Adults who eat more ultra-processed foods show measurably slower thinking on standardized cognitive tests, and the effect persists even when their broader diet scores well on established healthy-eating scales. That finding, drawn from multiple large cohort studies spanning thousands of participants in Australia and the United States, challenges the assumption that overall diet quality alone determines brain health. The data point to something specific about industrial food processing itself, separate from calorie counts or nutrient profiles, that tracks with weaker cognitive performance across attention, verbal fluency, and global cognition.
How processing quality, not just diet quality, predicts cognitive speed
The most striking aspect of recent findings is that high ultra-processed food intake appears to drag on cognitive function regardless of whether someone otherwise follows a Mediterranean, DASH, or MIND eating pattern. A study in the REGARDS cohort, which examined ultra-processed food consumption in relation to stroke and cognitive outcomes, found that the association between ultra-processed food intake and adverse brain outcomes held up after adjusting for all three of those diet-quality scores. In practical terms, a person could eat plenty of vegetables, whole grains, and fish, yet still face measurable cognitive consequences from the packaged snacks, sweetened beverages, or ready-to-eat meals filling the rest of their plate.
This distinction matters because most dietary guidance focuses on what to add, not what to subtract. Telling people to eat more leafy greens or lean protein does not address the separate signal coming from ultra-processed products. The REGARDS data suggest that food processing itself introduces something that healthy-pattern scores do not capture or offset.
One hypothesis gaining attention among researchers is that additives common in ultra-processed foods, such as emulsifiers, artificial sweeteners, and preservatives, alter gut-derived metabolites in ways that affect the brain. This idea draws support from the observation that the cognitive associations are independent of total energy density and broad nutrient composition. If the problem were simply excess calories or poor macronutrient ratios, adjusting for diet-quality scores should have weakened the link. It did not. That said, no study in the current evidence base has paired repeated dietary recalls with longitudinal biomarker data, such as inflammatory markers or neuroimaging, collected at the same time points. The additive-driven metabolite pathway remains plausible but unproven at the mechanistic level.
Cohort data linking ultra-processed intake to attention and dementia risk
A cross-sectional analysis of middle-aged Australian adults without dementia found that each 10% increase in ultra-processed food intake was associated with lower attention scores on the Cogstate Brief Battery, a validated tool used in dementia research. The study classified diets using the NOVA system, which groups foods by degree of industrial processing rather than nutrient content, and estimated dementia risk using the CAIDE score. The results showed a dose-response pattern: more ultra-processed food, worse attention performance.
The attention findings matter because they emerge in midlife, long before clinical dementia, suggesting that processing-heavy diets may erode cognitive efficiency while people are still working and otherwise functioning independently. The CAIDE risk estimates add a further warning signal that these subtle decrements may translate into higher long-term dementia risk, although the study design cannot prove causality.
A separate prospective analysis in the ASPREE/ALSOP cohort strengthened the case by adding a time dimension. Over a median follow-up of several years, older Australian adults who consumed four or more servings of ultra-processed food per day performed worse on the Modified Mini-Mental State Examination (3MS), the Controlled Oral Word Association Test (COWAT), and the Symbol Digit Modalities Test (SDMT) compared to those eating fewer than four servings daily. These tests measure global cognition, verbal fluency, and processing speed, respectively, meaning the deficit was not confined to a single cognitive domain.
Because the ASPREE/ALSOP analysis followed participants over time, it could examine how baseline diet related to later cognitive change. The consistent disadvantage across multiple tests suggests that ultra-processed intake is not merely associated with one narrow skill but with broader thinking abilities that support everyday tasks such as managing finances, following complex instructions, or sustaining focus in conversation.
In the United States, longitudinal data from the Health and Retirement Study tracked specific categories of ultra-processed food and subsequent cognitive impairment risk over roughly seven years. According to a Harvard T.H. Chan School of Public Health summary, adults in the highest ultra-processed food intake group faced a substantially higher dementia risk compared to those in the lowest group. That analysis also broke down which types of ultra-processed products carried the strongest associations, moving beyond a blanket category to examine individual food types such as processed meats, packaged sweets, and sugar-sweetened drinks.
While the precise risk estimates and category-level details come from that single U.S. dataset, the overall pattern echoes the Australian findings: higher consumption of industrially formulated products aligns with poorer cognitive trajectories. The convergence across continents and age groups, using different cognitive tests and dietary assessment tools, strengthens confidence that the signal is not a fluke of one study design or population.
Experimental evidence from a randomized inpatient crossover trial at the NIH Clinical Center adds another layer. When participants were given an ultra-processed diet matched for presented calories, macronutrients, sugar, sodium, and fiber against an unprocessed diet, they still consumed roughly 500 more kilocalories per day on the ultra-processed plan and gained weight over two weeks. While that trial measured energy intake rather than cognition directly, it demonstrated that ultra-processed foods drive overconsumption through mechanisms that nutrient matching alone does not neutralize. The same properties that push people to eat more-soft textures, rapid eating, hyper-palatability, and easy swallowability-could be relevant to the metabolic and inflammatory pathways suspected of affecting brain health.
Gaps in the evidence and what to watch next
Several important questions remain open. None of the primary studies measured additive-level exposure with enough granularity to isolate which specific ingredients, if any, drive the associations. NOVA categories treat a broad range of products as uniformly “ultra-processed,” even though a flavored yogurt, a plant-based meat analogue, and a packaged cookie may contain very different additive profiles. Future work will need more detailed ingredient coding and, ideally, objective biomarkers for common emulsifiers, colorants, and sweeteners.
Reverse causation is another concern. People in the early stages of cognitive decline might gravitate toward ready-to-eat foods because they are easier to prepare, which could exaggerate observed links between ultra-processed intake and poorer test scores. The Australian and U.S. cohorts attempted to limit this problem by excluding participants with baseline dementia and adjusting for education, income, and health conditions, but subtle preclinical changes are hard to rule out completely. Repeated cognitive testing starting earlier in adulthood, combined with time-updated diet data, would help clarify directionality.
Confounding by lifestyle and social factors also lingers in the background. Ultra-processed food consumption tends to cluster with other behaviors-less physical activity, more screen time, higher smoking rates-that can harm brain health. Statistical models can adjust for some of these variables, yet residual confounding is likely. Randomized trials that manipulate ultra-processed intake while holding overall nutrient quality constant, and then track short-term changes in cognition or neurophysiological markers, would offer stronger evidence, though they are logistically challenging and ethically constrained.
Mechanistic research is still at an early stage. Animal and cell studies hint that certain additives and processing byproducts may alter the gut microbiome, increase intestinal permeability, and trigger low-grade systemic inflammation that reaches the brain. Human cohort data linking ultra-processed diets to markers such as C-reactive protein and insulin resistance provide indirect support. However, no large study has yet integrated detailed dietary records, microbiome sequencing, metabolomics, and serial cognitive assessments in the same participants. Building such multimodal cohorts will be critical to move from correlation to plausible biological pathways.
Despite these gaps, the converging evidence base already has practical implications. For individuals, the studies suggest that improving diet quality is not just about adding more whole plant foods and healthy fats; it also means consciously trimming the share of calories that come from packaged, ready-to-heat, and fast-food items. For clinicians and public health agencies, the findings argue for updating dietary guidance to highlight processing level as a separate dimension of risk, alongside nutrients like sodium and added sugars.
At the policy level, clearer front-of-pack labeling that distinguishes minimally processed from ultra-processed products could help consumers make quicker decisions in real-world shopping environments. Support for cooking skills, workplace meal programs, and access to fresh ingredients may be just as important for brain health as campaigns focused on heart disease or diabetes. As longer-term and more mechanistic studies report results, recommendations can be refined. For now, the simplest message emerging from the data is that what happens to food before it reaches the plate appears to matter for how well the brain performs years down the line.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
