mA tDCS increased latencies and had no effect on binge eating episodes. These results suggest that tDCS monotherapy at 2 mA may improve eating disorder symptoms, possibly as a result of improved inhibition functions regarding rewarding food stimuli, while 1 mA stimulation may have a detrimental effect. Additionally, there is considerable interest in studying tDCS as an adjunct therapy for BED. In a phase-II sham-controlled, double-blind randomized control trial, Giel and colleagues assessed the effect of combining inhibitory control training with right-anodal 2 mA tDCS to the dlPFC vs. a sham in 41 patients with BED [8]. They found a significant reduction in binge eating frequency in both the treatment and the sham groups at 4 weeks and 12 weeks, with a statistically significant difference between the treatment and sham groups only at the 12-week timepoint. Gordon et al. are currently conducting a sham-controlled crossover RCT that examines the effect of a combined approach bias modification (ABM) training co-administered with either anodal tDCS to the right dlPFC or a sham in 66 patients with BED ^[42][71]. ABM is a learning method that reinforces avoidance behavior in response to food cues. The authors of the study have thus far published on the patients’ experience with ABM, and overall, patients have found ABM to be a worthwhile activity ^[66][95]. Full results from the work by the group are pending. The TANDEM trial by Flynn and colleagues is another exciting single-blind, sham-controlled randomized trial that will look at self-administered at-home anodal tDCS to the right dlPFC with concurrent ABM training ^[45][74].

2.1.3. Electroconvulsive Therapy (ECT)

ECT is the oldest form of non-invasive brain stimulation that has shown some efficacy in a variety of psychiatric disorders, including treatment-refractory depression ^[67][96], bipolar disorder ^[68][97], and schizophrenia ^[69][98]. ECT induces a brief, generalized tonic-clonic seizure through an external current ^[70][99]. It is generally a well-tolerated procedure with a low risk profile, and its use is mostly restricted to the acute inpatient treatment of severe, refractory psychiatric disorders (e.g., mania, schizophrenia, major depressive disorder, and suicidality) ^[68][71][97,100]. Despite its well-established use in multiple psychiatric conditions, there have been few studies investigating the efficacy of ECT in the context of eating disorders. Given its restricted use to acute psychiatric crises and the need for general anesthesia, ECT is unlikely to find significant use in the outpatient management of disordered eating when compared to other non-invasive brain stimulation methods.

2.2. Invasive Neuromodulation

As opposed to non-invasive brain stimulation, invasive neuromodulation involves surgical intervention and the placement of deeper-reaching stimulation devices. While these methods include both deep brain stimulation (DBS) and Vagal nerve stimulation (VNS), the vast majority of invasive neuromodulation for eating disorders involve DBS. The curation of potential targets for DBS largely stems from attempts to treat comorbid conditions, namely MDD ^[72][101] and OCD ^[73][102]. Today, studies primarily target the subcallosal cingulate (SCC) target ^[72][74][101,103] and the nucleus accumbens (Nacc) ^[75][76][77][104,105,106], the exploration of which have been largely led by Canadian and Chinese groups, respectively. Alternatively, Israël et al. reported one of the earliest cases of SCC DBS in a 52-year-old woman with comorbid depression ^[72][101] as the SCC is a target that was previously explored for the treatment of depression as well as OCD. Furthermore, studies of OCD patients who underwent Nacc DBS also reported improvement in anorexia symptoms ^[73][102].

2.2.1. Deep Brain Stimulation

DBS involves the implantation of an electrode into a deep brain structure ^[75][104]. Given this degree of invasiveness, DBS is generally an escalatory treatment option, usually reserved for those that are refractory to less invasive methods ^[74][75][78][103,104,107]. The Nacc and subcallosal cingulate (SCC) are the most common targets for DBS (Figure 1). The evidence is largely limited to a handful of case series and reports, which are summarized in Table 3. Many studies argued that DBS is warranted if patients are in a life-threatening situation due to a medically refractory disease ^[75][104] with no other reasonable treatment options, as the potential benefits would then outweigh the risks of surgery.