Science for a Healthier Tomorrow
In 46 children of healthy weight, a bigger ultra-processed share of calories tracked a stronger reward response to pictures of food — while the one-year fat-gain link appears non-significant.
In 46 children of healthy weight, a larger ultra-processed share of calories tracked a stronger dorsal striatal response to food pictures — a prospective, observational cohort within the BrainChild study: dietary recall and one fMRI scan at baseline, body fat a year later, no randomization, no control group. The coefficient was 0.34 per standard deviation (95% CI 0.07 to 0.61); the ventral striatum did not move. One-year body fat landed on the significance line (0.29, 95% CI 0.00 to 0.57, P=.05), which the authors call non-significant.
What this means
More ultra-processed calories went with a stronger dorsal striatal response to food cues, and only a borderline one-year fat signal. Nothing shows the food changed a child’s brain or caused that gain: no randomization, one brain scan, and 46 children. Whether striatal response mediates, confounds or merely accompanies it is unresolved.
One region moved, the other did not
The children were 7 to 10 at baseline (mean 8.7), all of healthy weight, in Southern California, 61% Hispanic/Latino, mean BMI 16.4 kg/m2; 67.7% of calories came from ultra-processed food by NOVA, a processing-based classification. Dorsal striatal response to food versus non-food pictures rose per standard deviation of intake (0.34, 95% CI 0.07 to 0.61). The ventral striatum showed nothing (0.16, 95% CI -0.14 to 0.46), so the reward system as a whole did not shift.
The one-year fat question
For body-fat change the ultra-processed share did not reach statistical significance: 0.29 (95% CI 0.00 to 0.57, P=.05), an estimate the authors themselves call non-significant and urge caution over. Waist circumference was the exception (0.38, 95% CI 0.11 to 0.67, P=.01). The striatal response also tracked one-year BMI (0.32, 95% CI 0.02 to 0.62, P=.04) — an association in 46 children, not the headline result. Adding striatal response cut that body-fat coefficient by about 44% (0.29 to 0.16), the authors’ change-in-estimate mediation result: a statistical pathway, not a demonstrated mechanism.
What it cannot say
Direction stays open: the brain was scanned once. Adjusting for added sugar left the dorsal association at 0.28 (95% CI -0.03 to 0.58, P=.07), hinting the exposure is partly a sugar signal, and reproducibility is thin at 46 children (Yokum 2021, six prospective studies). De Souza and Christoforou, in a same-issue commentary, accept the diet measurement as careful but argue that modeling change in BMI while adjusting for baseline BMI can create artifacts, including regression to the mean, without disputing the direction. In adults, an inpatient crossover trial found 508 kcal/day more intake on ultra-processed diets (Hall 2019), a route this cohort cannot test (see Ultra-processed food and appetite in adults 18-25, testing-ultra-processed-food-and-fullness-hormone-responses). Nothing covers children who already have overweight or obesity, or adults.
Terms explained
- Dorsal striatum — a deep brain structure tied to habit and reward learning; the region that moved.
- Ventral striatum — its neighbor, tied to immediate reward; the null result.
- Beta per standard deviation — a standardized coefficient: a 0.34 SD difference in outcome per 1 SD of intake.
Sources
- Alves JM, Morgan E, Angelo B, Le K, Smith J, Chow T, Jann K, Xiang AH, Page KA. Ultraprocessed Food Intake, Striatal Response to Food Cues, and Adiposity Gain in Children. JAMA Network Open. 2026;9(10):e2640252. doi:10.1001/jamanetworkopen.2026.40252 · PMID: 42912392 · full text verified, open access (CC-BY), 2026-10-24
- de Souza RJ, Christoforou A. Ultraprocessed Foods, Neuroimaging, and Interpretation of Observational Evidence. JAMA Network Open. 2026. doi:10.1001/jamanetworkopen.2026.40258 · invited commentary · full text verified, 2026-10-24
- Hall KD, Ayuketah A, Brychta R, 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.e3. doi:10.1016/j.cmet.2019.05.008 · PMID: 31105044 · abstract verified, 2026-10-24
- Yokum S, Gearhardt AN, Stice E. In search of the most reproducible neural vulnerability factors that predict future weight gain: analyses of data from six prospective studies. Social Cognitive and Affective Neuroscience. 2021:nsab013. doi:10.1093/scan/nsab013 · abstract verified, 2026-10-24



