Plate Nº 17 · recorded November 17, 2026

Health & Medicine ResearchReported finding

AI links movement in 3,328 adults to liver and sleep health

An AI analysis of movement in 3,328 adults linked whole-body motion to 1,887 of 3,210 health traits, including liver stiffness and sleep measures, while leaving diagnosis unproven.

By Nathan Brooks3 min read670 words

In brief

  1. Movement recordings from 3,328 adults yielded 1,887 statistically significant associations among 3,210 health characteristics.
  2. A depth camera tracked 26 joints in 3D during five walking, chair-rise and balance tasks.
  3. The model's statistical predictions of liver stiffness carried as much information as liver blood tests.
  4. Torso movement carried associations with sleep apnea, insomnia, exercise and previous-night sleep.
  5. Published in Nature Medicine in 2026, the cross-sectional study still requires independent clinical testing.
The Way You Walk Could Reveal Clues to Your Liver And Sleep Health, Study Finds
Plate Nº 17The Way You Walk Could Reveal Clues to Your Liver And Sleep Health, Study Finds — Netti_Nu_Nu / pixabay

By Nathan Brooks

An AI system found statistically significant links between whole-body movement and 1,887 of 3,210 health characteristics in 3,328 adults, including liver stiffness and measures of sleep. The count equals nearly six in 10 of the characteristics the team tested, but statistical significance does not mean the system can diagnose disease or identify which condition a person has.

The Weizmann Institute of Science team reported the results in Nature Medicine in 2026. Study author Adam Gabet, a computer scientist at the institute, said: "When we started this project, we believed that movement could tell a lot about your health."

What did the AI measure?

Researchers recorded how participants moved in a controlled setting. Doctors already watch changes in walking speed, balance and posture, and researchers have described walking speed as a "sixth vital sign."

The study included five tasks:

  • Walking on a treadmill at 3 kilometers per hour, about 1.9 miles per hour
  • Walking on a treadmill at each participant's preferred pace
  • Walking in place
  • Rising from a chair
  • Standing still to test balance

A depth-sensing camera recorded the positions of 26 joints in three dimensions. Researchers trained the model by hiding parts of the movement recordings and asking it to reconstruct what was missing, forcing the system to learn how different body parts move together.

After converting movement into numerical patterns, the researchers compared those patterns with laboratory results, body measurements, medical conditions and lifestyle characteristics. This gave the model a broader view than walking speed or stride length alone.

What did whole-body movement reveal?\n

The 1,887 statistically significant associations included age and body fat, alongside signals less obviously connected with walking. Movement patterns also provided information beyond familiar measurements such as walking speed and stride length.

  • Liver: The model predicted liver stiffness, which doctors use as an indicator of liver scarring. Within the study, those statistical predictions carried as much information as liver blood tests.
  • Sleep: Movement of the torso, the area between the shoulders and hips, carried information associated with sleep and exercise. The model predicted measures linked with sleep apnea, insomnia and how much sleep participants had received the previous night.
  • Broader predictions: Adding movement to a model that already accounted for age, body mass index, fat around internal organs and height improved predictions of diabetes, depression, insomnia and medication use.

Gabet said the effect of sleep quality on the body stood out to him. The findings suggest that sleep health can be reflected in movement, but they do not establish that the system can diagnose sleep disorders.

Conventional walking research has often concentrated on the legs, leaving upper-body movement relatively unexplored. The results suggest that examining the whole body may reveal information traditional walking measurements miss.

The researchers hope movement analysis could eventually complement routine health assessments. A brief test requiring no blood samples or invasive procedures might help doctors identify changes worth investigating further, rather than provide a diagnosis on its own.

What does the study not prove?

The analysis was cross-sectional, meaning it captured participants at a single point in time. It therefore cannot show whether sleep problems or diseases caused particular walking patterns, or whether changes in movement came before disease signs.

Participants were generally healthy and mostly Ashkenazi Jewish adults living in Israel. That limits how confidently the findings may apply to other populations.

The system also relied on a particular depth camera and its software. Newton Tech, which makes the camera system used in the study, co-authored the paper.

"Clinical studies that show clear separation with good specificity would be needed before medical deployment," Gabet said. Specificity means the ability to correctly identify people who do not have the condition being tested.

Researchers in Gabet's laboratory have previously produced results that held up in other groups, but this model still needs independent testing. The team now wants to investigate whether movement changes can appear before signs of disease and potentially reveal health problems earlier.

via sciencealert.com (Source)

Filed under

  • artificial-intelligence
  • movement-analysis
  • liver-health
  • sleep-health
  • digital-health
Share this article:

More from Nathan Brooks

Nathan Brooks

Show full bio

Market editor covering companies, deals and strategy in general science news at SciBeat.

456 articles

Nearby plates

« Previous articleNext article »