My Healthcare News & Research — October 5, 2026 · Ultra-Processed Meal Trial: More Insulin, Same Nutrients

On October 5, 2026, the journal Nature Metabolism published online a randomised crossover study from the Fralin Biomedical Research Institute at Virginia Tech in Roanoke, Virginia, that asked a narrow question: if an ultra-processed meal and a meal that is not ultra-processed carry the same calories, carbohydrate, fat, protein, fibre, salt and water, does the body handle them the same way? Thirty-two healthy-weight young adults ate each of two meals of about 300 calories on separate mornings while resting inside a sealed room that measures how much energy the body burns and which fuel it is burning. Blood was drawn before each meal and seven times in the three hours after it. On every nutritional measure the team matched, the two meals differed by 1.6% or less.

The answer, in this one experiment, was no. After the ultra-processed meal, blood insulin ran clearly higher from 40 to 120 minutes, even though total blood sugar over the three hours was the same after both meals. The body burned slightly more energy after the ultra-processed meal and made a smaller switch from burning fat to burning carbohydrate. On a separate day, brain scans in a larger group of 52 volunteers found that some brain responses to pictures of food differed between ultra-processed and other foods, even though people said they would pay the same for both kinds.

The kind of evidence matters. This is a single-meal laboratory study in healthy adults aged 18 to 45, with 32 people in the metabolic part, testing one pair of meals. It measures short-term body responses, not weight, disease or long-term health, and the authors write that it “was not designed to assess chronic outcomes.” Every figure below comes from the open-access full text of the paper, its figures and its trial registration, not from press coverage; where the coverage went further than the paper, this report says so.

Table of Contents

  1. At a Glance
  2. What Was Done: The Two Meals
  3. What Was Found, Number by Number
  4. What the Brain Scans Showed
  5. How Processing Might Change the Response
  6. How It Fits With Earlier Research
  7. What the Researchers and the Coverage Said
  8. What It Cannot Show, and What Remains Unknown
  9. Sources and Primary Documents
  10. Key Research Papers
  11. Connections

At a Glance

Back to Table of Contents

What Was Done: The Two Meals

Volunteers were recruited from the university and the surrounding community in Roanoke between June 2023 and December 2025; the meal sessions ran from June 2023 to March 2025, and participants were paid for their time. To take part, people had to be 18 to 45 years old with a body mass index between 18.5 and 25. People with food allergies or restricted diets (such as vegetarian or ketogenic diets), with metabolic, neurological or psychiatric disorders or taking medicines for them, who were pregnant, who used inhaled nicotine or recreational drugs (other than occasional cannabis), or who had an impaired sense of taste or smell were not eligible. On average, the volunteers’ usual diets drew 54.4% of their calories from ultra-processed foods, close to the US national average of 55.0% that the paper cites.

Each meal session followed the same script. Participants arrived in the morning after an overnight fast, having done no strenuous exercise the day before, and entered a small whole-room indirect calorimeter: a sealed room that measures the oxygen a person uses and the carbon dioxide they breathe out, from which energy burned and the fuel mix can be calculated. After a 50-minute baseline, they had ten minutes to eat the whole meal. Measurements then continued for three hours. Blood was drawn through an intravenous line 60 minutes before the meal and at 5, 20, 40, 60, 90, 120 and 180 minutes after it, for blood sugar (glucose) and insulin. Movement was tracked with wrist and ankle sensors, and people rated their hunger, fullness and thirst, and how much they liked and wanted the meal.

There was no fixed gap between the two sessions. Seven people had them within three days of each other, 14 between four days and two weeks, seven between 15 days and four weeks, and four more than four weeks apart. Because these sessions were demanding, only 32 of the 57 volunteers did them; all 57 did the brain-scan session on a separate day.

The two meals

The meals were assembled by a computer search through combinations and portion sizes of 28 foods from the team’s validated food-picture set, looking for the closest nutritional match between a meal that was not ultra-processed and one that was, using a standard research nutrient database. The paper classifies foods with the Nova system, which sorts foods into four groups by level of industrial processing; group 4 is ultra-processed. The “non-ultra-processed” side covered Nova groups 1 to 3, so it was not limited to unprocessed foods.

How closely they matched, non-ultra-processed versus ultra-processed, with the percentage difference the paper prints:

These are values calculated from the nutrient database, not measured in the laboratory. The glycaemic index figures, for example, come from database values rather than from testing each meal’s effect on blood sugar.

Back to Table of Contents

What Was Found, Number by Number

Most of the results below are reported as an “area under the curve” (AUC): the total rise above each person’s own starting level, added up over the three hours after the meal. The insulin analysis included 31 people; the others included 32.

Starting points were the same. Before the meals, blood sugar (P = 0.17), insulin (P = 0.97), resting energy burn (P = 0.95), the fuel-mix measures (P = 0.17 to 0.23) and movement (P = 0.10) did not differ between the two sessions, and neither did time since the last meal (P = 0.20) or calories eaten the day before (P = 0.20).

Eating and appetite. The ultra-processed meal took on average 1.5 minutes longer to eat (P < 0.001); the authors judge that gap “unlikely to produce a physiologically meaningful effect.” Liking (P = 0.55) and wanting (P = 0.47) of the two meals did not differ, and neither did the pattern of hunger (P = 0.56), fullness (P = 0.10) or thirst (P = 0.86) over the session.

Blood sugar. Total blood sugar over three hours was no different (P = 0.92; the 95% confidence interval for the difference ran from −738 to +813 units, spanning zero). The shape of the curve did differ (P = 0.001). Blood sugar rose faster after the non-ultra-processed meal and was higher at 20 minutes (P = 0.02 after correction for multiple tests). Both curves peaked at 40 minutes. Later, blood sugar after the ultra-processed meal was “trending higher at 120 and 180 min” and did not return toward the starting level in the same way, in the paper’s words. That later difference is described as a trend, not as a statistically significant result.

Insulin. Total insulin over three hours was higher after the ultra-processed meal (P < 0.001), which the paper calls “a markedly greater insulinaemic response.” Insulin levels were similar at 5 and 20 minutes, higher after the ultra-processed meal at 40, 60, 90 and 120 minutes (all P < 0.05), and back near the starting level after both meals at 180 minutes. In short, a similar overall blood-sugar load was met with more insulin after the ultra-processed meal.

Energy burned. Metabolic rate over three hours was higher after the ultra-processed meal (P = 0.02). Both curves peaked at about 40 minutes; after that, energy burn fell back more after the non-ultra-processed meal. The paper reports this as an area-under-the-curve difference and does not convert it into extra calories burned.

Which fuel was burned. The respiratory exchange ratio, which compares the carbon dioxide breathed out with the oxygen taken in and shows the mix of carbohydrate and fat being burned, was higher after the non-ultra-processed meal (P = 0.01). Carbohydrate burning showed a larger rise, peaking around minute 40, after the non-ultra-processed meal; its area under the curve was lower after the ultra-processed meal (P = 0.02). Fat burning showed the mirror image: its area under the curve was higher after the ultra-processed meal (P = 0.003), so fat burning stayed higher. The authors sum this up by saying that the ultra-processed meal “blunts the post-ingestive shift from fat to carbohydrate oxidation.” One figure legend in the paper words the carbohydrate result the other way round; the abstract, the main text and the plotted bars all agree with the direction given here.

Habitual diet. Adding each person’s usual intake of carbohydrate, fat and protein, and their usual share of calories from ultra-processed food, to the analyses had no effect on these results.

Back to Table of Contents

What the Brain Scans Showed

On a separate day, after about three hours without food, all 57 volunteers had functional MRI scans, which track changes in blood flow as a stand-in for brain activity; 52 were included in the analyses after exclusions for head movement, task performance and one equipment failure. During the scan they saw pictures of 28 foods, 14 ultra-processed and 14 not, and for each bid what they would pay for it, from $0 to $5. The picture set, published by the same team in 2024, was matched on visual features, nutrition and how people rate the foods.

These brain findings are associations, measured on a different day from the meals, in a subset of people. The paper tested six metabolic measures against the scans separately and states that “No additional correction for multiple comparisons across metabolic measures was applied.” Results were unchanged when hunger, fullness, thirst or body mass index were taken into account; age and sex were included in every model.

Back to Table of Contents

How Processing Might Change the Response

The study measured what happened, not why. The authors’ explanation, which they present as a hypothesis, centres on food structure: the same nutrients can be more or less available to the gut depending on how a food is built. Their discussion draws on several lines of earlier evidence:

The authors read the result itself in two ways. First, more insulin for a similar peak in blood sugar is “a pattern consistent with a relative reduction in insulin sensitivity.” Second, a delayed shift from burning fat to burning carbohydrate after a meal is also seen in insulin resistance and type 2 diabetes, where researchers call the slow switch between fuels “metabolic inflexibility.” The authors suggest that repeated exposure could be “a plausible physiological mechanism” behind the metabolic problems seen in population studies. That link is a proposal; healthy young adults in this study showed a short-term difference after one meal, not a disease.

For the brain results, the team builds on “flavour-nutrient learning,” the idea that the gut reports what a food delivered after it is eaten, and the brain’s reward system, including the striatum, gradually learns to value that food accordingly. On this view, foods that release their nutrients differently could, over many meals, come to be valued differently. This study did not follow anyone over time and cannot test the learning step.

Back to Table of Contents

How It Fits With Earlier Research

The 2019 NIH inpatient trial. In the best-known earlier experiment, 20 weight-stable adults lived at the US National Institutes of Health Clinical Center and ate an ultra-processed or an unprocessed diet for two weeks each, matched for the calories, energy density, macronutrients, sugar, sodium and fibre presented. They could eat as much or as little as they liked. On the ultra-processed diet they ate 508 ± 106 more calories a day (p = 0.0001), from extra carbohydrate and fat but not protein, gained 0.9 kg, and lost 0.9 kg on the unprocessed diet. That trial showed people eat more when meals are ultra-processed; it left open whether the body also handles the same food differently once it is eaten.

The 2025 UPDATE trial. In England, 55 adults with a body mass index of 25 to under 40 who usually ate at least half their calories from ultra-processed foods followed two eight-week diets, both built to the UK Eatwell Guide, one from minimally processed and one from ultra-processed foods, eating freely. Weight fell on both, by 2.06% on the minimally processed diet and 1.05% on the ultra-processed diet; the difference, −1.01 percentage points (95% CI −1.87 to −0.14), meaning greater weight loss on the minimally processed diet, was statistically significant (P = 0.024).

Population studies. A 2024 umbrella review in The BMJ brought together 45 pooled analyses from earlier meta-analyses, covering 9,888,373 people, and graded as “convincing” the link between more ultra-processed food and higher risk of type 2 diabetes (dose-response risk ratio 1.12, 95% CI 1.11 to 1.13), among other outcomes. These are observational associations. A 2025 Lancet review series argued that the totality of evidence supports ultra-processed diets as a driver of chronic disease and listed soft texture and “disrupted food matrices” among the drivers of overeating it describes. The site’s June 26, 2026 report, Ultra-Processed Food and an Estimated 3.4 Million Deaths, covers a modelling study of the global toll.

Where this study adds something. The feeding trials above let people eat freely, so differences in intake and differences in how the body responds are mixed together. Here each meal was a fixed 300 calories eaten in full, so overeating was taken out of the picture. The authors note that earlier studies reporting different metabolic rates after more and less processed meals were hard to interpret because those meals also differed in calories and macronutrients. The trade-off is scale and duration: one small meal, once, in 32 healthy young people.

Policy context. In July 2025 the US Food and Drug Administration and the US Department of Agriculture asked the public for information to help write the first uniform federal definition of ultra-processed foods; the site’s explainer, FDA and USDA Seek a Definition of Ultra-Processed Foods, reports that request and where it stands. This study, like most in the field, used the Nova system to decide which foods counted as ultra-processed.

Back to Table of Contents

What the Researchers and the Coverage Said

The university’s announcement, reproduced by the medical news site News-Medical on October 5, 2026, quoted the senior author, neuroscientist Alexandra DiFeliceantonio: “We know that the majority of ultraprocessed foods are high in fat, they’re high in sugar, they’re low in fiber, and they’re low in protein. But if we artificially hold all of those things constant, is there something about the processing that leads to a different outcome?” The first author, Zach Hutelin, said of the meals: “If these meals had a nutritional label, they would be practically identical.” On next steps, DiFeliceantonio said: “We want to examine specific factors, such as commercial additives or specific processing steps, that might lead to different metabolic responses.”

The same coverage quoted Carlos A. Monteiro of the University of São Paulo, whose work led to the Nova classification and who is listed in the paper as one of its peer reviewers, as telling Nature Metabolism: “This is a novel and important contribution, as the brain ultimately regulates eating behavior.” The journal’s companion research briefing is behind a paywall, so that quotation could be checked only as printed in the coverage.

Some statements in the coverage go further than the paper:

The coverage did correctly report the study’s own caveats: one pair of 300-calorie meals, young and healthy volunteers, different protein sources in the two meals, more additives in the ultra-processed meal, and no direct measurement of food structure.

Back to Table of Contents

What It Cannot Show, and What Remains Unknown

The authors list their own limitations, and the design adds others:

What remains unknown

Whether the same pattern appears with other food pairings, larger meals, mixed diets over weeks, or in people who already have insulin resistance; which feature of processing is responsible; whether higher insulin after such meals, repeated over years, contributes to the metabolic disease seen in population studies; and whether the brain differences reflect learning, as the authors propose. The trial registration lists blood glucose, insulin, energy expenditure, fuel use and food valuation among its outcomes, and the team has deposited its data and brain images in public repositories for others to examine.

Back to Table of Contents

Sources and Primary Documents

Every figure attributed to the study above was taken from the open-access full text, its Figure 1 (meal composition and results) and its Methods, and checked against the PubMed and Crossref records. Press coverage was used only for the quotations, which are labelled as such.

  1. Hutelin Z, Ahrens M, Baugh ME, Nartey E, Herald DL 3rd, Hanlon AL, DiFeliceantonio AG (2026). Metabolic and neural responses to ultraprocessed foods: a randomized, controlled, crossover study. Nature Metabolism, published online October 5, 2026. — DOI: 10.1038/s42255-026-01619-4 · PubMed PMID: 42834140
  2. Research Briefing (2026). Ultraprocessed foods evoke distinct metabolic and neural responses. Nature Metabolism, published online October 5, 2026 (subscription). — DOI: 10.1038/s42255-026-01618-5 · PubMed PMID: 42834141
  3. ClinicalTrials.gov. The Effect of Processing on Food Reward (official title: Metabolic and Physiological Effects of Processing on Food Reward Encoding). Status completed; 57 enrolled. — NCT06017986
  4. Hutelin Z (2026). Study data and code, Virginia Tech Data Repository. — DOI: 10.7294/32077803
  5. News-Medical (October 5, 2026), reproducing the Virginia Tech announcement. — Ultraprocessed meals produce distinct metabolic and brain responses

Back to Table of Contents

Key Research Papers

Background papers on ultra-processed food trials, food structure and insulin responses. Each entry was checked against PubMed and Crossref before it was written down.

  1. Hall KD, Ayuketah A, Brychta R, Cai H, Cassimatis T, Chen KY, et al. (2019). Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake. Cell Metabolism 30(1):67-77.e3. — PubMed PMID: 31105044
  2. Dicken SJ, Jassil FC, Brown A, Kalis M, Stanley C, Ranson C, et al. (2025). Ultraprocessed or minimally processed diets following healthy dietary guidelines on weight and cardiometabolic health: a randomized, crossover trial. Nature Medicine 31(10):3297-3308. — PubMed PMID: 40760353
  3. Lane MM, Gamage E, Du S, Ashtree DN, McGuinness AJ, Gauci S, et al. (2024). Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses. BMJ 384:e077310. — PubMed PMID: 38418082
  4. Monteiro CA, Louzada ML, Steele-Martinez E, Cannon G, Andrade GC, Baker P, et al. (2025). Ultra-processed foods and human health: the main thesis and the evidence. The Lancet 406(10520):2667-2684. — PubMed PMID: 41270766
  5. Monteiro CA, Cannon G, Levy RB, Moubarac JC, Louzada ML, Rauber F, et al. (2019). Ultra-processed foods: what they are and how to identify them. Public Health Nutrition 22(5):936-941. — PubMed PMID: 30744710
  6. Tosh SM, Chu Y (2015). Systematic review of the effect of processing of whole-grain oat cereals on glycaemic response. British Journal of Nutrition 114(8):1256-1262. — PubMed PMID: 26330200
  7. Cai M, Tejpal S, Tashkova M, Ryden P, Perez-Moral N, Saha S, et al. (2025). Upper-gastrointestinal tract metabolite profile regulates glycaemic and satiety responses to meals with contrasting structure: a pilot study. Nature Metabolism 7(7):1459-1475. — PubMed PMID: 40542296
  8. Wolever TM, Zurbau A, Koecher K, Au-Yeung F (2024). The Effect of Adding Protein to a Carbohydrate Meal on Postprandial Glucose and Insulin Responses: A Systematic Review and Meta-Analysis of Acute Controlled Feeding Trials. Journal of Nutrition 154(9):2640-2654. — PubMed PMID: 39019167
  9. Corpeleijn E, Saris WH, Blaak EE (2009). Metabolic flexibility in the development of insulin resistance and type 2 diabetes: effects of lifestyle. Obesity Reviews 10(2):178-193. — PubMed PMID: 19207879
  10. Hutelin Z, Ahrens M, Baugh ME, Oster ME, Hanlon AL, DiFeliceantonio AG (2024). Creation and validation of a NOVA scored picture set to evaluate ultra-processed foods. Appetite 198:107358. — PubMed PMID: 38621591

PubMed Topic Searches

  1. PubMed: Ultra-processed food and insulin in randomised trials
  2. PubMed: Food structure and post-meal blood sugar
  3. PubMed: Metabolic flexibility and insulin resistance
  4. PubMed: Flavour-nutrient learning and the striatum
  5. PubMed: Controlled feeding trials of ultra-processed food

Back to Table of Contents

Connections

Back to Table of Contents