A UCSF Trial Cut Ultra-Processed Food by 85 Percent. Depression Scores Fell With It.

Twenty adults walked into a University of California, San Francisco trial with a depression diagnosis already on their chart, a high body mass index already on their chart, and in fourteen of the twenty cases, an anxiety diagnosis on top of it. For four weeks, each person changed almost nothing else about how they lived except one thing: they cut ultra-processed food out of their diet by more than 85 percent. Then the two study groups swapped diets, and the researchers watched what happened to their mood scores.

That specific test, removing ultra-processed food and nothing else from people who already had depression, had not really been run before. Most of what links processed food to mood is observational: track what people eat, track how they say they feel, find a correlation, argue for years about which way the arrow points. This trial, published September 23 in JAMA Psychiatry, tried to get closer to an actual answer by controlling the diet itself.

What the trial actually did

The design was a crossover: one group ate their normal diet for four weeks while the other switched to a low ultra-processed diet, then the two groups traded places for four more weeks. Every participant spent time on both sides. The average age was 44, 70 percent of participants were women, and everyone came in with an elevated BMI and at least one metabolic condition, hypertension, diabetes, or high LDL cholesterol. That is not a random sample of people who feel a little anxious. It is a group already living with depression and a body already under some metabolic strain.

• Four-week crossover design, each participant served as their own control
• Average participant age 44, 70 percent women
• All 20 participants had an elevated BMI and at least one metabolic condition
• 14 of 20 participants also carried an anxiety diagnosis
• Ultra-processed food intake fell more than 85 percent during the low-UPF phase

The gut theory

Corresponding author Nyasha Chagwedera, MD, PhD, an assistant professor at UCSF who founded the university's Metabolic and Mental Health Program, has a working explanation for why cutting the food, not adding anything else, moved the numbers. "Ultra-processed food triggers dopamine, the pleasure hormone, which promotes craving," she said. Her theory points at the gut lining next: "Junky processing may harm your gut in ways that ripple up to your brain," she said, describing a chain that researchers think runs from gut damage to low-grade inflammation to depression itself.

“The pilot successfully established the viability of carrying out this research in a larger-scale trial.”

Where this doesn't reach

Twenty people is a pilot, not a verdict, and the researchers say so themselves: the point of this round was to prove the study could be run at all, not to settle the question. There was no way to blind participants to which diet they were eating, so some of the improvement could reflect people's expectations rather than their biology. Everyone recruited already had depression and a metabolic condition, so the results say nothing about whether cutting processed food would do anything for someone who is metabolically healthy and not depressed. The team is not enrolling for a larger follow-up trial until 2028.

None of that makes the underlying link new. Marion Nestle, the New York University food-politics researcher, has pointed out there are literally hundreds of studies tying ultra-processed food to obesity, cancer, cardiovascular disease, and mortality. What is new here is the method: an interventional trial that isolated removing the food itself, in people who already had depression, rather than another dataset showing that people who eat more of it also report feeling worse. Observational correlation research has piled up for over a decade. Controlled tests of the actual mechanism are still rare.

That is the gap this pilot is aimed at, not a verdict on every clean-eating claim currently circulating online. It does not tell someone with depression to empty their pantry tomorrow. It tells a research team, and whoever funds the follow-up in 2028, that the experiment millions of people already run on themselves, cutting out packaged food and waiting to see how they feel, is specific and testable enough to be worth running properly, with a bigger group and a design built to rule out the placebo effect this one could not.