Artificial Intelligence meets Nutrition - Interview with Maria Vasiloglou

Technology is drastically changing the world we live in. There are a growing number of health and nutrition apps, which are reaching more and more people with a new kind of nutrition communication. In this interview, we speak to Maria Vasiloglou, clinical dietitian and researcher on innovative technologies for dietary monitoring and assessment at the University of Bern, about how these tools can be used in nutrition practice and how we can bridge the gap between nutrition and technology. BTW: Personalization and digitalization in nutrition therapy are among our Top 10 nutrition trends of 2021.


speech by a young woman, Maria Vasiloglou

Maria, we are very curious to find out more about you and your research. What are your current research projects?

My journey started in the department of Nutrition and Dietetics (BSc) in Greece and then continued with a Master in Human Nutrition with a specialization in clinical nutrition from the University of Glasgow. After that, I did a traineeship in JRC of the European Commission where I mostly worked on science-based policy support in the field of nutrition, contributing to a foresight study and evaluating the fitness of the current food safety and nutrition regulatory system, and especially assessing the effect of merging innovation and technology in the food and pharmaceutical area. I returned back to Athens and worked for some years in an outpatients’ clinic of a private hospital, working with children and adolescents with diabetes type 1, eating disorders, obesity, and metabolic diseases while in parallel I was teaching nutrition-related topics to a vocational school. From 2017, I live in Bern Switzerland pursuing my Ph.D. in innovative technologies for dietary monitoring and assessment. I work with a team of engineers and computer scientists looking at our research hypothesis from different perspectives. I believe that multidisciplinary work is fundamental to create a complete system that meets user needs and at the same time can be useful for research purposes as well as clinical practice.


"I believe that multidisciplinary work is fundamental to create a complete system that meets user needs." - Maria Vasiloglou


You have worked on mobile health apps and investigated food imaging-based dietary assessment by utilizing these apps. What were the results and your experience?

In our lab, we have created an image-based app (goFOOD) that automatically translates meal images/video into nutrients with the use of artificial intelligence. Till now it is only used for research purposes and we intend to use and validate it in different clinical settings.


As a dietitian, I am mostly working on the user-perspective and clinical aspects of image-based apps. Thus, I was curious to gather healthcare professionals‘ opinions, who work with nutrition-related diseases, since they are the key players in recommending nutrition apps. This is why we conducted an international survey with 1001 participants from 73 countries and 6 continents. We found out that almost 46% of them have recommended nutrition apps to their clients or patients. Around a quarter of those who have not yet recommended an app, do not know of their existence! Another main area of research falls into the observation that major loss of data, and thus impaired quality of them, can be attributed to the common mistakes made by users of image-based nutrition apps.


"Healthcare professionals are the key players in recommending nutrition apps." - Maria Vasiloglou

I believe that this research is the most exciting so far. It is true that current research on apps is mainly focused on the optimization of the algorithms behind an app - which is absolutely essential - but we put aside lots of data that are excluded because people did not use the app correctly. In our study conducted among 48 Swiss residents in free-living conditions, around 13% of the data that we gathered had to be discarded due to errors in the capturing procedure. For example, people did not include the whole plate in the picture, or even they have taken photos from the packaged food’s box. We concluded that we need to train people to correctly taking photos and following the instructions of each app. By detecting what is being done wrongly, we can create better instructions, we can optimize the apps to validate obscure input, we can guide people better by designing better studies, and thus, improve the quality of data!


What are some of the misunderstandings and confusions surrounding nutrition apps and their effectiveness that you have encountered during your research?

It is true that there is an abundance of nutrition apps in the app stores and the majority of them ist not validated and there is no scientific proof, that they work. As a result, the effectiveness of those apps has not been tested in detail. Apart from that, most of the nutrition apps include a feature to track food intake, are asking the user to manually insert portion size information of what they have consumed. However, we do know from previous research that people are not trained enough (or even not at all) in estimating portion sizes and therefore, can make mistakes. As a result, one mistake brings the other. That is where AI comes and can solve those problems by literally estimating the portion size of the food/drink that is about to be consumed. Machine learning is a way in which hundreds of thousands of photos are gathered with the goal that the system will start learning to recognize the foods and be able to estimate the different portion sizes.


"There is an abundance of nutrition apps in the app stores and the majority of them is not validated and there is no scientific proof, that they work." - Maria Vasiloglou

young woman working on her laptop, Maria Vasiloglou

Where do you see the applications of these technologies in the future? How can these help people make better food choices?

Smartphone apps can be integrated into our everyday life if large studies are conducted to first prove their effectiveness. Apart from this, I believe that such technologies would essentially help people who do not have easy access to healthcare facilities because they live in rural areas or even countries whose healthcare systems cannot support frequent visits to healthcare professionals. I deeply believe that the engagement of those apps can substantially improve the quality of people’s life. If those apps also include behavioral change techniques instead of only tracking food intake, then we could aim at long-term beneficial effects. Gamification techniques that include motivation elements could probably make users use the apps longer. Multidisciplinary teams are needed to design those apps taking into consideration user needs and clinical well-structured studies are needed to test them. Recent technology can definitely assist with this!

Personalized nutrition technologies are a new paradigm for dietetic practice and training in the digital transformation era. How does your research cater to this growing trend?

My research is mainly focused on dietary monitoring and assessment by the use of image-based apps, meaning apps that use a meal image/video as an input and via AI features can translate this picture/video into nutrients. Ideally, an image-based app can be individualized according to the user’s needs, conditions, preferences and also be connected with other apps aiming for example at tracking physical activity. We should first, though, ensure that validation and scientific evidence is behind those apps and that’s why I’m involved in such a research team.


What are the main barriers you’ve experienced when conducting research on nutrition apps, and where do you see gaps and opportunities in this sector? What are some learnings from them that you would like to communicate to all our readers?

I truly think that research on mHealth apps needs researchers from different disciplines to work together if we want to produce a complete system, a complete app. It should be a team of dietitians, doctors, behavioral scientists, psychologists, AI experts, computer scientists, and of course the end-users.


I believe that it is challenging to coordinate with people coming from different scientific backgrounds since time is needed to understand each other and speak a “common language”. At the same time that is one of the greatest parts of interdisciplinary research because it broadens your horizons and makes you think in a more complex way. Moreover, what is also challenging is the rapid advancement of mHealth and eHealth domains which require a constant need for being up to date with the related literature and the ongoing trials.


Well, I would like to mention that young researchers should take advantage of their fresh way of thinking and seek opportunities in interdisciplinary teams. To my knowledge, there are very few dietitians in the domain and we definitely need more in order to be able to communicate with each other and build strong collaborations.

What excites you about nutrition and what's your vision for our foods and nutrition system by 2030? What role will nutrition experts play?

I love the fact that nutrition is a rapidly evolving science and that different studies are taking place around the world. I can only speculate that the nutrition system by 2030 would be much more automatized, the connection from “farm to fork” would be easier and the availability of a variety of foods would be wider. Hopefully, we would be following more sustainable dietary patterns and we will learn to care more about the environment, too. Nutrition experts are the fundamental players of it who need to be always well-informed and trained in order to provide evidence-based advice. I believe that nutrition experts can substantially benefit from the integration of technology by making their life a bit easier in terms of time spending in re-training patients or automatizing procedures. For example, if there is an app that can automatically estimate the amount and type of food, replacing the conventional food diary, much time will be saved and can be dedicated to coping with the behavioral aspect of nutritional change or the training for a specific disease-related issue.

And last question: Who is your role model?

I get inspired by people, by the power of human beings. I was inspired by amazing children with diabetes type 1 that I met in the camp in Valencia some years ago where I volunteered. There, I understood that people not only need good clinical assistance but also a big hug. I have also been inspired by scientists that I met in conferences or meetings. It’s usually people who even though they have climbed the career ladder, still have these sparkling eyes of creation. Great people can inspire me with their way of thinking and their perspective on life. And somewhere there, I realize that this research world is totally worth it!

If you’d like to find out more about Maria or her research, you can connect and network with her via LinkedIn, Twitter, or Google Scholar (for publications).


This interview was carried out by Bhargavi Arvind. Bhargavi is a science editor at Nutrition Hub and is an Indian clinical and holistic nutrition expert.

Bhargavi Arvind, editor at Nutrition Hub