The research team, led by Zach Hutelin and Alexandra DiFeliceantonio of Virginia Tech’s Fralin Biomedical Research Institute, recruited 57 healthy-weight adults between 18 and 45 years old with body mass indices in the normal range. Participants’ habitual diets contained roughly 54 percent ultraprocessed foods, mirroring the national average of about 55 percent, which the authors argue makes the sample representative of typical Western eating patterns. Thirty-two participants completed both metabolic testing sessions, while 52 completed the functional magnetic resonance imaging arm of the study, and everyone who finished the metabolic sessions also underwent brain scanning.

The metabolic experiment was deliberately rigorous. On two separate days, in randomized order, participants entered a whole-room indirect calorimeter for four hours and consumed a roughly 300-kilocalorie meal composed entirely of either ultraprocessed foods or minimally processed alternatives. The two meals were algorithmically matched on weight, energy, energy density, total and available carbohydrate, fat, protein, glycaemic index, glycaemic load, fibre, sodium and water, with deviations of no more than 1.6 percent between conditions. Participants had to finish each meal within ten minutes, and blood was drawn through an intravenous catheter at eight time points spanning from an hour before eating to three hours afterward.

The results were striking. Blood glucose area under the curve did not differ between conditions, exactly as expected given the matched glycaemic indices. But the insulin story was entirely different: the ultraprocessed meal evoked a markedly larger insulinaemic response, with insulin concentrations significantly elevated between 40 and 120 minutes after eating compared with the non-ultraprocessed condition. Post-ingestive metabolic rate was also greater after the ultraprocessed meal, while the respiratory exchange ratio was higher after the whole-food meal, indicating that the body partitioned substrates differently despite identical macronutrient loads. Carbohydrate oxidation rose more robustly after the non-ultraprocessed meal, and fat oxidation was correspondingly greater after the ultraprocessed one.

The authors interpret this pattern as a blunted post-ingestive shift from fat to carbohydrate oxidation following ultraprocessed food consumption, a dynamic that resembles the delayed substrate switching seen in insulin resistance and type 2 diabetes. They suggest a plausible physical mechanism: many ultraprocessed products, though eaten as solids, are engineered to fracture easily in the first bite and combine rapidly with saliva into a lubricated, semi-fluid bolus, accelerating digestion kinetics and increasing nutrient exposure at the proximal small intestine, the site of greatest absorption. Manufacturing techniques designed primarily to optimize oral sensory experience may therefore unintentionally alter downstream metabolic signalling, producing higher insulin output even when peak blood glucose looks the same.

Crucially, the study went beyond metabolism to ask whether these processing-related physiological differences are reflected in the brain. In a separate imaging session, participants viewed pictures of 14 ultraprocessed and 14 non-ultraprocessed foods that had been validated in an independent cohort to be matched on 26 characteristics, including nine visual properties, eleven nutritional characteristics and six perceptual attributes such as liking, familiarity and expected satiety. Participants then bid real money for each item in a Becker–DeGroot–Marschak auction inside the scanner, a task that reliably quantifies subjective food value.

When the researchers regressed the between-condition difference in peak carbohydrate oxidation onto brain responses to non-ultraprocessed versus ultraprocessed food cues, they found whole-brain-corrected associations in the left superior temporal gyrus and, within a predefined striatal region of interest, significant effects in the right caudate and left ventral striatum. A larger carbohydrate oxidation response to whole foods relative to ultraprocessed foods was negatively associated with the neural contrast, extending previous findings from sugar-sweetened beverages to whole foods and reinforcing the idea that carbohydrate metabolism specifically conveys nutrient information from gut to brain through flavour-nutrient learning.

Perhaps the most intriguing finding concerned subjective value. Willingness to pay did not differ between the two food categories after adjusting for perceived healthiness and frequency of consumption, yet the neural encoding of value diverged dramatically. Value signals in the left fusiform gyrus and lingual gyrus, regions of higher-order visual cortex that include the so-called fusiform food area, were positive for non-ultraprocessed items but negative for ultraprocessed ones. The same opposing pattern appeared in the right putamen and caudate, striatal regions central to reward valuation. In other words, even when participants bid the same amounts, their brains represented the worth of ultraprocessed and whole foods in fundamentally different ways, hinting that processing level engages valuation circuitry through mechanisms independent of macronutrient content.

The authors are careful about limitations. Protein sources differed between meals even though total protein was matched, which could have influenced insulin secretion; the study did not directly measure the physical and chemical structure of the food matrix beyond total fibre; and the modest 300-kilocalorie meal size may not generalize to larger loads. Eating rate also differed slightly, with the ultraprocessed meal consumed about 1.5 minutes more slowly, though the team argues this is unlikely to be physiologically meaningful. Because the study was acute and conducted in healthy-weight adults, it cannot establish whether repeated exposure to these altered post-ingestive responses causes chronic disease, only that a plausible mechanism exists.

Even with those caveats, the implications are considerable. The findings provide experimental support for the hypothesis that altered nutritional availability, driven by changes in food structure rather than nutrient labels, is a candidate mechanism linking ultraprocessed food consumption to both metabolic dysfunction and overconsumption. If the gut-to-brain signals that teach us the value of food are systematically amplified or distorted by industrial processing, then two products with identical nutrition facts panels may not be metabolically or neurologically equivalent after all. As regulators and health agencies worldwide grapple with how to respond to the ultraprocessed food era, this study suggests that the degree of processing itself, as captured by the Nova classification, carries physiological consequences that calories and macronutrients alone cannot explain.

Subject of Research: Effects of food processing level on post-ingestive metabolism and brain responses to nutritionally matched meals

Article Title: Metabolic and neural responses to ultraprocessed foods: a randomized, controlled, crossover study

Article References: Hutelin, Z., Ahrens, M., Baugh, M. E., Nartey, E., Herald, D. L., III, Hanlon, A. L., & DiFeliceantonio, A. G. (2026). Metabolic and neural responses to ultraprocessed foods: a randomized, controlled, crossover study. Nature Metabolism. https://doi.org/10.1038/s42255-026-01619-4

Image Credits: AI Generated

DOI: 10.1038/s42255-026-01619-4

Keywords: ultraprocessed foods, Nova classification, insulin response, indirect calorimetry, carbohydrate oxidation, striatum, fMRI, food reward, flavour-nutrient learning, metabolism, nutrition, gut-brain signalling

The framing of the new work draws heavily on a hypothesis articulated by Monteiro and colleagues in a 2025 review in The Lancet, which laid out the principal mechanisms by which ultraprocessed foods might damage health. Among the candidate explanations is the idea that these foods are not merely calorie-dense but are engineered in ways that interact with our metabolic and neural machinery in fundamentally distinct manners. The new study provides direct experimental support for that proposition, showing that the body’s immediate physiological reaction to an ultraprocessed meal differs measurably from its reaction to less-processed alternatives, and that these metabolic differences are mirrored in patterns of brain activity.

The experimental design matters here. Because the study was randomized, controlled and crossover, each participant experienced the different dietary conditions in varying orders, allowing the researchers to separate the effects of the foods themselves from individual differences in metabolism, preference or circumstance. That within-person comparison is the gold standard for questions of this kind, since metabolic responses to food vary enormously between individuals. By holding each person constant across conditions, the design isolates what the ultraprocessed foods themselves are doing to the body’s post-meal physiology.

What the researchers found is that ultraprocessed foods evoke a distinct metabolic signature. The acute responses that follow consumption, the cascade of hormonal and substrate changes that govern how the body processes an incoming meal, differ from those elicited by less-processed foods. This is a consequential finding because it shifts the debate away from a simple accounting of calories. Two meals with identical energy content may not be metabolically equivalent, and the new results suggest that the degree of processing, and not merely the macronutrient arithmetic on the label, shapes what the body does with the food it receives.

Equally striking is the neural half of the study. The investigators asked how the brain encodes food value, and they did so along two dimensions at once: the nutritional value of what is eaten and its monetary worth. This dual-axis approach reflects a growing recognition in the neuroscience of eating that the brain’s valuation systems do not treat food as a single undifferentiated reward. Instead, they integrate signals from multiple sources, including sensory input from the mouth and, crucially, post-ingestive signals rising from the gut that report on what the food is doing to the body’s metabolism once it has been swallowed.

That gut-to-brain framework was set out in a 2020 review by de Araujo, Schatzker and Small in the Annual Review of Psychology, which argued that food reward must be rethought to include the metabolic consequences of eating rather than taste alone. The new Nature Metabolism study operationalizes that idea. Its central figure shows a correlation between the brain’s response and the metabolic response to ultraprocessed foods, suggesting that the neural valuation of these foods is not independent of the body’s metabolic handling of them. In other words, what the gut tells the brain about a meal appears to be woven into how the brain values that meal.

There is important precedent for this kind of metabolic-neural coupling. A 2017 study published in Current Biology by Veldhuizen and colleagues demonstrated that the integration of sweet taste and metabolism determines carbohydrate reward, and that carbohydrate oxidation, the body’s actual burning of the sugar for fuel, was associated with neural activity in the nucleus accumbens, a key hub of the brain’s reward circuitry. That finding established that the brain’s reward regions are sensitive not just to the flavor of a carbohydrate but to its metabolic fate. The new study extends this logic into the ultraprocessed food domain, where engineered combinations of ingredients may produce especially potent metabolic signals for the brain to read.

Another strand of prior evidence concerns the notorious appeal of foods that combine fat and carbohydrate, a combination that whole foods rarely offer but that ultraprocessed products deliver routinely. Work by DiFeliceantonio and colleagues, published in Cell Metabolism in 2018, found supra-additive effects of combining fat and carbohydrate on food reward, meaning that the reward value of the combination exceeded the sum of its parts. That study also included observational findings supporting the idea that macronutrient composition influences how the brain encodes reward value independently of caloric load. The new results slot into this picture: ultraprocessed foods, with their carefully optimized macronutrient profiles, appear to engage neural valuation systems in a distinctive way, and that engagement tracks the metabolic response the foods provoke.

The significance of the work lies in how it connects the epidemiological and mechanistic levels of the ultraprocessed food debate. The large-scale observational studies can show associations between ultraprocessed food consumption and poor health, but they cannot by themselves explain why. Hypotheses have ranged from the energy density and hyperpalatability of these products to their additive and emulsifier content, to the speed with which they can be eaten. The new study contributes a piece of mechanistic evidence: the acute metabolic response to ultraprocessed foods is distinct, and that distinctiveness is visible in the brain’s response. This supports the hypothesis, outlined in the Lancet review, that the processing itself, through the metabolic and neural signals it generates, may be a driver of the health harms, rather than a mere bystander.

It is worth being precise about what the study does and does not establish. The findings concern acute responses, the short-term physiology and brain activity that follow consumption in a controlled experimental setting. Chronic disease outcomes, from obesity to cardiovascular disease, unfold over years and decades, and bridging the gap between a distinct post-meal metabolic signature and long-term health will require further work. The correlation between brain response and metabolic response reported in the study is just that, a correlation, and the direction of any causal influence between the two remains a question for future research. Nonetheless, the crossover design gives the acute findings unusual rigor, and the alignment of the results with a coherent pre-existing framework of gut-to-brain reward signalling gives them explanatory weight.

The broader implications reach into public health policy and food formulation alike. If the degree of processing changes the metabolic and neural conversation between gut and brain, then debates about food labeling, such as whether processing level deserves a place on nutrition panels alongside fat, sugar and salt, gain fresh scientific grounding. For the food industry, the results suggest that reformulation aimed solely at macronutrient targets may not neutralize the effects of processing if the engineered structure of the food itself drives distinct physiological responses. And for the millions of people trying to understand why ultraprocessed foods are so hard to stop eating, the study offers a compelling if sobering picture: these products may be speaking to the brain in a language that whole foods never use, a language written in metabolic signals that the brain’s reward circuitry has evolved to find irresistible. The new work, published as Hutelin and colleagues’ randomized, controlled, crossover study in Nature Metabolism, marks a step toward decoding that language, and with it, perhaps, toward designing foods and policies that work with our biology rather than against it.

Subject of Research: Acute metabolic and neural responses to ultraprocessed food consumption

Article Title: Ultraprocessed foods evoke distinct metabolic and neural responses

Article References: Ultraprocessed foods evoke distinct metabolic and neural responses. (2026). Nature Metabolism. https://doi.org/10.1038/s42255-026-01618-5

Image Credits: AI Generated

DOI: 10.1038/s42255-026-01618-5

Keywords: ultraprocessed foods, metabolism, food reward, gut-brain signalling, nucleus accumbens, nutrition, neuroscience, food processing, public health, crossover study, Nature Metabolism, macronutrients