Two meals can match on the label and differ in the body
A controlled experiment tests food processing beyond calories and macronutrients.
A nutrition label is useful, but it does not describe every property of a meal. A randomized crossover experiment asked whether nutritionally matched ultraprocessed and non-ultraprocessed meals produced different metabolic and brain responses.
The study involved 57 healthy-weight adults. In a crossover design, participants receive different conditions at different visits, allowing each person to contribute to both sides of the comparison. This can reduce some between-person differences, although it does not remove every source of bias or carryover.
What the experiment measured
The ultraprocessed meals produced greater insulin and energetic responses, with lower carbohydrate oxidation relative to the comparison meals. Differences in peak carbohydrate oxidation were associated with striatal responses to food cues.
Participants’ subjective food valuation did not differ between conditions. Certain brain-response relationships did. The abstract does not supply a single numerical effect that can fairly summarize all of these different physiological and imaging outcomes.
Why an acute result is not a long-term verdict
The experiment examined responses to meals, not years of eating. It did not establish future diabetes incidence, weight trajectories or lifespan. A short-term difference in insulin response is not automatically evidence of harm, and a brain-imaging result is not proof that a food is addictive.
The category “ultraprocessed” also contains many products. A result from matched experimental meals does not demonstrate that every food within that category has the same effect. Reproducibility across meals, populations and longer exposures matters.
What the study adds
The useful contribution is a direct experimental test of a question that observational nutrition studies cannot easily isolate: can processing-related properties matter even when calories and major nutrients are matched? The findings support further investigation rather than a universal clinical prescription.
For someone reading nutrition news, the methodological lesson travels well. Ask whether a study measured intake, weight, an acute biological response or a clinical event. Each outcome answers something different. An interesting mechanism should not silently become a claim about disease prevention.
For Arising Health, this belongs beside—not in place of—the clinical literature on dietary patterns and metabolic treatment. It is possible to take food structure seriously without suggesting that a nutrition label is useless or that one short experiment resolves all the choices at a grocery store.
Sources and evidence
Metabolic and neural responses to ultraprocessed foods: a randomized, controlled, crossover study
October 2026 journal issue · Acute randomized crossover experiment