Artificial Intelligence May Be Able to Help Determine the Best Diets for Everyone

This interview was edited and condensed to make it more concise.

CNN: There is already a lot information on healthy eating. What is the difference between your approach and others?

Holly Nicastro There is a lot to consider and no one-size fits all approach to staying healthy. NPH’s goal is to develop algorithms that can predict individual responses to food or dietary patterns using AI.

Because we are looking at many factors, which is not often done in nutrition science, our approach is unique. NPH will study how genes, biology, physiology, environment, lifestyle, psychology, and other social determinants of one’s health affect individual diet responses. We will also be studying the largest and most diverse group of participants in a precision nutrition study.

CNN: Who is your audience?

Nicastro NPH participants will be recruited from All of Us Research Program which is managed by NIH. All of Us invites one million Americans to join the effort to build the largest health database in American history. Many of the participants are from underrepresented groups in biomedical science. All of Us participants provided information via surveys, electronic health records and biological samples. NPH will be able to study factors like age, sex and race in the All of Us Research Program cohort.

CNN: What data do you collect and how do you analyze it?

Nicastro: PH includes three modules. Module 1 will collect information about all participants’ daily diets. Module 2 will see a subset from Module 1 eat three diets chosen by researchers. Module 3 will see a subset of Module 1 participants participate in a two week study at research centers, where their food will have been carefully monitored by researchers.

Each module will end with a meal challenge test. Participants will then fast for several hours and then eat a standard breakfast or drink to allow us to examine their blood sugar levels over the next few hours.

Mobile photography apps and wearable devices will be used to passively record information about what people eat. Continuous glucose monitors and accelerometers will be worn by participants to track physical activity, sleep, and sedentary time. Researchers will also examine biomarkers such as blood lipids, hormone levels, and stool microbiome. 

NPH will use AI to analyze data from All of Us and link it into the NPH study. This method offers unprecedented opportunities to advance nutrition research towards personalized nutrition. AI, unlike human researchers can quickly process large amounts of data and convert connections between data points into algorithms. These algorithms can predict how an individual will react to food and their dietary habits, taking into consideration genes, proteins, metabolism, environmental, and lifestyle factors.

CNN: Who is most likely to profit from precision nutrition?

NicastroSome people who are at high risk of developing diabetes or have difficulty controlling their blood sugar levels could see the first benefits. The skin-attached blood glucose monitors allow us to see how a person’s blood sugar changes after eating certain foods, food groups, and meals. We can then predict the responses based upon the individual’s characteristics. We can then develop customized plans to prevent big blood sugar swings.

We are also using precision nutrition methods to predict how other diets will affect our mood, cognition, and blood cholesterol levels.