Short answer: Eating a lot of ultra-processed food is linked to faster cognitive decline and higher dementia risk. Large cohorts in Brazil and the UK show this, with risk rising as intake rises, but no trial has tested it, so we grade the evidence Moderate. A reasonable first step: replace sugary drinks, packaged snacks and processed meats with whole foods.

Diets high in ultra-processed food (UPF) are linked to faster cognitive decline and a higher risk of dementia. The evidence comes from large observational studies that agree with each other, which makes it worth acting on, but it can’t prove that UPF is the cause.

You don’t need to cut out every packaged food. The categories that stand out are sugary drinks, packaged snacks, processed meats and mass-produced baked goods. Start by swapping those for water, eggs, oats, nuts, fruit, fish and beans.

UPF now makes up about 58% of calories in the US, in a national survey from 2009–2010 (Martínez Steele 2016), and around 57% in the UK (Rauber 2018).

What counts as ultra-processed food?

The most widely used way to sort food by processing is the NOVA system. It was developed by researchers at the University of São Paulo and is used in reports by the Food and Agriculture Organization of the United Nations. It has four groups:

  • Group 1, unprocessed or minimally processed: fresh fruit, vegetables, eggs, plain milk, beans, nuts, fresh meat or fish. These may be dried, frozen, pasteurized or fermented.
  • Group 2, cooking ingredients: olive oil, butter, sugar, salt, flour and vinegar. People rarely eat these on their own.
  • Group 3, processed foods: Group 1 foods with Group 2 ingredients added, by canning, bottling or baking. Canned beans in salt water, artisan bread, simple cheeses and cured meats. They usually have two or three ingredients.
  • Group 4, ultra-processed foods: industrial products made mostly from substances extracted from foods, plus additives, with little or no whole food left. Soft drinks, packaged snacks, mass-produced bread, instant noodles, reconstituted meat and most breakfast cereals.

Almost all food is processed in some way. What sets Group 4 apart is industrial reformulation. These products contain things you wouldn’t find in a home kitchen: high-fructose corn syrup, hydrogenated oils, protein isolates, emulsifiers, humectants (additives that keep food moist), flavor enhancers and colorings.

The evidence linking UPF to cognitive decline and dementia

The Brazilian study: faster decline

In the ELSA-Brasil cohort, Gomes Gonçalves and colleagues (2023) followed 10,775 Brazilian public servants aged 35 to 74 for a median of eight years. They measured diet and tested thinking skills at the start and at follow-up.

People whose share of calories from UPF was above the lowest quarter of the group had a 28% faster rate of overall cognitive decline. Their executive function (planning, focus and self-control) declined 25% faster than in the lowest quarter. The analysis adjusted for demographic, lifestyle and health factors.

The study’s strength is that it tracked change over time in a large group. It shows that people who ate more UPF declined faster, not only that they scored lower at one point. As with any observational study, it can’t show that UPF caused the difference.

The UK Biobank: higher dementia risk

Li and colleagues (2022) studied 72,083 UK Biobank participants aged 55 and older, none with dementia at the start. Over a median of ten years, 518 developed dementia, including 287 cases of Alzheimer’s disease and 119 of vascular dementia.

Each 10% increase in UPF as a share of daily food (by weight) was linked to a 25% higher dementia risk. Replacing 10% of UPF with unprocessed or minimally processed food was linked to a 19% lower risk. These links held after adjusting for social factors, lifestyle and diet quality.

The replacement figure is a statistical estimate from the same observational data. It suggests what might happen if people swapped UPF for whole foods, but it isn’t a trial and can’t show cause and effect.

Other studies

An umbrella review of 45 pooled analyses (Lane 2024) linked higher UPF intake to 32 health outcomes. The most convincing evidence was for cardiovascular death, type 2 diabetes, anxiety and common mental disorders. All of the underlying studies were observational.

A French study followed 26,730 adults for about five years (Adjibade 2019). Each 10% increase in the share of UPF in the diet was linked to a 21% higher risk of depressive symptoms. Depression is closely linked to poorer thinking.

No randomized trial has tested cutting UPF over the long term with cognitive decline as the main outcome. Such a trial would be very hard to run. The evidence rests on observational findings that are consistent across countries and show more risk with more intake. That is why we grade it Moderate rather than Strong.

How ultra-processed food might affect the brain

Several pathways probably interact. Much of this evidence comes from animal and lab studies.

Inflammation

Long-term, low-grade inflammation in the body is one of the best-documented drivers of poorer thinking (see our guide to neuroinflammation and diet). Several common UPF ingredients activate inflammatory pathways, mostly in animal and laboratory studies. These include trans fats, emulsifiers such as carboxymethylcellulose and polysorbate-80, excess omega-6 fats from industrial seed oils, and AGEs (compounds formed when food is cooked at high heat).

Inflammatory signals from the body can reach the brain and activate microglia, the brain’s resident immune cells. When microglia stay switched on, they release substances that damage neurons and synapses (Heneka 2015). This kind of chronic microglial activation is a hallmark of Alzheimer’s disease and age-related cognitive decline.

Blood sugar

UPF tends to be high in fast-digesting carbohydrates and low in the fiber, protein and fat that slow glucose absorption. The result is bigger blood sugar spikes after meals, followed by dips.

Glucose matters for the brain even below the diabetes range. Mortby 2013 studied 210 Australians aged 68–73 without diabetes. Higher fasting blood glucose within the normal range was linked to smaller volumes in parts of the frontal cortex and poorer cognitive performance. This was a single observational study of average glucose, not of spikes.

Gut bacteria

UPF-heavy diets may change gut bacteria in unhelpful ways, though much of this evidence comes from animal studies. Emulsifiers, artificial sweeteners, preservatives and a lack of fiber reduce microbial diversity and make the gut lining leakier. They also cut production of short-chain fatty acids such as butyrate, which have anti-inflammatory and neuroprotective effects (Dalile 2019). Our gut-brain axis guide explains this link.

Crowding out better food

The simplest pathway may be the most important. If half or more of your calories come from UPF, you eat less of the foods that supply omega-3 fats, polyphenols, choline, magnesium and B vitamins. Supplements don’t easily make up for this. The benefits seen in diet pattern studies such as the Mediterranean and MIND diets seem to come from many nutrients eaten together in whole foods.

Which ultra-processed foods to cut first

The evidence for ranking UPF types is still developing, but four categories stand out.

  • Sugary drinks: soft drinks, energy drinks, sweetened juices and sweetened teas and coffees deliver fast sugar with no fiber and almost no micronutrients. In a cross-sectional analysis of about 4,000 Framingham Heart Study participants (Pase 2017), more sugary drinks were linked to lower total brain volume and poorer episodic memory. Daily fruit juice was also linked to a smaller hippocampus, a brain region key to memory. A study at one point in time can’t show which came first.
  • Packaged snacks: crisps, flavored crackers, extruded snacks and sweets combine refined carbohydrates, industrial fats and lots of salt. They are designed to override normal fullness signals.
  • Processed meats: hot dogs, nuggets and sausages. In about 494,000 UK Biobank participants followed for eight years (Zhang 2021), each extra 25 g a day of processed meat was linked to a 44% higher risk of dementia. These products are high in sodium, nitrates and saturated fat.
  • Mass-produced baked goods: industrial bread, pastries, cakes and biscuits often contain hydrogenated oils, high-fructose corn syrup and emulsifiers, on a refined flour base with little fiber.

How to eat less ultra-processed food

The major studies suggest the highest risk comes with the highest intake, and partial reductions are linked to benefit. Both the Brazilian and UK studies found cognitive differences at levels well above zero. A diet that is 20% ultra-processed is very different from one that is 55%.

Check the ingredient list. If a product contains things you wouldn’t find in a home kitchen, such as emulsifiers, flavor enhancers, maltodextrin, modified starches or protein isolates, it’s ultra-processed. A peanut butter with peanuts, salt and oil is processed. One with fifteen ingredients, including mono- and diglycerides and hydrogenated rapeseed oil, is ultra-processed.

Swap rather than ban. Focus on the highest-impact changes:

  • Drinks: sugary drinks to water, tea or black coffee. This is among the easiest changes to make.
  • Breakfast: packaged cereals to oats, eggs, or yogurt with fruit and nuts. Most commercial cereals count as UPF, even “whole grain” ones.
  • Lunch: ready meals and packaged sandwiches to quick assemblies from canned fish, pre-washed greens, hummus and bakery wholegrain bread.
  • Snacks: nuts, fruit, cheese or dark chocolate instead of packaged snacks.

Cook simply. Preparing more of your own food is the most effective way to cut UPF. Five to ten easy meals are enough for most people: scrambled eggs with greens, a bean and vegetable stew, baked salmon with roasted vegetables, a grain bowl. Batch cooking at the weekend gives you ready options for the week.

Aim for about 80/20. Getting about 80% of your calories from minimally processed foods is a realistic target. The goal is a lasting shift in your overall pattern, not purity.

What to eat instead

The foods most often linked to better brain health are ordinary staples:

  • Fatty fish: salmon, mackerel, sardines and anchovies, two to three servings a week, for omega-3 fats, especially DHA. Try our omega-3 power bowl.
  • Vegetables and leafy greens: as much variety as you can, including deeply colored options like spinach, kale, broccoli, peppers and sweet potatoes.
  • Berries: blueberries, strawberries and blackberries are among the most studied foods for brain health and are rich in polyphenols.
  • Nuts and seeds: walnuts, almonds, flaxseeds and chia seeds supply healthy fats, magnesium, vitamin E and fiber.
  • Beans and lentils: slow-release carbohydrates, fiber and plant protein.
  • Eggs: one of the best food sources of choline, which the brain uses to make acetylcholine, a messenger involved in memory.
  • Olive oil: the main fat in the Mediterranean diet, rich in polyphenols with anti-inflammatory properties.
  • Fermented foods: yogurt, kefir, sauerkraut, kimchi and miso support gut microbial diversity.

This list largely overlaps with the Mediterranean diet and the MIND diet, the two eating patterns most studied for brain health.

Frequently asked questions

Can a better diet reverse damage from years of UPF?

The evidence is encouraging but incomplete. The UK Biobank replacement analysis (Li 2022) suggests the link is modifiable. In a sub-study of the PREDIMED trial (Valls-Pedret 2015), 447 older Spanish adults ate a Mediterranean diet with extra olive oil or nuts. After a median of about four years, they had better cognitive scores than a low-fat group. No trial has tested cutting UPF itself.

Is it the additives or the overall diet?

Probably both. Some emulsifiers and artificial sweeteners have been shown in controlled studies to disrupt gut bacteria and make the gut leakier. But much of the harm likely comes from what UPF-heavy diets lack: fiber, omega-3 fats, polyphenols and micronutrients. The whole package, not a single villain, is the likely explanation.

Are “healthy” ultra-processed foods like protein bars OK?

They beat the worst offenders but don’t replace whole foods. A protein bar usually brings emulsifiers, sweeteners and refined carbohydrates that Greek yogurt with nuts does not. Fortified cereals add back a few vitamins alongside fast-digesting starch and additives. Use them as occasional conveniences, not staples.

Sources

  • Adjibade, M., et al. (2019). Prospective association between ultra-processed food consumption and incident depressive symptoms in the French NutriNet-Santé cohort. BMC Medicine, 17(1), 78. doi:10.1186/s12916-019-1312-y
  • Dalile, B., Van Oudenhove, L., Vervliet, B., & Verbeke, K. (2019). The role of short-chain fatty acids in microbiota-gut-brain communication. Nature Reviews Gastroenterology & Hepatology, 16(8), 461-478. doi:10.1038/s41575-019-0157-3
  • Gomes Gonçalves, N., et al. (2023). Association between consumption of ultraprocessed foods and cognitive decline. JAMA Neurology, 80(2), 142-150. doi:10.1001/jamaneurol.2022.4397
  • Heneka, M.T., et al. (2015). Neuroinflammation in Alzheimer’s disease. The Lancet Neurology, 14(4), 388-405. doi:10.1016/S1474-4422(15)70016-5
  • Lane, M.M., et al. (2024). Ultra-processed food exposure and adverse health outcomes: umbrella review of epidemiological meta-analyses. BMJ, 384, e077310. doi:10.1136/bmj-2023-077310
  • Li, H., et al. (2022). Association of ultraprocessed food consumption with risk of dementia: a prospective cohort study. Neurology, 99(10), e1056-e1066. doi:10.1212/WNL.0000000000200871
  • Martínez Steele, E., Baraldi, L. G., Louzada, M. L., Moubarac, J. C., Mozaffarian, D., & Monteiro, C. A. (2016). Ultra-processed foods and added sugars in the US diet: evidence from a nationally representative cross-sectional study. BMJ Open, 6(3), e009892. doi:10.1136/bmjopen-2015-009892
  • Mortby, M.E., Janke, A. L., Anstey, K. J., Sachdev, P. S., & Cherbuin, N. (2013). High “normal” blood glucose is associated with decreased brain volume and cognitive performance in the 60s: the PATH through life study. PLoS One, 8(9), e73697. doi:10.1371/journal.pone.0073697
  • Pase, M.P., et al. (2017). Sugary beverage intake and preclinical Alzheimer’s disease in the community. Alzheimer’s & Dementia, 13(9), 955-964. doi:10.1016/j.jalz.2017.01.024
  • Rauber, F., et al. (2018). Ultra-processed food consumption and chronic non-communicable diseases-related dietary nutrient profile in the UK (2008-2014). Nutrients, 10(5), 587. doi:10.3390/nu10050587
  • Valls-Pedret, C., et al. (2015). Mediterranean diet and age-related cognitive decline: a randomized clinical trial. JAMA Internal Medicine, 175(7), 1094-1103. doi:10.1001/jamainternmed.2015.1668
  • Zhang, H., Greenwood, D. C., Risch, H. A., Bunce, D., Hardie, L. J., & Cade, J. E. (2021). Meat consumption and risk of incident dementia: cohort study of 493,888 UK Biobank participants. American Journal of Clinical Nutrition, 114(1), 175-184. doi:10.1093/ajcn/nqab028