The Neuroscience of Eating Disorders: How Brain Circuits Drive Anorexia, Bulimia, and Binge Eating

Eating disorders — including anorexia nervosa (AN), bulimia nervosa (BN), and binge eating disorder (BED) — are among the most lethal psychiatric conditions. Anorexia nervosa carries the highest mortality rate of any mental health disorder. Yet for decades, eating disorders were mischaracterized as lifestyle choices, vanity-driven behaviors, or failures of willpower. Neuroscience has fundamentally overturned that narrative.

Modern neuroimaging, genetics, and computational neuroscience reveal that eating disorders arise from measurable differences in brain structure, function, and neurochemistry. These are not disorders of wanting to be thin. They are disorders of reward processing, interoception (the brain's ability to sense internal body signals), cognitive control, and habit formation — all mediated by identifiable neural circuits.

Understanding the neuroscience of eating disorders matters for several reasons. It reduces stigma by grounding these conditions in biology. It explains why recovery is so difficult and why relapse rates are high. And it is beginning to open the door to targeted, neuroscience-informed treatments that go beyond traditional talk therapy alone.

This article examines the key brain systems implicated in eating disorders, the neurotransmitters and hormones involved, current research findings, and what the science means for clinical care.

Eating disorders involve disruptions across multiple interconnected brain networks rather than a single "broken" region. The most consistently implicated systems include:

• Insula: The insular cortex is the brain's primary hub for interoception — the perception of internal bodily states such as hunger, fullness, pain, and heartbeat. Neuroimaging studies consistently show altered insular activity in individuals with anorexia nervosa and bulimia nervosa. In anorexia, the insula appears to process hunger and satiety signals abnormally, which helps explain how individuals can override starvation signals that would be intolerable for most people. Reduced insular responsiveness has been documented even after weight restoration, suggesting this is a trait-level vulnerability rather than simply a consequence of malnutrition.
• Striatum (caudate, putamen, nucleus accumbens): The striatum is central to reward processing and habit formation. In anorexia nervosa, research shows blunted reward responses to food in the ventral striatum but heightened activity in the dorsal striatum — a pattern consistent with the shift from goal-directed behavior to compulsive, habit-driven behavior. In binge eating disorder and bulimia, the ventral striatum shows heightened reactivity to food cues, similar to patterns observed in substance use disorders.
• Prefrontal Cortex (PFC): The dorsolateral and ventromedial prefrontal cortex govern executive function, decision-making, and cognitive flexibility. In anorexia nervosa, enhanced prefrontal control is thought to enable extreme dietary restriction by overriding reward and hunger signals. In bulimia and binge eating disorder, impaired prefrontal regulation is associated with difficulties inhibiting binge-eating urges.
• Amygdala: This region processes threat and emotional salience. In eating disorders, the amygdala often shows exaggerated responses to food images, body-image stimuli, and social evaluation. This heightened emotional reactivity to food and body-related cues is a consistent finding across eating disorder diagnoses.
• Hypothalamus: As the brain's master regulator of energy homeostasis, the hypothalamus integrates signals from hormones like leptin, ghrelin, and insulin to regulate appetite and metabolism. Chronic starvation in anorexia nervosa disrupts hypothalamic function, and emerging evidence suggests that some hypothalamic abnormalities may predate the illness rather than simply result from it.

Critically, these regions do not operate in isolation. Eating disorders involve disrupted connectivity between these areas — particularly between the prefrontal cortex, insula, and striatum. The balance (or imbalance) between cognitive control circuits and reward circuits is a central theme in the neuroscience of all eating disorders.

Several neurotransmitter and hormonal systems are dysregulated in eating disorders, contributing to the characteristic patterns of behavior, mood, and cognition associated with these conditions.

Serotonin (5-HT): Serotonin is one of the most studied neurotransmitters in eating disorder research. Individuals with anorexia nervosa show alterations in serotonin receptor binding — particularly at 5-HT1A and 5-HT2A receptors — that persist after weight recovery. Serotonin is involved in satiety signaling, anxiety regulation, and harm avoidance. One influential theory proposes that individuals vulnerable to anorexia experience a chronically dysphoric state driven by excessive serotonergic activity, and that food restriction paradoxically reduces serotonin synthesis (because tryptophan, serotonin's precursor, comes from food), temporarily relieving anxiety. This creates a neurochemical reinforcement loop: restriction reduces distress, which reinforces further restriction.

Dopamine: The dopaminergic reward system is implicated across all eating disorder subtypes, though in different directions. In anorexia nervosa, altered dopamine signaling in the striatum is associated with reduced pleasure from food and an enhanced sense of reward from restraint and weight loss. In bulimia nervosa and binge eating disorder, dopaminergic dysregulation mirrors patterns seen in addictive behaviors — heightened anticipatory reward responses to food cues combined with diminished satisfaction from actual consumption, which drives escalating binge episodes.

Ghrelin and Leptin: Ghrelin (the "hunger hormone") and leptin (the "satiety hormone") are profoundly disrupted in eating disorders, especially anorexia nervosa. Ghrelin levels are elevated during starvation as the body attempts to stimulate eating, yet individuals with anorexia appear to have a blunted brain response to these signals. Leptin, which is produced by fat tissue, drops to very low levels in underweight states, contributing to amenorrhea, bone loss, and further metabolic disruption. These hormonal changes also feed back into brain circuits, affecting mood, cognition, and reward processing.

Cortisol and the HPA Axis: The hypothalamic-pituitary-adrenal (HPA) axis, the body's stress response system, is hyperactivated in anorexia nervosa. Elevated cortisol levels contribute to anxiety, cognitive rigidity, and further metabolic disturbance. Chronic stress activation also impairs hippocampal function, which has implications for memory, learning, and emotional regulation.

Endogenous Opioids: There is emerging evidence that the endogenous opioid system, which modulates pleasure and pain, is altered in eating disorders. Some researchers hypothesize that restrictive eating or purging behaviors trigger opioid release, creating a neurochemical basis for the "addictive" quality of these behaviors.

Eating disorders are substantially heritable. Twin studies estimate heritability at 50–80% for anorexia nervosa, 55–65% for bulimia nervosa, and approximately 45–57% for binge eating disorder. These are comparable to the heritability estimates for schizophrenia and bipolar disorder, underscoring that eating disorders are strongly biological in origin.

Genome-wide association studies (GWAS), particularly the landmark Anorexia Nervosa Genetics Initiative (ANGI) and the Psychiatric Genomics Consortium, have identified significant genetic loci associated with anorexia nervosa. A pivotal 2019 study published in Nature Genetics identified eight genetic loci significantly associated with anorexia nervosa and found that the genetic architecture of anorexia overlaps with both psychiatric traits (obsessive-compulsive disorder, depression, anxiety, schizophrenia) and metabolic traits — including genetic correlations with fasting insulin levels, leptin, type 2 diabetes (inversely), and body composition measures.

This was a paradigm-shifting finding. It suggests that anorexia nervosa is not purely a psychiatric disorder but a metabo-psychiatric condition — one in which both psychological and metabolic genetic factors contribute to risk. This may help explain why weight restoration is so physiologically difficult for some individuals with anorexia: their metabolic genetics may actively work against maintaining a higher body weight.

Genetic research in bulimia nervosa and binge eating disorder is less advanced but progressing. Common genetic variants implicated in impulsivity, reward sensitivity, and obesity risk appear to contribute to vulnerability for these conditions.

One of the most important and often overlooked aspects of eating disorder neuroscience is the profound effect of malnutrition itself on brain structure and function. Starvation is not just a symptom of anorexia nervosa — it is a perpetuating mechanism that remodels the brain in ways that make recovery harder.

Structural neuroimaging studies show that individuals with active anorexia nervosa exhibit significant reductions in gray matter volume and white matter integrity. Global gray matter reductions of 5–10% are commonly reported, with particularly pronounced loss in the prefrontal cortex, insula, and parietal regions. These are the very brain areas needed for accurate body perception, flexible decision-making, and interoceptive awareness.

The encouraging finding is that most of these structural changes are at least partially reversible with nutritional rehabilitation. Studies tracking brain volume during weight restoration show significant recovery of gray matter, though some subtle differences in white matter structure and cortical thickness persist for months or even years. Whether complete neurological recovery occurs in all cases remains an active area of investigation.

Functionally, starvation impairs cognitive flexibility, enhances rigid thinking, narrows attentional focus, and intensifies anxiety — all of which reinforce eating disorder behaviors. This creates a vicious neurobiological cycle: the eating disorder causes brain changes that make the eating disorder harder to overcome. This is one of the strongest neuroscience-based arguments for early intervention and the priority of nutritional rehabilitation in treatment.

Starvation also disrupts the gut-brain axis. The gut microbiome, which communicates with the brain via the vagus nerve and neuroendocrine pathways, is dramatically altered in eating disorders. Emerging research suggests that these microbial changes influence mood, anxiety, appetite regulation, and even autoimmune processes that affect brain function, though this field is still in its early stages.

Eating disorders share neurobiological features with several other psychiatric conditions, which has significant implications for understanding and treating them.

Addiction: Binge eating disorder and bulimia nervosa show striking neurobiological parallels with substance use disorders. These include heightened cue reactivity in the ventral striatum, impaired prefrontal inhibition, and a pattern of escalation and compulsive behavior despite negative consequences. Some researchers have proposed a "food addiction" model, though this remains debated. The neural overlap is real, but food is not a substance that can be abstained from entirely, which makes the clinical picture more complex than substance addiction.

Obsessive-Compulsive Disorder (OCD): Anorexia nervosa in particular shares significant overlap with OCD — both genetically and neurologically. Both conditions involve hyperactivity in cortico-striatal-thalamic circuits, excessive behavioral rigidity, and a compulsive quality to core behaviors. Research estimates that 25–69% of individuals with anorexia nervosa also meet criteria for OCD at some point in their lives. Genetic studies confirm shared genetic liability between the two conditions.

Anxiety Disorders: Anxiety disorders frequently precede eating disorder onset, and the neuroscience suggests shared vulnerability in amygdala reactivity, serotonergic function, and threat processing circuits. In many cases, disordered eating behaviors begin as maladaptive strategies for managing intolerable anxiety — strategies that become entrenched through neurochemical reinforcement and habit formation.

According to the DSM-5-TR, eating disorders are classified in their own diagnostic category (Feeding and Eating Disorders), separate from anxiety, OCD, and substance use disorders. However, the neurobiological boundaries between these categories are far less distinct than the diagnostic manual implies. Recognizing this overlap is essential for comprehensive treatment planning.

The neuroscience of eating disorders is advancing rapidly across several frontiers:

• Computational Psychiatry: Researchers are using computational models to quantify how individuals with eating disorders process reward, make decisions, and learn from feedback. These models go beyond simple brain scan images to characterize the specific algorithms the brain uses — and how those algorithms differ in eating disorders. For example, computational studies have shown that individuals with anorexia nervosa exhibit an exaggerated "model-based" (planning-oriented) decision style, consistent with excessive cognitive control overriding basic reward signals.
• Interoception Research: A growing body of work examines how individuals with eating disorders perceive (or fail to perceive) internal body signals. Altered interoceptive accuracy and interoceptive confidence — the ability to detect and trust internal states like hunger, fullness, and emotion — is emerging as a transdiagnostic feature across anorexia, bulimia, and binge eating disorder. Interventions targeting interoceptive retraining are currently in development.
• Neuromodulation: Transcranial magnetic stimulation (TMS) and transcranial direct current stimulation (tDCS) targeting the dorsolateral prefrontal cortex and other regions are being studied as potential treatments. Early results for repetitive TMS in anorexia nervosa and binge-type eating disorders are cautiously promising, with some studies showing reductions in binge-purge frequency and improvements in mood. Deep brain stimulation (DBS) has been explored in treatment-resistant anorexia nervosa in small case series, targeting the nucleus accumbens or subcallosal cingulate, though this remains highly experimental.
• Neuroinflammation: There is growing interest in the role of neuroinflammation in eating disorders. Chronic malnutrition, gut dysbiosis, and HPA axis overactivation all promote inflammatory processes in the brain. Some researchers hypothesize that neuroinflammation contributes to the cognitive rigidity, anhedonia, and treatment resistance seen in chronic anorexia nervosa.
• Epigenetics: Beyond inherited genetic variants, researchers are investigating how eating disorder behaviors themselves alter gene expression through epigenetic mechanisms. Starvation, purging, and chronic stress leave epigenetic marks — changes in DNA methylation and histone modification — that can affect brain function and potentially be transmitted to offspring. This is an early but compelling area of inquiry.

Neuroscience has not yet revolutionized eating disorder treatment in the way that many researchers hope, but it is already informing clinical practice in important ways.

Nutritional rehabilitation is neurological rehabilitation. Because starvation directly impairs brain structure and function — degrading the very cognitive capacities needed for therapy to work — nutritional rehabilitation is not merely a medical priority; it is a neurological one. Psychotherapy delivered to a starving brain is less effective than psychotherapy delivered to a nourished brain. This understanding reinforces clinical guidelines that prioritize weight restoration and nutritional stabilization as foundational to recovery.

Habit-based models inform treatment design. The neuroscience of habit formation — particularly the shift from ventral (goal-directed) to dorsal (habitual) striatal control — has influenced treatments like habit reversal approaches and exposure-based interventions. If eating disorder behaviors become automatic habits encoded in the dorsal striatum, then treatment needs to do more than change beliefs and motivations; it needs to actively disrupt and replace habitual behavioral routines.

Transdiagnostic approaches are neuroscience-informed. Because eating disorders share neural circuitry with anxiety, OCD, and addiction, treatments that address these shared mechanisms — such as cognitive flexibility training, emotion regulation skills, and interoceptive exposure — have a neuroscience rationale that extends across diagnoses.

Pharmacotherapy remains limited but evolving. Currently, fluoxetine is the only FDA-approved medication specifically for an eating disorder (bulimia nervosa), and lisdexamfetamine is approved for binge eating disorder. No medications are currently approved for anorexia nervosa. Neuroscience research into specific receptor systems and circuit abnormalities is guiding the development of novel pharmacological targets, including glutamatergic agents, oxytocin, and drugs targeting ghrelin signaling.

Neuroimaging is not yet diagnostic. While brain scans reveal meaningful group-level differences, no neuroimaging biomarker can currently diagnose an eating disorder in an individual. Brain-based findings are best used to understand mechanisms and guide treatment development, not to replace clinical assessment.

Despite advances in neuroscience, several misconceptions persist:

• "Eating disorders are a choice." This is the most damaging misconception. Neuroscience demonstrates that eating disorders involve measurable brain differences in reward processing, interoception, cognitive control, and habit circuitry. No one chooses to have these neurobiological vulnerabilities. While behavior change is part of recovery, characterizing eating disorders as choices ignores their biological basis and increases stigma.
• "It's just about body image." Body dissatisfaction is a cultural risk factor and a common symptom, but eating disorders are driven by deeper disruptions in how the brain processes reward, detects internal signals, and forms habits. Many individuals with eating disorders, particularly those with binge eating disorder or avoidant/restrictive food intake disorder (ARFID), have minimal body image disturbance.
• "Brain scans can diagnose eating disorders." Neuroimaging research identifies patterns at the group level that illuminate underlying mechanisms. These findings are not reliable for individual-level diagnosis. Eating disorders are diagnosed through clinical evaluation based on DSM-5-TR criteria — patterns of behavior, cognition, and physical health.
• "If it's biological, therapy won't help." This reflects a false mind-body dichotomy. Psychotherapy changes the brain. Cognitive-behavioral therapy, family-based treatment, and other evidence-based approaches produce measurable changes in brain structure and function. Recognizing the neuroscience of eating disorders does not diminish the importance of psychological treatment — it reinforces why comprehensive treatment addressing both brain and behavior is essential.
• "Only young, thin, white women get eating disorders." Eating disorders affect people across all demographics — all genders, ages, races, ethnicities, body sizes, and socioeconomic backgrounds. The neurobiology of these conditions is not confined to any demographic group. This misconception leads to systematic under-diagnosis in men, people of color, individuals in larger bodies, and older adults.

If you or someone you know is experiencing patterns consistent with an eating disorder — such as persistent restriction of food intake, recurrent binge eating, compensatory behaviors like purging or excessive exercise, intense preoccupation with weight or shape, or a noticeable disconnect between eating behavior and hunger/fullness signals — it is important to seek professional evaluation.

Eating disorders are serious medical and psychiatric conditions with significant health consequences, including cardiovascular complications, electrolyte imbalances, bone density loss, organ damage, and death. Early intervention is associated with better outcomes. The neuroscience is clear: the longer an eating disorder persists, the more deeply entrenched its neural pathways become, and the more starvation-related brain damage can accumulate.

A comprehensive evaluation should ideally include a mental health professional experienced in eating disorders, a medical provider for physical assessment, and often a registered dietitian specializing in eating disorder recovery. Treatment typically involves a combination of psychotherapy (such as cognitive-behavioral therapy for eating disorders, family-based treatment, or dialectical behavior therapy), nutritional rehabilitation, medical monitoring, and sometimes medication.

In the United States, the National Eating Disorders Association (NEDA) helpline (1-800-931-2237) and the Crisis Text Line (text "NEDA" to 741741) provide immediate support and referrals. For life-threatening emergencies, call 911 or go to the nearest emergency department.