Dr. Barbara Cardoso did not set out to find a reason for people to feel guilty about snacking. She is a senior lecturer in nutrition and dietetics at Monash University, and the question her team was investigating was specific: Does the proportion of ultra-processed food in a person’s diet predict cognitive function, independent of whether that diet is otherwise healthy? The answer, published in April 2026 in Alzheimer’s & Dementia: Diagnosis, Assessment & Disease Monitoring, a journal of the Alzheimer’s Association, was yes. And the size of the effect was not theoretical. It was attached to a number anyone can picture.
“To put our findings in perspective, a 10 percent increase in UPFs is roughly equivalent to adding a standard packet of chips to your daily diet,” Cardoso said. For every increment of that size, the study found a measurable drop in attention and a measurable increase in dementia risk factors. Not in people eating poorly across the board. In people whose diets, by every conventional measure, looked fine.
The Study That Controlled for the Excuse
Every previous study linking ultra-processed food to cognitive decline faced the same objection: maybe it is not the processing. Maybe it is everything else that tends to come with a high-UPF diet, the lower fiber, the higher sugar, the lower fruit and vegetable intake, the general pattern of eating that nutritionists have warned about for decades. If someone eats a diet heavy in packaged snacks and ready meals, they are probably also skipping the foods that protect the brain. The UPF correlation might just be a proxy for an absence of good nutrition, not evidence that processing itself does anything.
Cardoso’s team, analyzing data from 2,192 dementia-free Australian adults aged 40 to 70 enrolled in the Healthy Brain Project, designed the study specifically to test this objection. Participants completed detailed food frequency questionnaires that allowed researchers to calculate not just how much ultra-processed food they ate, but how that intake interacted with overall diet quality, including adherence to the Mediterranean diet, the dietary pattern with the strongest existing evidence for protecting cognitive health.
The result eliminated the easy explanation. The relationship between UPF intake and attention held regardless of how well participants otherwise followed a Mediterranean-style diet. People eating plenty of vegetables, fish, olive oil, and whole grains, who also had a higher proportion of ultra-processed food in their daily calories, showed the same measurable decline in attention as people with less healthy diets overall. Cardoso told CNN that this association was not changed by adherence to the Mediterranean diet, indicating that the effect is linked to food processing rather than simply food displacement.
That distinction is the entire point of the study. Food displacement means UPFs are bad because they take the place of foods that would otherwise protect you. Food processing means something about the processing itself, independent of what it replaces, is doing something to the brain. The study’s design was built to separate these two explanations, and the data came down on the side of processing.
The Two Numbers That Matter
The study measured two outcomes: attention, using standardized cognitive tests, and a 20-year dementia risk score based on established risk factors, including cardiovascular health, blood pressure, and metabolic markers.
For attention, the research found that every 10 percent increase in the proportion of a participant’s diet made up of ultra-processed food was associated with a drop of approximately 0.05 points on standardized attention tests. For dementia risk, every 10 percent increase in UPF proportion added approximately 0.24 points to the validated 20-year dementia risk score.
These numbers sound small in isolation, and they are small for any single 10 percent shift. But the participants in this study were already consuming an average of approximately 41 percent of their daily energy from ultra-processed foods, closely matching the broader Australian national average of 42 percent. The scale that matters is not one packet of chips added to an otherwise low-UPF diet. It is the cumulative difference between someone whose diet sits at 30 percent UPF and someone whose diet sits at 60 percent UPF, a difference that, according to this data, compounds into a substantial gap in measured attention and dementia risk trajectory.
The study found no association between UPF intake and memory specifically. This matters because it sharpens what the finding actually is. Attention is not memory. Attention is the gatekeeping function, the capacity to filter relevant information from irrelevant information, to sustain focus on a task, to process incoming information at speed. Memory consolidation happens downstream of attention. A measurable decline in attention, even without a corresponding decline in memory, represents degradation in one of the foundational processes that everything else depends on.
What Counts as Ultra-Processed and Why That Definition Matters
The NOVA food classification system, developed by researchers at the University of São Paulo and used as the basis for this study’s UPF measurements, defines ultra-processed foods as industrial formulations made primarily from substances extracted from whole foods, oils, starches, sugars, and proteins combined with additives, including flavors, colors, emulsifiers, and preservatives, that are rarely or never used in home cooking. Soft drinks, packaged salty snacks, ready-made meals, mass-produced bread, breakfast cereals, and reconstituted meat products fall into this category.
The classification is not about calorie content, sugar content, or fat content in isolation. A product can be relatively low in sugar and still be classified as ultra-processed if its formulation relies on industrial extraction and synthetic additives rather than recognizable whole-food ingredients. This is why the Mediterranean diet control in Cardoso’s study matters so much. A person can follow Mediterranean dietary principles emphasizing vegetables, legumes, fish, and olive oil while still consuming a substantial proportion of their daily calories from ultra-processed bread, packaged snacks between meals, and convenience products that fill the gaps of a busy schedule. The study found that this gap-filling UPF intake carried a cognitive cost that the healthy core of the diet did not offset.
Dr. Euridice Martinez Steele from the University of São Paulo, one of the study’s co-authors and a leading researcher in the development of the NOVA classification system, brought to the analysis the framework that allows researchers to separate “what is this food made of” from “how was this food made.” The Cardoso study’s finding that how matters independent of the what is precisely the question that the classification system was designed to make answerable.
How This Connects to What Is Happening Inside the Gut
The mechanism connecting ultra-processed food to cognitive function is not fully established, but the directions of current research point toward a system that has been documented extensively in other contexts: the gut-brain axis.
Research mapping the effects of ultra-processed food on the gut microbiome has found that these foods consistently reduce populations of bacteria associated with gut barrier integrity and increase populations associated with inflammation. The gut microbiome communicates with the brain through multiple pathways, including the vagus nerve and the production of metabolites that influence neuroinflammation. If ultra-processed food consumption is driving low-grade systemic inflammation through changes in gut bacteria, and that inflammation extends to the brain, a cognitive effect that operates independently of conventional nutritional content becomes mechanistically plausible.
This does not mean the Cardoso study proves a gut-to-brain causal chain. The study is observational; it measures associations in a single population at a single point in time, and it cannot establish that UPF consumption causes the attention decline rather than being associated with some other factor that causes both. What it does is add a specific, quantified data point to a growing body of research in which the gut microbiome keeps appearing as the connective tissue between diet and outcomes that, on the surface, seem to have nothing to do with digestion.
The Steelman: What This Study Cannot Tell Us
The honest limitations of this research are significant, and Cardoso herself acknowledged the central one directly: the study could only show an association, not a direct cause and effect.
Cross-sectional studies measure a population at one point in time. They cannot establish the sequence of events. It is possible that people experiencing early, subtle declines in attention for reasons unrelated to diet gravitate toward ultra-processed foods because those foods require less planning, less preparation, and less of the executive function that early cognitive decline erodes first. In that scenario, the causal arrow would point in the opposite direction: declining attention leads to higher UPF consumption, not the other way around.
The study population was also demographically narrow. The 2,192 participants were predominantly white, middle-aged to older Australian adults. Whether the same relationship holds across different populations, age groups, genetic backgrounds, and dietary cultures has not been established by this study alone.
The effect sizes, while statistically significant across a population of this size, represent small shifts for any individual person at any single point in time. A 0.05-point drop on an attention test for a 10 percent increase in UPF proportion is not something an individual would notice in daily life. The significance of the finding lies in its consistency across a large sample and its independence from overall diet quality, not in its magnitude for any one person.
These limitations do not undermine the finding. They define what kind of finding it is: a population-level signal that food processing itself, independent of nutritional content, correlates with a measurable cognitive marker. That is exactly the kind of finding that drives further research, the kind that, in other domains, has eventually led to changes in dietary guidelines once enough independent studies converge on the same signal.
The Packet of Chips Framing Is the Point
Cardoso’s choice to frame the finding in terms of a single packet of chips was not incidental. It translates a statistical effect size of 0.05 points on an attention test per 10 percent UPF increase into something that maps onto an actual daily decision. Most people do not think in percentages of daily caloric intake. They think in terms of whether they grab a snack at 3 pm.
The same research that has documented how dollar stores expand into communities and become the primary food source available describes an environment in which the “extra packet of chips” framing is not really a choice in any meaningful sense for a significant portion of the population. When the available food environment is dominated by shelf-stable, ultra-processed products, the daily decision Cardoso’s framing implies should I eat this extra packet of chips is not the decision most people are actually making. The decision was made further upstream, based on what is stocked, what is affordable, and what is convenient in the place where someone buys food.
The broader pattern connecting corporate food strategy to consumer outcomes shows that the food environment itself is shaped by decisions made far from the kitchen table, decisions about packaging, formulation, distribution, and pricing that determine what ends up being the path of least resistance for a tired person at the end of a long day. Cardoso’s research adds a cognitive dimension to a body of evidence about what that path of least resistance costs.
What the Number 41 Percent Means
The participants in this study consumed approximately 41 percent of their daily energy from ultra-processed foods, a figure that closely tracks the broader Australian national average of 42 percent. This is not a study of an extreme population eating an unusual diet. It is a study of ordinary middle-aged and older adults eating roughly what their national population eats on average.
That ordinariness is what makes the finding land differently from research on extreme diets or rare conditions. The participants whose attention scores correlated with higher UPF intake were not outliers. They were, on average, the country’s typical adult. If the relationship Cardoso’s team identified holds, it describes something happening inside a population-wide pattern of eating, not a fringe behavior affecting a small subset of people.
The study did not measure whether the cognitive effects are reversible. It did not establish a threshold below which UPF consumption carries no measurable cognitive cost. What it established is that the relationship between food processing and attention exists at the levels of consumption that describe how most people in a developed country currently eat, and that the relationship persists even when the rest of the diet looks, by every conventional standard, healthy.
A packet of chips, added to a diet that otherwise includes vegetables, fish, and whole grains, registered in the data. Cardoso’s team measured it. The number was small. The number was also there, in a population of 2,192 people whose diets looked, on paper, like they were doing everything right.
