Is interoception the same as hunger?

No. Interoception is the broader nervous-system capacity to detect internal body states, which includes hunger, thirst, fullness, fatigue, thermal regulation, heart rate, bladder signals, and the distinction between emotional states and physical ones. Hunger is one channel within that system. Nancy Zucker's laboratory at Duke has documented that interoceptive accuracy varies substantially across individuals, and that low accuracy on the hunger channel can appear alongside normal or above-average accuracy on other channels. A person who can feel her heartbeat clearly may still be unable to reliably detect when she has not eaten for ten hours. The condition is channel-specific rather than global.

How do you build appetite when the hunger signal is absent?

The clinical answer, developed inside the low-interest track of CBT-AR by Jennifer Thomas and Kamryn Eddy, is that appetite is not built through waiting. It is built through structure. The protocol treats eating as a scheduled task, with meals and snacks timed to the clock rather than to hunger, while interoceptive awareness training runs alongside as a separate skills track. Over weeks and months, the combination produces two results that matter. First, nutritional intake stabilizes, which reverses the downstream consequences of chronic under-eating. Second, the interoceptive channel sometimes, though not always, comes online as the body's nutritional state becomes more predictable. When it does, hunger signals start to arrive on a schedule the patient can read. When it does not, the scheduled approach continues as the long-term adaptation, and the patient learns to treat eating as a task rather than a response to a signal that never arrives.

Why does low-interest ARFID overlap so often with autism in adults?

The overlap runs through the shared substrate of interoceptive processing. Sarah Garfinkel at University College London, Nancy Zucker at Duke, and colleagues across the autism and eating-disorder research networks have documented reduced interoceptive accuracy as a feature that appears in both populations at rates well above the general population baseline. In late-diagnosed autistic adults, particularly women who were missed in childhood because the diagnostic criteria were calibrated on male samples, lifelong low appetite often turns out to be one expression of a broader interoceptive pattern that also includes difficulty detecting fatigue, thirst, temperature change, and the physiological markers of emotion. The clinical finding that an adult presenting for ARFID evaluation screens positive for autism at rates several times the general population is consistent across clinic samples. The two conditions are not the same condition. They share a neurological floor.

I Forget to Eat: Low-Interest ARFID and the Interoception Problem

TL;DR: Low-interest ARFID, the third axis in the three-dimensional model developed by Jennifer Thomas and Kamryn Eddy, is not caused by willful restriction, by body-image distortion, or by suppressed appetite. It is produced by a measurable deficit in interoceptive accuracy, the nervous-system capacity to detect internal body states such as hunger. Rachel Bryant-Waugh identified the lack-of-interest presentation in the ARFID diagnostic development work that preceded the DSM-5 entry. Nancy Zucker at Duke has built the interoception research that explains why the signal is missing. The overlap with late-diagnosed autism is significant and clinically load-bearing.

The Alarm Her Partner Set

A twenty-nine-year-old doctoral student in cognitive science at the University of Pittsburgh notices, one Tuesday in October, that her partner has set three alarms on her phone. The alarms are labeled Breakfast, Lunch, and Dinner. Her partner has set them because two weeks earlier, after coming home from a six-hour lab meeting, the student realized she had not eaten since a granola bar at eight the previous morning. The realization was intellectual. It arrived through a calculation of elapsed hours, not through a bodily sensation. She was not dizzy, not shaky, not distracted. Her body had not sent a signal she could read. The granola bar had been more than twenty-four hours earlier. Her partner, who has watched this pattern across four years of their relationship and has stopped believing it will self-correct, has installed the alarms.

The student’s pediatric records, which she requested from her mother earlier that year, document a growth-chart note from age seven that reads “eats poorly; small appetite; parents report child never requests food.” Her college physical at nineteen listed her as underweight by two pounds. Her graduate school physical at twenty-five listed her as underweight by seven. Her current primary care physician ran thyroid, complete metabolic panel, and ferritin in the spring. Thyroid was normal. Ferritin was 14 ng/mL. The physician prescribed an iron supplement and suggested she eat more. The alarms were the mechanism by which her partner translated “eat more” into something a nervous system that does not register hunger could actually execute.

The Signal That Does Not Arrive

Interoception, the neurological capacity to detect internal body states, operates through a network that includes the vagus nerve, the insular cortex, and the anterior cingulate cortex. Nancy Zucker’s laboratory at Duke has built the most developed clinical research program in the United States on interoception and eating, documenting that accuracy on interoceptive tasks varies substantially across individuals, that accuracy is measurable through behavioral paradigms such as the heartbeat detection task, and that reduced accuracy predicts patterns of eating that the patient experiences as “not noticing” rather than as restricting. Sarah Garfinkel at University College London has extended this work across eating disorders and autism, documenting reduced interoceptive accuracy in both populations relative to neurotypical controls without an eating disorder.

The clinical translation for the graduate student is specific. Her body does produce the physiological markers of hunger. Gastric motility shifts when she has not eaten in six hours. Ghrelin rises. Blood glucose drops. The signals are being generated. What is reduced is the central processing that converts those signals into the conscious experience of wanting food. The experience, in a neurotypical system, is something like a pull toward the kitchen, a thought about what sounds good, a restlessness that resolves when the person eats. In her system, the signal is either not registering or is registering with such low volume that the cognitive tasks she is doing at any given moment override it without effort. She does not feel herself choosing work over food. She simply does not feel the food channel.

Bryant-Waugh and the Lack-of-Interest Presentation

Rachel Bryant-Waugh, working at Great Ormond Street Hospital in London through the 2000s and early 2010s, was one of the core clinicians whose research fed into the DSM-5 framing of ARFID. Her contribution, consolidated in the Pica, ARFID, and Rumination Disorder Interview that she developed with colleagues, was to insist that the ARFID diagnosis required a dimensional structure rather than a single syndrome model. One of the three dimensions she named was lack of interest in eating or food, which the DSM-5 entry preserved as one of three non-exclusive pathways to the diagnosis. Bryant-Waugh’s clinical descriptions of the lack-of-interest presentation emphasized that these patients often did not fit the other ARFID categories. They were not selective about textures. They were not fearful of choking or vomiting. They simply did not experience the motivational pull toward food that organizes most people’s relationship to eating.

The lack-of-interest pathway has been less studied than the sensory and fear axes, partly because the presentation is quieter and less likely to produce crisis referrals. A child with a five-food sensory repertoire generates pediatric visits. A child who eats a normal range of foods but eats small amounts and never requests food may go unnoticed until growth-chart drift or a laboratory finding prompts attention. Adults with the lack-of-interest presentation often go undiagnosed for decades, arriving at eating-disorder clinics in their thirties or forties through a primary care physician who has exhausted somatic explanations for low body weight or persistent iron deficiency.

Thomas, Eddy, and the Three-Dimensional Model

Jennifer Thomas and Kamryn Eddy’s 2017 paper in Current Psychiatry Reports, which operationalized the three-dimensional ARFID model, treats lack of interest as the third axis alongside sensory and fear. Their CBT-AR manual, published through Cambridge in 2018, includes a dedicated treatment module for the low-interest presentation that differs substantially from the sensory and fear protocols. The sensory module is built on graduated exposure to specific food properties. The fear module is built on interoceptive exposure to autonomic sensations. The low-interest module is built on scheduling, on interoceptive awareness training, and on the recognition that appetite may not come online in a way the patient can rely on.

The clinical rationale for the scheduling approach rests on a distinction that matters. A patient with a sensory presentation has a nervous system that is over-responding to inputs and can be trained to tolerate a wider range. A patient with a fear presentation has a nervous system that has installed a threat response to eating and can be trained to extinguish it. A patient with a lack-of-interest presentation has a nervous system that is under-responding to a specific interoceptive channel, and the training target is less about modifying the response and more about building external structures that compensate for the missing signal.

Hilde Bruch, Fifty Years Early

Hilde Bruch, whose 1973 book Eating Disorders: Obesity, Anorexia Nervosa, and the Person Within established much of the clinical vocabulary for how eating disorders intersect with self-experience, named the failure of interoceptive awareness as a central feature of eating pathology decades before the neuroscience caught up to her observation. Her clinical descriptions of patients who could not accurately identify hunger, satiety, or the physiological markers of emotion anticipated what Zucker and Garfinkel would later measure with behavioral paradigms and neuroimaging. The citation matters for late-diagnosed adults because it locates their experience in a long intellectual lineage. What the twenty-nine-year-old graduate student is living with has been described, in clinically specific terms, since at least 1973. What has changed is the measurability of the phenomenon and the treatment protocols built around it.

The Autism Overlap

The graduate student has not been evaluated for autism. Her partner, who works in a software role, has mentioned it three times across their four-year relationship, and she has deflected each time. Her primary care physician has not asked. Her graduate advisor has not asked. The pattern in her pediatric records, including the age-seven note about never requesting food, the childhood difficulty with social reciprocity that a teacher flagged in third grade and that her mother attributed to shyness, the narrow range of clothing fabrics she has tolerated since adolescence, and the lifelong difficulty with reading fatigue until she has crossed into exhaustion, maps onto the late-diagnosed autistic woman profile that Sally Ozonoff at UC Davis and others have worked to make visible inside a diagnostic system that missed her population for two decades.

The screening rate in adult ARFID clinics supports the overlap. Adults presenting for ARFID evaluation screen positive for autism at rates several times the general-population baseline, particularly on the low-interest axis. The mechanism that connects the two conditions runs through the interoceptive substrate rather than through the surface features of either diagnosis. An autistic adult with reduced interoceptive accuracy on the hunger channel will present as lacking interest in food. An ARFID patient on the low-interest axis will often also show the interoceptive pattern across non-hunger channels that autism research has documented. The conditions are not identical. They are connected at the neurological floor.

A clinician who recognizes the overlap adds an autism screening instrument to the ARFID evaluation. The Autism Quotient or the Ritvo Autism Asperger Diagnostic Scale-Revised can flag the pattern in an evaluation visit. A positive screen prompts a recommendation for a full autism diagnostic evaluation, which in Pittsburgh runs through the University of Pittsburgh Medical Center or through one of several private providers. The autism diagnosis does not change the ARFID treatment plan in its broad outlines, but it sharpens the treatment delivery. Literal language at intake. Sensory accommodations. Pacing that respects the autistic nervous system. Coordination with occupational therapy if the interoceptive profile extends beyond hunger.

What a Treatment Plan Actually Looks Like

For the graduate student, a treatment plan would include six components running in parallel. First, medical coordination with her primary care physician to monitor ferritin, vitamin D, B12, and weight on a quarterly schedule while the behavioral work is happening. Second, an external eating structure: meals and snacks scheduled to the clock, with content that does not require her to decide what to eat in the moment, because decision-making at low blood glucose is where the system fails. Third, interoceptive awareness training, which uses paradigms from Zucker’s work to help her detect and label internal states. The training does not assume the hunger channel will come online. It treats interoception as a skill that can be partially developed on the channels where the nervous system is capable of improvement. Fourth, an autism evaluation, which will recontextualize the broader pattern and allow for treatment delivery that accounts for the neurology she has. Fifth, a consultation with a nutritionist who can help her construct meals dense enough in iron and other nutrients that the scheduled eating does not require her to eat larger volumes than her body tolerates. Sixth, couples work with her partner, who has been running an unpaid case-management function for four years and needs a framework that allows the alarms to be a piece of clinical infrastructure rather than a private negotiation in the relationship.

The treatment window for this kind of plan is typically twelve to eighteen months, with the scheduled-eating piece continuing indefinitely because the underlying interoceptive deficit often does not fully resolve. The goal is not to produce a neurotypical relationship to hunger. The goal is to construct a reliable eating life on top of a nervous system that will not signal hunger in the way neurotypical nervous systems do. Her weight will stabilize. Her ferritin will correct. Her work will benefit from the fact that she is no longer running on caloric deficit through her six-hour lab meetings. Her partner will stop needing to set alarms because the alarms will have become part of the infrastructure of her day.

What the Alarms Are Telling You

The graduate student’s partner has accidentally built, through an iPhone setting and four years of patience, the external structure that the CBT-AR low-interest module teaches patients to construct deliberately. The alarms are not a sign that she is incapable of feeding herself. They are a sign that the specific interoceptive channel that tells most people to eat is either missing in her or arriving at a volume her central processing cannot register above the other signals competing for her attention. The clinical translation of the alarms, from a phenomenon her partner has been quietly maintaining to an intervention her clinical team can formalize, is most of what treatment would do for her. The rest is reframing her lifelong relationship to food as a neurological pattern with a name, a research literature, and a treatment protocol, rather than as a personal failing she has been apologizing for since she was a child whose parents noticed she never asked for lunch.

Further reading on the sensory axis of the three-dimensional ARFID model is available in the post on ARFID and sensory sensitivity. The broader pattern of adult ARFID, including the late-diagnosed autistic presentation, is covered in ARFID in adults and in ARFID and autism. A full treatment overview is available in ARFID treatment. The eating disorders assessment and the autism screening assessment can serve as a starting point for adults whose pattern resembles the one described above. The ARFID course covers the three-dimensional model in depth across eleven modules.

If the pattern in this post matches the one you have been living with, a consultation is the next step.