A large new study suggests that diets high in ultra-processed foods are linked to increased risk of stroke and cognitive impairment in adults, independent of overall diet quality. The findings, published in Neurology, are based on data from over 30,000 participants in the United States and indicate that food processing itself may play an important role in brain health, beyond traditional measures of nutritional quality.
Ultra-processed foods make up the majority of calories consumed in the United States. These foods are often manufactured with ingredients not typically used in home cooking—such as emulsifiers, artificial flavors, and stabilizers—and are designed to be hyper-palatable and convenient. Examples include packaged snacks, sweetened beverages, processed meats, and ready-to-eat meals.
Past research has linked high intake of these foods to metabolic and cardiovascular problems like obesity, hypertension, and heart disease. Some studies have also suggested connections to poorer brain health, but it remained unclear whether these links were simply due to poor overall diet quality or whether the processing itself might be harmful.
To address this question, researchers analyzed whether ultra-processed food consumption predicted risk of stroke or cognitive decline independently of adherence to three well-established healthy eating patterns: the Mediterranean diet, the DASH diet, and the MIND diet.
“My research group has been interested in studying risk factors for stroke and cognitive impairment, including biomarkers that can be measured years in advance and predict subsequent neurologic disease,” said study author W. Taylor Kimberly, chief of the Division of Neurocritical Care at Mass General Hospital and an associate professor of neurology at Harvard Medical School.
“As part of that work, we identified new biomarkers of stroke, but the markers were independent of the usual risk factors such as high blood pressure, diabetes, smoking, and high cholesterol. That got us thinking about what these markers might be a reflection of. As we considered a variety of potential factors, we found that diet was by far the strongest. As we studied the role of diet and brain health in more detail, we then recognized that the degree of food processing was an independent predictor of stroke and cognitive impairment.”
The study was part of the REasons for Geographic and Racial Differences in Stroke (REGARDS) project, a long-term observational study designed to investigate why stroke rates are higher in some parts of the United States. It includes a diverse sample of Black and White adults aged 45 and older who were enrolled between 2003 and 2007.
Participants provided detailed dietary information using a standardized food frequency questionnaire. Foods were categorized using the NOVA classification system, which groups items by level of processing. The researchers focused on NOVA1 (unprocessed or minimally processed foods) and NOVA4 (ultra-processed foods), analyzing their intake as a percentage of total daily food weight.
Separate groups were followed over time for the development of cognitive impairment or stroke. The cognitive impairment cohort included 14,175 people who had no evidence of cognitive issues at the start of the study and who completed follow-up cognitive testing. The stroke cohort included 20,243 participants without a prior history of stroke.
The research team used statistical models that accounted for a wide range of variables—including age, sex, race, income, education, blood pressure, diabetes, physical activity, depression, alcohol use, and more—to isolate the relationship between food processing and neurological outcomes. They also accounted for adherence to the Mediterranean, DASH, and MIND diets to see whether the level of food processing added predictive value beyond traditional dietary patterns.
People who consumed more ultra-processed foods had higher rates of both stroke and cognitive impairment during the study period. For every 10% increase in the proportion of ultra-processed food in the diet (by weight), there was a 16% higher risk of cognitive impairment and an 8% higher risk of stroke, even after controlling for other health and lifestyle factors.
In contrast, a higher proportion of unprocessed or minimally processed foods was associated with lower risk: a 12% reduction in stroke risk and a similar reduction in the likelihood of cognitive impairment.
“Even relatively small differences in the consumption of ultra-processed foods (UPFs) were associated with a differing risk to one’s brain health,” Kimberly told PsyPost. “In other words, swapping one or two meals each week from a UPF meal into less processed foods can have a substantive change in risk of stroke and cognitive impairment. This is encouraging because these types of changes sound achievable, and that it could modify one’s brain health in the long term.”
Importantly, these associations were not explained away by the participants’ overall adherence to healthy diets. Even among people who followed the Mediterranean, DASH, or MIND diets, eating more ultra-processed foods was still linked to poorer outcomes. Adding ultra-processed food intake to models that already included these diet scores significantly improved the ability to predict who would go on to experience stroke or cognitive decline.
The study also revealed that the negative impact of ultra-processed foods on stroke risk was greater for Black participants than for White participants, though no such racial difference was observed for cognitive outcomes.
“Prior studies had shown that diet patterns are important for maintaining good brain health (e.g., eating fruits, vegetables, whole grains, etc),” Kimberly explained. “Our study adds to that knowledge by reporting that the degree of food processing is also important. It’s not just what type of food you eat, but how it’s processed and prepared that also matters.”
The findings could help refine public health recommendations and food policy. While the Mediterranean, DASH, and MIND diets remain effective strategies for supporting long-term brain health, the results suggest that reducing ultra-processed food intake may provide additional protection—even for those who already follow a healthy diet.
While the study’s design is robust and includes a large, diverse sample with long-term follow-up, it does have limitations. Because it is observational, it cannot prove that ultra-processed foods cause cognitive decline or stroke. There may be unmeasured factors that explain the relationship. “While we can show there is a strong and consistent association between ultra-processed food intake and brain health, our study does not prove causation.”
In addition, dietary intake was self-reported, which can introduce inaccuracies, and the classification of foods by processing level (using the NOVA system) has been criticized for being too broad or imprecise in some cases. The researchers also noted that results differed depending on whether they analyzed food intake by weight or by calories, highlighting the need for standardization in future research.
“We’re now studying whether participants who had a change in their diet over the observation period also had a change in their risk of subsequent cognitive impairment,” Kimberly said. “Although this is not the same as running a randomized clinical trial, it builds further evidence to suggest that changing one’s diet is associated with a change in risk for subsequent brain health outcomes.”
“Long term, we hope to define which types of ultra-processed foods are associated with the greatest risk. Ultra-processed foods have some notable advantages—convenience, shelf life—and it’s likely that not all ultra-processed foods confer the same risk. The next step is to figure that out, which will allow policy makers to make informed decisions in the future about our food supply.”
The study, “Associations Between Ultra-Processed Food Consumption and Adverse Brain Health Outcomes,” was authored by Varun M. Bhave, Carol R. Oladele, Zsuzsanna Ament, Naruchorn Kijpaisalratana, Alana C. Jones, Catharine A. Couch, Amit Patki, Ana-Lucia Garcia Guarniz, Aleena Bennett, Michael Crowe, Marguerite R. Irvin, and W. Taylor Kimberly.
