Same calories, same carbs, same salt: the ultra-processed plate still pulled more insulin

Most of the case against ultra-processed food rests on two kinds of evidence. Large population studies link eating a lot of it to obesity, diabetes and other illness, and feeding trials show people eat more when their diet is ultra-processed. What neither settles is why. Ultra-processed foods also tend to be richer in calories, sugar, fat and salt, so it is hard to tell whether the processing itself matters or just the nutrients that come with it. A study published on 5 October 2026 in Nature Metabolism tried to take the nutrients out of the question.

A team at Virginia Tech's Fralin Biomedical Research Institute in Roanoke assembled two meals of about 300 kcal. The ultra-processed one, as the researchers described it to the press, was small portions of deli turkey, instant mashed potatoes, frosted wheat cereal, veggie chips, white bread, a chocolate chip cookie and a bite of peanut butter and jelly. The other used less processed foods such as egg, cheese, sliced banana and dried cranberries. The two plates were matched to within 1.6% on weight, calories, energy density, protein, fat, carbohydrate, glycaemic index and load, fibre, sodium and water. Thirty-two healthy-weight adults, average age 26, ate each meal on a separate morning in random order, finishing within ten minutes, and then spent three hours in a sealed room that measured every breath while a catheter drew blood eight times.

Blood sugar over the three hours came out almost identical. The shape differed a little: glucose rose faster after the less processed meal and was higher at 20 minutes, while after the ultra-processed meal it lingered and was slower to come back down. Insulin told a clearer story. It was similar for the first 20 minutes, then ran higher after the ultra-processed meal at 40, 60, 90 and 120 minutes, and the total insulin response over three hours was significantly larger. People also burned slightly more energy after the processed plate, and their bodies kept burning relatively more fat and less carbohydrate, as if the meal were being handled in a different way even though its nutrients were the same. Hunger, fullness and how much people liked or wanted the food did not differ.

On a separate day, 52 of the volunteers looked at pictures of 14 ultra-processed and 14 less processed foods in a brain scanner and bid up to $5 for each. They valued the two groups about equally, but people whose carbohydrate burning differed most between the two meals also showed the biggest differences in how reward areas of the brain, including the striatum, responded to the pictures. The authors read this as a possible route by which the body learns to prize these foods, but it is an association, not proof. The study has clear limits. It tested one small meal, once, in young healthy-weight volunteers. The protein came from different foods on each plate, which could itself change insulin, and the ultra-processed plate had more additives and a different food structure that was not measured. It cannot say whether the bigger insulin response, repeated meal after meal, leads to weight gain or diabetes. What it does show is that two meals with the same nutrition label don't necessarily land the same way in the body.