New gold standard in diet tracking: How AI, wearable cameras, and biomarkers transform nutrition research

In a new analysis, scientists suggest combining different diet-tracking methods — from wearable cameras to dietary biomarkers — to paint a more reliable picture of what people eat and drink, rather than focusing on a single tool. Accurately capturing these trends has been one of the biggest challenges in nutrition research. 

The researchers stress that traditional self‑reported food diaries have often relied on memory and guesswork, placing a heavy time burden on participants. These limitations have made it difficult for researchers and policymakers to definitively link certain diets to health outcomes.

On a broader scale, these issues have kept scientists from learning how people’s food choices are evolving in response to global food system changes.

Speaking to Nutrition Insight, Dr. Thomas Wilson, a co‑author of the paper from the UK Aberystwyth University’s Department of Life Sciences, tells us more about why accurate dietary measurements remain such a persistent challenge, despite decades of methodological advances.

“There have been numerous advancements in the development of modern tools such as wearable cameras, sensors, and biomarkers of food intake. However, when it comes to its widespread and consistent adoption, these tools are still in their infancy.”

“A key challenge is that despite methodological advances, they have mostly centered around self-reported tools,” he continues. “Even with years of advances in improving the accessibility and usability of these tools, the fundamental issue remains — people find it difficult to accurately remember every specific detail of what they eat, and how much.”

Wilson adds that no single tool is perfect. “Self-reported tools such as Food Frequency Questionnaires and 24-hour dietary recalls suffer from misreporting errors and recall bias.”

“By integrating modern tools — such as biological biomarkers and digitally assisted reporting — we can dramatically improve accuracy while reducing the burden on participants. This opens the door to much more reliable research and helps us better understand the role of diet in long‑term health.”

Tech advances trim errors in diet tracking

In their review published in Nature Food, the team combined emerging evidence from nutrition science, metabolomics, microbiome research, computer vision, and sensor technologies.

New technological advances are helping participants in diet studies reduce memory‑related errors in food logging.The paper highlights how these new technological advances are helping users reduce memory‑related errors. Examples include smartphone apps and wearable cameras that capture meals in real time, identifying foods, and estimating portion sizes with the help of AI.

The authors also highlight biomarkers of food intake as a promising advancement in dietary assessment. These biomarkers detect chemicals in urine, blood, or stool that correspond to specific foods or dietary patterns, offering objective insights into what people have eaten.

Biomarkers are already being used in large-scale studies and are becoming more important as the impact of inaccurate dietary intake data on population health studies is more widely recognized.

“In the context of personalized nutrition, modern methods such as food intake biomarkers are essential,” says Winston. “These methodologies allow us to measure diet on a population scale, but also examine the effect of dietary intake on the individual level.”

Limitations of biomarker tracking

Wilson points out that implementing biomarker technology at population scale is not without challenges. “Biomarkers are objective measures (as opposed to subjective) but not every food has a validated biomarker available, and some bio-fluids are better suited to different food groups.”

“For example, urine is great for detecting the intake of fruits and vegetables, but it can’t tell us anything about the intake of food groups like dairy and polyunsaturated fats. Using blood-based biomarkers helps us detect food groups not detectable in urine.”

He adds that not every biomarker has the sufficient validation data, such as dose-response and long-term reproducibility, to demonstrate it is suitable in free living scenarios.

“Also, the analytical methods used are costly and complex, which can sometimes be a limiting factor in accessing dietary biomarker analysis for population studies.”

Evolving research design

The authors argue that emerging dietary assessment tools will be essential for advancing precision nutrition, improving dietary recommendations, and supporting evidence‑based policies for human and planetary health.

Winston goes into further detail about how an integrated dietary assessment framework could improve how scientists pinpoint diet-related health outcomes.

“It will mean that we can accurately link health outcomes with dietary intake,” he notes. “Now, we know that self-reported questionnaires are imprecise and this will ultimately impact diet and health associations when it comes to analyzing data from these studies.”

“Using objective methodologies in an integrated way will ensure that scientists have the most accurate representation for dietary intake and can then have a greater level of confidence in the subsequent health associations made.”

He adds that the implications of the findings for evidence-based food policy are huge, especially given how rapidly the food system is evolving. Popularized plant-based food and beverages, for instance, have been the subject of research into potential nutrient gaps.

“Dietary assessment methodologies must evolve at the same rate, and modern technologies such as biomarkers of food intake and data from wearable cameras will allow policy makers to fully understand the effect of food composition on long term health,” he says.

“Utilizing methods that have more flexibility will make it easier to scale across diverse populations but also tailor to specific research questions.”