Using AI to Uncover What Drives How Much We Eat — a clinical trial (ClinicalTrials.gov)
An observational study that measures healthy adults extensively over two years and applies AI to understand why people differ in their tendency to gain weight.
Trial overview (primary data)
- StatusNot yet recruiting
- ConditionsHealthy Volunteer
- SponsorNational Institute of Diabetes and Digestive and Kidney Diseases (NIDDK)
- Target enrollment800 participants
- Period2026-09-23 〜 2037-07-01
Key points
- An observational study in healthy adults exploring factors linked to food intake and future weight gain.
- Repeatedly collects multifaceted data over two years: samples, body composition, glucose, metabolism, cognition, and eating behavior.
- Plans to apply AI to the complex, high-dimensional data to surface interactions between factors.
- It is not a trial testing treatment effects, but research aimed at understanding individual differences in weight gain.
- Blood, hair, urine, stool, body fat, activity, continuous glucose, gastric emptying, cognition and feeding tests over two years from one participant.
1Why the same diet produces different weight
Overweight and obesity are common health challenges in the United States, yet people who eat similarly can differ greatly in how much weight they gain. Untangling why is hard, because appetite, metabolism, the gut, brain function, sleep, and activity all interact in complicated ways.
AI (machine learning) is well suited to handling many different measurements at once and surfacing patterns and combinations that are difficult for people to spot. This study aims to apply that strength to the search for the determinants of how much we eat and how weight changes over time.
2Layered data from each participant
What stands out is the breadth of data collected repeatedly from each participant over two years: blood, hair, urine, and stool samples; DXA scans of body fat; a wrist activity monitor; continuous glucose monitoring; mixed-meal and gastric-emptying tests; resting metabolic rate; cognitive tasks measuring attention and memory; breakfast and lunch eating tests; and questionnaires.
By integrating this complex, high-dimensional data with AI, the study seeks to reveal interactions between factors that simple, piecemeal measures would miss. It is an observational study and does not test the effect of any specific treatment or intervention.
3AI that helps understanding before treatment
Seen more broadly, this study is an example of how medical AI can add value first at the stage of understanding, before treatment. If the reasons for weight gain differ from person to person, mapping them could lay the groundwork for prevention and dietary guidance tailored to individuals.
The approach of reading multifaceted measurement data with AI is one many expect to extend beyond obesity to other lifestyle-related health questions.
4How wide the data from one participant runs
What distinguishes the study is the breadth of data collected repeatedly from one participant over two years, aimed at surfacing interactions that no sum of fragmentary measures would show.
| What is collected | Content |
|---|---|
| Fluids and specimens | Blood, hair, urine, stool |
| Body composition and activity | DXA body fat measurement, wrist-worn activity monitor |
| Metabolism and digestion | Continuous glucose monitoring, mixed-meal and gastric emptying tests, resting metabolic rate |
| Cognition and behaviour | Attention and memory tasks, breakfast and lunch feeding tests, questionnaires |
People eating similarly gain weight very differently, and the difference involves appetite, metabolism, the gut, brain function, sleep and activity in combination, which conventional methods struggle to separate. This is an observational study and does not test the effect of any treatment or intervention.
Why it matters
The design of integrating multifaceted health data with AI offers signals for individualized prevention and nutrition guidance, and for building data foundations in obesity-related research.
FAQ
What is this study trying to learn?
What do participants do?
Sources (primary)
Source: ClinicalTrials.gov (U.S. NIH/NLM, public domain). This site does not provide medical advice. Verify the latest and exact details with the official source. This site is not endorsed or certified by the NIH/NLM.
- ClinicalTrials.gov (study record, original)
- NCT ID: NCT07637656