Using wearable sensors and AI to spot sitting and understand sedentary habits
Leveraging deep learning to classify sitting posture and measure sedentary patterns from accelerometer data in diverse cohorts
This project uses data from wrist and hip activity trackers plus artificial intelligence to detect when people are sitting and describe how sedentary time is accumulated, especially for adults with or at risk for type 2 diabetes.
Quick facts
| Grant type | R01 grant |
|---|---|
| Study type | NIH-funded research |
| Funding institution | University of California, San Diego NIH-funded |
| Lab location | 1 site (La Jolla, United States) |
| Project ID | NIH-11229614 on NIH RePORTER |
What this research studies
I will use raw accelerometer data from wrist and hip trackers collected in several large studies to teach an AI model how to tell sitting from other activities. The team will build a deep-learning model (CNN combined with Bi-LSTM) and apply transfer learning so the model works across diverse age groups, body sizes, and device types. They will train and test the models on data from about 6,390 people and more than 200,000 hours of device wear with concurrent posture labels to improve accuracy and reproducibility. The outputs will characterize sedentary patterns like long uninterrupted sitting bouts versus fragmented sitting with many short breaks.
Who could benefit from this research
Good fit: Ideal candidates are adults with or at risk for type 2 diabetes who wear or can wear a compatible wrist or hip activity tracker, or who have previously contributed wearable data to research.
Not a fit: People who do not use wearable activity monitors, cannot wear a device, or whose health concerns are unrelated to sedentary behavior may not directly benefit from this project.
Why it matters
Potential benefit: If successful, this work could make activity trackers much better at measuring sitting patterns so doctors and patients get clearer information to guide steps that reduce diabetes and cardiovascular risk.
How similar studies have performed: Previous research shows wearables can estimate activity levels and the investigators have related prior work, but using deep transfer learning to improve posture-based sedentary measurement is a newer and less-tested approach.
Where this research is happening
La Jolla, United States
- University of California, San Diego — La Jolla, United States (Active)
Researchers
- Principal investigator: Natarajan, Loki — University of California, San Diego
- Study coordinator: Natarajan, Loki
About this research
- This is an active NIH-funded research project — typically early-stage science, not a clinical trial accepting patient enrollment.
- Some NIH-funded labs run parallel clinical studies or seek volunteers for related work. To check, contact the principal investigator or institution listed above.
- For full project details, budget, and progress reports, visit the official NIH RePORTER page below.