
How you walk could reveal clues about your liver and sleep health, study finds: ScienceAlert
Before you’ve said a word about why you came, your doctor may have already noticed something about your health. The way you walk across the room may offer a first clue.
Scientists have now developed an AI system that looks at how the whole body moves, and have discovered links to health that go far beyond the legs. Some concern sleep, diabetes and even the liver.
In a new study published in Natural medicineResearchers at the Weizmann Institute of Science in Israel analyzed the movement recordings of 3,328 adults to see what those movements might reveal.
Doctors have long monitored changes in walking speed, balance and posture. Researchers have even described walking speed as a “sixth vital sign.”
“When we started this project, we thought movement could tell a lot about your health,” study author Adam Gabet, a computer scientist at the Weizmann Institute of Science, told ScienceAlert.
The team wanted to know if a computer could recognize some of the subtle signs doctors noticed.
In a controlled environment, participants performed five movement tasks: walking on a treadmill at 3 kilometers (about 1.9 miles) per hour and at their preferred pace, walking in place, getting up from a chair, and standing still to test their balance.
A depth-sensing camera recorded the position of their joints in three dimensions. The researchers then trained an AI model by masking parts of the movement recordings and asking it to reconstruct what was missing.
To fill in the gaps, the system had to learn how different parts of the body move together.

It converted these movements into digital models, which researchers compared with detailed health information, including laboratory results, body measurements, medical conditions and lifestyle characteristics.
The model identified statistically significant associations with 1,887 of the 3,210 health characteristics examined. Movement patterns also provided information beyond familiar measures such as walking speed and stride length.
Some links concerned age and body fat. Others pointed out health characteristics that might seem unrelated to walking.
One involved the liver. Doctors can assess liver scarring by measuring the stiffness of the organ. The researchers found that movement patterns could predict liver stiffness and reported that the predictions were as informative as liver blood tests.
Sleep was another surprise: Movements in the torso, the area between the shoulders and hips, conveyed information associated with sleep and exercise.
“One finding that struck me is the extent to which sleep quality affects our bodies,” Gabet said.
The model could predict metrics related to sleep apnea, a condition in which breathing repeatedly stops or becomes restricted during sleep. The study also found associations with insomnia and the amount of sleep participants had the previous night.

These results suggest that sleep health is reflected in movement patterns, although the study does not establish that the system can diagnose sleep disorders.
Gabet said conventional research on walking has often focused on the legs, leaving upper body movements relatively unexplored.
The team’s results suggest that examining the whole body could reveal information that traditional gait measurements miss.
Adding movement patterns to a model that already accounted for age, body mass index, fat around internal organs and waistline improved predictions of diseases such as diabetes, depression and insomnia, as well as medication use.
Researchers hope that movement analysis could eventually complement routine health assessments. A brief test, requiring no blood draws or invasive procedures, could help doctors identify changes that merit further study.
However, the results do not establish that sleep problems or illnesses cause particular walking habits. The analysis was cross-sectional, meaning it captured participants at one point in time, so researchers can’t tell which changes occurred first.
Participants were generally healthy Ashkenazi Jewish adults living in Israel, so the results may not apply equally to other populations. The system also relied on a particular depth camera and its software. Newton Tech, which makes the depth camera system used, is a co-author of the study.
Gabet noted that the lab’s related research has produced results that have held up to other groups, but that this model still requires independent testing.
“Clinical studies showing clear separation with good specificity would be necessary before medical deployment,” Gabet said.
Researchers now want to determine whether changes in movement can appear before signs of illness, potentially revealing health problems earlier.
The study was published in Natural medicine.
This article was fact-checked by Rebecca Dyer and edited by Rebecca Dyer. Although we are proud of our process, we are only human. If you spot an error, please let us know.
Gn Health