A Meal Is More Than a Label

Food, body signals, and what the research can tell us

The Arising · Nutrition · November 2026 · John Griffin, MD

Dinner rarely arrives as a scientific question. It arrives when people are tired, hungry, short of time, or looking forward to sitting down together.

Then a headline enters the kitchen. Processed food changes the brain. A particular diet affects cancer. The language is urgent, the conclusion sounds simple, and an ordinary meal becomes another decision that feels easy to get wrong.

The research is usually more interesting than the headline. It asks a narrower question, uses a particular comparison and measures something specific. Understanding those details does not require becoming a nutrition scientist. It begins with asking what actually happened.

This issue follows two very different experiments. One compared short-term responses to carefully matched meals in people. The other examined intestinal tumors in genetically susceptible mice. Their findings should not be merged into a single verdict about what everyone should eat. Together, they show why the ingredients, the way a food is made, the person or animal studied, and the timescale all matter. [1,2]

There is room here for curiosity without turning every meal into a test of personal discipline. The aim is to understand the evidence well enough to use it, rather than let a label make every decision.

What a nutrition label can tell us

A nutrition label gives useful information about a food's contents. It does not describe every feature of the food or every response the body might have to it.

Two meals can contain similar amounts of energy and major nutrients while differing in how their ingredients are put together and processed. That creates a question researchers can test: when familiar nutritional quantities are matched, do the body's responses still differ?

The word “ultraprocessed” is used to classify foods by features of their industrial formulation and processing. It is not another word for any food that has been cooked, frozen or packaged. It also groups many different products together. A study of two experimental meals cannot tell us that every product in the category behaves identically.

Hutelin and colleagues set out to compare an ultraprocessed meal with a non-ultraprocessed meal matched on nutritional properties. Their experiment examined immediate responses and relationships with brain responses to food cues. It did not follow people for years to see who developed diabetes or heart disease. [1]

The people behind the study number

The study included 57 healthy-weight adults altogether. That does not mean 57 people completed every part of the experiment. [1]

Thirty-two completed both meal-testing sessions, with insulin measurements available for 31. Fifty-two were included in the brain-imaging analysis, and 29 contributed to the combined analysis linking the metabolic and brain measurements. Those smaller groups matter when interpreting the results. [1]

In the meal experiment, participants tried both conditions on different visits, with the order assigned by chance. This is called a randomized crossover design. Each person helps provide their own comparison, reducing some of the differences that would arise if two entirely separate groups ate the meals.

The meals supplied about 300 calories, and researchers followed responses for three hours afterward. Although nutritional properties were matched, food and protein sources differed, and the people collecting and analyzing the data were not blinded to the conditions. The experiment therefore cannot isolate one additive or one processing step as the explanation. This was a close look at what happened after a meal, not a test of a complete eating pattern over weeks or months. [1]

What the body measurements showed

The ultraprocessed meal produced a greater insulin response and greater energy expenditure after eating, alongside lower carbohydrate oxidation. [1]

Those terms describe different parts of the body's handling of food. Insulin is a hormone involved in controlling blood sugar and moving nutrients into use or storage. Energy expenditure describes the energy the body uses. Carbohydrate oxidation means using carbohydrate as fuel. A change in one does not automatically tell us whether a meal is better or worse for long-term health.

The total blood-glucose response over the observation period did not differ significantly between the meal conditions. This means the analysis did not clearly distinguish a difference; it does not prove the responses were identical. That result belongs beside the insulin finding. The study did not show that these meals caused diabetes, or that a higher insulin response in this experiment established insulin resistance, a reduced response to insulin. [1]

This is a useful example of why a biological response is not a verdict. Researchers can detect a difference worth investigating without having shown that the difference causes harm.

What the brain measurements showed

The researchers also looked at brain responses to food cues, including activity in the striatum, a set of brain structures involved in processing rewards and helping guide behavior. Differences in carbohydrate use were associated with some of those responses. Participants' subjective valuation of the foods did not differ significantly between conditions. [1]

The brain-imaging session took place on a separate day. The relationship between the measurements was therefore not proof that eating the ultraprocessed meal immediately changed the brain in the way a dramatic headline might imply. Nor does a brain response to a food cue establish that a food is addictive. [1]

An association means measurements varied together. It is a clue about a possible connection, not a complete explanation of how one caused the other.

Why the experiment matters anyway

Keeping the limits visible does not make the experiment unimportant.

Nutrition research often asks people what they usually eat and then watches what happens to their health. That can reveal valuable patterns, but the people eating different diets may also differ in income, health, activity, access to care and many other ways. Carefully controlled meal experiments help researchers examine a more focused question.

Here, the findings support further investigation of food properties that a simple nutrient comparison might miss. They do not show that calorie or nutrient information is useless. They also do not provide a universal shopping rule. [1]

To move from this experiment to a claim about long-term health, more questions remain. Do the findings recur with other meals? Do they apply to other populations? What happens when people choose and eat foods over longer periods? Do differences in the body's responses lead to a change in an outcome that matters to people?

That last step is essential. A three-hour insulin response, a weight change over six months, and a diagnosis years later are not different names for the same result.

A different experiment: diet and tumors in mice

The second study, by Shay and colleagues, concerned ketogenic diets and intestinal tumor formation. It is important to put the species and the setting in the first sentence: these were experiments focused on small-intestinal adenomas in genetically susceptible mice, not a trial measuring cancer risk in people choosing a diet. [2]

A ketogenic diet greatly limits carbohydrate and supplies much of its energy from fat. It can change several things at once, including how the body uses fat and produces ketones, molecules that can serve as fuel. The diet's name alone does not tell us which of those changes caused a particular experimental result.

In the studied mouse models, the ketogenic diet increased intestinal tumor burden and shortened survival. The researchers then examined possible explanations rather than assuming the ketones were responsible. [2]

An intestinal adenoma is a growth arising from the intestine's lining. These animal models make particular steps in tumor development easier to study. Their genetic susceptibility is part of what the experiment tests, not a detail that can be removed when describing the result.

Following the pathway rather than the diet's name

The researchers changed the animals' ability to produce or use ketones. Those manipulations did not explain the diet-associated tumor effects. The stronger signal involved dietary fats and the pathways that use them as fuel. [2]

They examined intestinal stem cells, cells that help renew the gut lining. Removing a set of regulators called PPARs weakened the diet-associated increase in these cells' numbers and activity. PPARs help control how cells respond to and handle fats. [2]

They also disrupted CPT1A, a protein involved in using long-chain fats for energy. That limited adenoma formation specifically under the ketogenic diet. These experiments helped connect a dietary exposure to a cellular process in the studied models. [2]

This is the kind of work that can make a biological explanation more convincing: change one component and see whether the result changes. It still does not make a mouse pathway a personal dietary instruction.

What the human boundary means

The study did not estimate how much a ketogenic diet changes cancer risk for a person. It did not compare every version of such a diet, establish a safe or unsafe duration for human use, or test a universal recommendation to stop eating a particular way. [2]

It would also be a mistake to use “not ketones” as an endorsement of ketone supplements. Showing that one proposed mechanism did not explain an animal result is not the same as demonstrating that a supplement is safe or beneficial for people.

For someone with an inherited risk of intestinal cancer, an important illness or a medically supervised diet, the useful next conversation belongs with the team that knows their situation. A mouse finding can help researchers choose their next questions. It cannot replace the evidence and clinical context needed for that person's care.

The paper also includes an observational analysis of human blood samples. That does not turn it into a human ketogenic-diet trial or supply a causal estimate of cancer risk. Several authors disclose biotechnology interests, and one reports consulting for Nestlé. Such disclosures belong in the context of the work; they do not, by themselves, establish whether a finding is right or wrong. [2]

The scientific opportunity is to understand whether the pathway matters in humans, for whom, and under which dietary conditions. The answer cannot be read straight from the name of the diet.

Two studies, two kinds of insight

The meal experiment and the mouse experiment are not competing answers to one question.

The human study asks whether matched meals produce different short-term responses. The mouse study asks which biological pathways contribute to tumors under a particular diet in susceptible animals. One offers a controlled glimpse of human metabolism; the other tests mechanisms in a disease model. Neither alone supplies a complete eating plan. [1,2]

That distinction helps with nutrition news more broadly. Before deciding what a headline means for dinner, identify four things:

• Who was studied? People with particular characteristics, animals with a specific genetic change, or cells in a dish are different starting points.
• What was actually eaten? A diet name does not provide the ingredients, amounts, preparation or comparison meal.
• What changed, and over how long? A laboratory measurement after a meal is different from symptoms, daily function or illness over time.
• What decision does the evidence support? An interesting mechanism may justify further research without justifying a treatment or a rule for everyone.

These questions make the science easier to use without demanding a perfect answer before anyone can enjoy lunch.

Food still has to fit a life

A workable eating pattern involves more than the argument over which label wins.

Can the food be obtained, stored and prepared? Does it fit the person's budget, preferences, culture and time? Are there medical needs or restrictions that change what is appropriate? Is eating becoming a source of connection and nourishment, or another daily source of fear?

The two studies do not answer all those questions. They do remind us why a narrow experimental finding should not erase them.

A practical conversation about food can begin with what is already happening. Which meals are difficult to manage? What feels satisfying? What gets skipped because there is no time or energy? What advice has been given for a diagnosed condition? Those details help identify the actual problem before a new headline supplies a solution.

Food can be taken seriously without treating every product as a moral category. Research can help clarify choices, but it works best when it is connected to the circumstances in which people eat.

Three everyday situations

A shopper compares two packages with similar nutrition labels. The meal experiment gives them a reason to be curious about more than nutrient totals. It does not prove that either product in their hands will reproduce the experimental findings. The useful next step is to understand the actual foods and their place in a broader pattern, not assume that a study settled the choice between two untested products.

Someone sees a headline saying a ketogenic diet causes cancer. They look for who was studied and discover genetically susceptible mice. That changes the meaning of the headline. The experiment may be important for understanding a pathway, but it supplies no personal risk estimate. If the person is following a diet for a medical reason, they bring the question to their clinical team rather than making an abrupt change from the headline alone.

A busy family feels judged by the word “processed.” Their immediate challenge may be getting a workable meal on the table. These studies do not measure their care for one another or give them a score for effort. Understanding the research can create room for thoughtful choices without pretending time, cost and access have disappeared.

These are illustrative situations, not reports of actual participants. They show how a finding can improve a question without issuing an order.

A place for curiosity at the table

The body responds to food in ways that nutrition research is still learning to describe. Food structure, fuel use and cellular pathways are worth understanding. So are pleasure, practicality and the different needs people bring to a meal.

The most useful research does not have to explain everything. It needs to explain clearly what was tested, what was found and how far that finding can travel.

A meal is more than a label. So is a life.

Sources and further reading

[1] Zach Hutelin, Monica Ahrens, Mary Elizabeth Baugh and colleagues. Metabolic and neural responses to ultraprocessed foods: a randomized, controlled, crossover study. Nature Metabolism. 2026;8(10):2152–2164. Primary publication record. Full text.

[2] Jessica E. S. Shay, Fangtao Chi, Constantine N. Tzouanas and colleagues. Ketogenic diet mediates intestinal tumorigenesis through lipids not ketones. Nature. 2026;656(8127):474–483. Primary publication record. Publisher record.

Continue exploring

Read Muscle & strength, GLP-1 & metabolic health and Fitness, or explore the magazine collection.