Categorizing UPFs: Researchers create tool to analyze addictive properties of foods

06 Jun 2024 --- Ultra-processed foods (UPFs) comprise 58% of the calories consumed in the US, Canada and the UK and are linked to a wide range of adverse health outcomes. Scientists from Fralin Biomedical Research Institute at Virginia Tech Carilion (VTC) studying UPFs created a tool to assess the properties that make these foods rewarding and reinforcing.

“It might seem strange to use pictures to assess foods, but we are interested in asking questions about what goes on in the brain when we are evaluating these foods,” Alexandra DeFeliceantonio, associate director of the Fralin Biomedical Research Institute’s Center for Health Behaviors Research at VTC, tells Nutrition Insight.

“The current most common tool to measure brain activity non-invasively in people is functional magnetic resonance imaging (fMRI) and you can’t eat in an fMRI scanner, but you can look at pictures.”

The study is based on the NOVA classification system, which groups foods into four categories based on their level of processing. The scale was developed by nutrition researchers at the University of São Paulo, Brazil while analyzing the rapidly rising obesity rates in the region. 

Matching macronutrients to visual aspects
The research, published in the journal Appetite, provides a collection of curated images of minimally processed and UPFs matched on 26 characteristics, including macronutrients, sodium, dietary fiber, calories, price and visual characteristics such as color and portion size.

“This is really a tool set to help us start to ask more in-depth questions on how processing, independent of nutrient content, affects health,” DeFeliceantonio explains.

“We matched the pictures on macronutrients (protein, carbohydrates, fats), fiber, calories and energy density. By surveying over 100 people, we also matched them based on estimated calories, estimated price, liking and familiarity. The only difference is the level of processing according to the NOVA scale,” she says.

The NOVA system assigns food to categories unprocessed or minimally processed (fresh fruit, legumes, plain yogurt), processed culinary ingredients (cooking oils, butter, salt), processed foods (cheese, canned vegetables, baked bread) and ultra-processed foods (soft drinks, flavored yogurt, processed meat, packaged bread, additives).

“There is very little experimental research on UPFs, and part of what’s been holding us back is better tools for measuring and assessing their effects. The more tools we can provide, the more we can learn,” says DiFeliceantonio.

Neuroimaging studiesThe visual tool allows scientists to analyze food through fMRI using pictures.
To develop the picture set, a team of psychologists, neuroscientists and registered dietitians selected foods to represent either minimally processed or ultra-processed foods. The foods were prepared in a lab, visually represented through professional photography and controlled for consistency. 

Researchers also gathered price, food weights and nutritional information, including calories, macronutrients, sodium and dietary fiber.

“With this food picture set we can start to infer that any differences between food pictures is due to the degree of food processing, and not all these other factors that we know are potentially impactful,” says Zach Hutelin, Ph.D. graduate student, translational biology, medicine and health and lead author of the study.

The Virginia Tech team makes the pictures and associated data accessible through the Virginia Tech Data Repository of the university libraries, allowing scientists to test hypotheses in behavioral economic and neuroimaging studies.

The images are being used with functional magnetic resonance imaging (MRI) to reveal associated brain activity, with the pictures isolating the effects of food processing from other characteristics. 

Meanwhile, the World Health Organisation (WHO) developed a data-driven dietary impact assessment (DIA) for policymakers, researchers and practitioners in Europe to answer questions about the health and environmental impacts of diets. 

Rating the images
Study participants rated images across various qualities to generate a final set of 28 pictures matched across 26 characteristics. To objectively measure NOVA classification, researchers recruited 67 nutrition professionals and asked them to classify the foods as minimally or ultra-processed.

“A major criticism of the NOVA scale is that it’s difficult to use or that foods are classified differently by different people. We found that people with education in nutrition generally agreed on the food classifications, providing some data that might not be a valid criticism,” says DiFeliceantonio.

The study was funded by a National Science Foundation graduate research fellowship, the National Institute of Diabetes and Digestive and Kidney Diseases of the National Institutes of Health and a grant from the Seale Innovation Fund, which supports innovative pilot research projects at the Fralin Biomedical Research Institute. 

DiFeliceantonio received a grant from the fund to investigate metabolic, neural and behavioral data to better understand how our brains process nutrient availability and food preferences.

In addition, a 30-plus-year US study found that higher consumption of most ultra-processed foods is linked to a “slightly higher” risk of death. The researchers found the strongest association for ready-to-eat meat, poultry and seafood-based products, sugary drinks, dairy-based desserts and highly processed breakfast foods. A higher diet quality mitigates these associated risks.

Furthermore, the latest WHO Europe report, based on data from 44 countries participating in the Health Behaviour in School-aged Children study, shows a rise in unhealthy eating habits, rates of overweight and obesity, and low levels of physical activity among young people — disproportionately affecting those from less affluent families. Based on the findings, WHO calls for urgent action to address these trends.

By Inga de Jong