a table filled with food

Does Processing Change How Food Is Metabolized?

A small crossover matched two meals on calories and nutrients. Insulin, energy use and fuel choice still differed — an acute lab signal, not weight or disease.

Two matched meals produced different insulin and fuel-use responses depending only on how processed they were.

Two meals matched on calories and nutrients to within about 1.6% still produced different insulin, energy-burning and fuel-use responses in a small crossover study of healthy-weight young adults — an acute lab signal, not evidence about weight or disease.

What this means

In a small crossover feeding study, healthy-weight young adults ate two meals of about 300 kcal matched on calories, macronutrients, fiber, sodium and glycemic load to within about 1.6% — and their bodies responded differently depending only on how processed the meal was.

The ultraprocessed meal triggered a much larger insulin release and a higher energy-burning rate, with less carbohydrate and more fat burning; brain reward regions responded differently to food pictures. The boundary is the story: this is an acute three-hour lab measure in 18- to 45-year-olds, with 32 of the 57 enrolled contributing metabolic data and exploratory brain findings. It shows processing changes short-term physiology — not that ultraprocessed food causes weight gain or disease.

A meal built to isolate processing

Does the degree of processing change how the body handles a meal even when calories and nutrients are matched? A randomized, controlled, crossover feeding study with a neuroimaging arm, published in Nature Metabolism on October 5, 2026, was built to test that. Each participant ate two meals of about 300 kcal, in random order on separate days: a Nova-matched ultraprocessed meal and a minimally processed one. The two were matched on weight, energy, energy density, macronutrients, available carbohydrate, glycemic index and load, total fiber, sodium and water — all within less than 1.6% deviation.

The 57 people enrolled were healthy-weight adults aged 18 to 45 (BMI 18.5 to 25 kg/m2), recruited in Roanoke, Virginia, at the Virginia Tech Fralin Biomedical Research Institute. The group was 31.6% male, with a mean age of 26.21 (± 6.85) years, a mean BMI of 22.75 (± 1.87) kg/m2 and habitual intake of 54.4% (± 18.4%) of calories from ultraprocessed food. No one had metabolic disease or took relevant medications. The prespecified primary behavioral outcome was willingness to pay, tested in a separate fMRI session with a Becker-DeGroot-Marschak auction task; the metabolic session used 50 minutes of baseline plus three hours of whole-room indirect calorimetry, with serial blood draws. Registered as NCT06017986.

What changed — and what didn’t

Insulin was where the meals separated most sharply. Over the three hours after eating, blood insulin area under the curve was higher after the ultraprocessed meal (t(28.67) = 5.48; P<0.001; 95% CI −2254.02 to −1032.03), and insulin levels were higher between 40 and 120 minutes after the ultraprocessed meal (χ²(7) = 45.43; P < 0.001; insulin analysis n = 31).

Blood glucose did not follow. Glucose area under the curve did not differ between meals (t(27.15) = −0.11; P = 0.92; 95% CI −738.2 to 813.21), though glucose stayed elevated later after the ultraprocessed meal. The meals did not move every metric the same way.

Energy use and fuel choice did diverge. After the ultraprocessed meal, metabolic rate area under the curve was greater (t(30.73) = 2.39; P = 0.02; 95% CI −0.036 to −0.003). The respiratory exchange ratio — a marker of how much of the body’s fuel comes from carbohydrate rather than fat — was higher after the minimally processed meal (t(30.75) = −3.04; P = 0.01; 95% CI 0.004 to 0.022). Carbohydrate oxidation was lower after the ultraprocessed meal (t(30.76) = −2.47; P = 0.02; 95% CI 0.002 to 0.019), and fat oxidation was greater (t(30.84) = 3.26; P = 0.003; 95% CI −0.01 to −0.002).

The brain findings are exploratory

In the imaging session, activity in the striatum (caudate and ventral striatum) and the superior temporal gyrus while viewing food cues tracked the between-meal difference in peak carbohydrate oxidation. The superior temporal gyrus cluster reached whole-brain significance (MNI −44, −24, −8; t = 7.33; P_FWE = 0.013), as did small-volume-corrected clusters in the right caudate (t = 6.11; P_FWE = 0.007) and left ventral striatum (t = 5.20; P_FWE = 0.045), in 29 participants.

The prespecified primary outcome did not move: willingness to pay did not differ between meals (t(32.41) = 0.96; P = 0.34) after adjusting for healthiness and frequency. The study treats these brain-metabolism links as exploratory and ad hoc-powered, and their direction is inconsistent with some earlier work — the paper attributes that to visual versus oral cue tasks.

Why this is a signal, not a verdict

Several limits define what this study can say. It is acute: outcomes were measured over three hours after a single meal in a metabolic chamber, using surrogate markers rather than weight, disease or clinical endpoints. The population is narrow — healthy-weight adults aged 18 to 45, not children, older adults, or people with overweight or obesity. The metabolic sessions were completed by a subset of 32 of the 57 enrolled (the paper’s own CONSORT counts report the metabolic arm at n = 32 and the fMRI arm at n = 52), and the effect sizes rest on 29 to 32 participants. The meals also differed in protein source and additive content, food matrix was not measured beyond total fiber, and only one load of about 300 kcal was tested. The link from these acute differences to long-term health is hypothesized, not shown.

Where this sits

Two rigorous human crossover trials already showed that freely eating ultraprocessed diets raises energy intake and weight relative to minimally processed diets — an inpatient NIH trial in 20 people reported roughly 500 kilocalories a day more eaten and weight gain on the ultraprocessed diet, and an 8-week crossover in 55 UK adults found greater percent weight loss on minimally processed than ultraprocessed diets (difference in percent weight change −1.01; 95% CI −1.87 to −0.14; P = 0.024). Hall’s meals were matched for presented calories and macronutrients but still allowed large differences in eating rate and food form; Dicken’s diets followed national dietary-guidance nutrient targets rather than being matched at the individual-meal level. The paper presents this study as directly addressing the nutrient-confounding limitation of those ward diet studies, not as the first matched-meal demonstration — a mechanistic complement to the weight outcomes, not a replacement, and no long-term answer.

Nutritious News has covered adjacent questions, including a calorie-matched ultraprocessed-diet trial in men (upf-calorie-matched-men), a null satiety-hormone finding in lean young adults (testing-ultra-processed-food-and-fullness-hormone-responses) and a trial that cut ultraprocessed food in older adults (does-cutting-processed-foods-improve-health).

Terms explained

  • Ultraprocessed food (Nova category 4): the most heavily processed tier of the Nova classification — industrial formulations built largely from refined ingredients and additives rather than whole foods. Here the ultraprocessed meal was Nova-matched against a minimally processed meal.
  • Crossover trial: every participant completes each meal or diet being compared, in randomized order, on separate days — so each person serves as their own comparison.
  • Whole-room indirect calorimetry (metabolic chamber): a sealed room that measures oxygen use and carbon dioxide production to estimate how much energy the body is burning and which fuel it is using; here, over three hours after the meal.
  • Respiratory exchange ratio: the ratio of carbon dioxide produced to oxygen consumed. A higher value means more carbohydrate is being burned; a lower value means more fat.
  • Insulinemic response and area under the curve (AUC): the rise in the hormone insulin after eating, which moves glucose from the blood into cells; AUC summarizes a measurement over time — here, the total insulin, glucose or energy response across three hours.
  • Willingness to pay (BDM auction): an economic task asking how much someone would pay for a food, used as a proxy for how much they value it. It was this study’s prespecified primary outcome.

Sources

  1. Primary source: Hutelin Z, Ahrens M, Baugh ME, Nartey E, Herald DL, Hanlon AL, DiFeliceantonio AG. “Metabolic and neural responses to ultraprocessed foods: a randomized, controlled, crossover study.” Nature Metabolism, published 2026-10-05. DOI: 10.1038/s42255-026-01619-4 · PMID: 42834140. Registered trial: NCT06017986. Full-text access.
  2. Trial registration: ClinicalTrials.gov NCT06017986 (randomized, controlled, crossover feeding study; IRB 21-1052, Virginia Tech). https://clinicaltrials.gov/study/NCT06017986
  3. Preprint: Hutelin Z, et al. medRxiv preprint, posted 2026-04-11. DOI: 10.64898/2026.04.10.26350599.
  4. Context (prior human crossover): Hall KD, et al. “Ultra-Processed Diets Cause Excess Calorie Intake and Weight Gain: An Inpatient Randomized Controlled Trial of Ad Libitum Food Intake.” Cell Metabolism 2019;30(1):67–77. DOI: 10.1016/j.cmet.2019.05.008 · PMID: 31105044. Abstract-level access.
  5. Context (prior human crossover): Dicken SJ, et al. “Ultraprocessed or minimally processed diets following healthy dietary guidelines on weight and cardiometabolic health: a randomized, crossover trial.” Nature Medicine 2025;31(10). DOI: 10.1038/s41591-025-03842-0 · PMID: 40760353. Abstract-level access.
  6. Context (same lab, reward framework): “Supra-Additive Effects of Combining Fat and Carbohydrate on Food Reward.” Cell Metabolism 2018;28(1):33–44. DOI: 10.1016/j.cmet.2018.05.018. DiFeliceantonio AG, Coppin G, Rigoux L, Edwin Thanarajah S, Dagher A, Tittgemeyer M, Small DM. Snippet-level access. (Author list verified via Crossref 2026-10-06.)
  7. Funding: NIH/NIDDK DK132389; National Science Foundation 2235205. Competing interests (per the paper): “The authors declare no competing interests.”

Related Articles