BE RU EN

Your Gait Can Reveal More About Your Health Than Medical Tests

  • 19.11.2026, 9:46

Scientists have discovered some unexpected patterns.

Sometimes, it takes a doctor just a few seconds—while the patient is walking across the exam room toward the chair—to suspect certain health issues. The way a person walks is often the very first sign that indicates hidden changes in the body, even before the conversation with the specialist begins. How can a person’s gait indicate a problem?

Science Alert reports on this.

Walking style may indicate hidden liver and sleep problems—a new study<

Scientists recently developed an artificial intelligence system that analyzes the motor activity of the entire human body. The algorithm analyzes the collected data and identifies correlations that are not directly related to muscles or joints. Specifically, it can detect sleep disorders, the onset of diabetes, and even liver health based on a person’s gait.

As part of a recent study, the results of which were published in the journal *Nature Medicine*, researchers at Israel’s Weizmann Institute analyzed movement recordings from 3,328 adult volunteers.

In clinical practice, observing balance, posture, and walking speed is nothing new—specialists have long used these markers, and walking speed has even been informally dubbed the “sixth vital sign” in the scientific community.

“When we started this project, we believed that movement could reveal a lot about your health,” said Adam Gabet , a co-author of the study and a computer scientist at the Weizmann Institute .

The developers faced the challenge of determining whether computer vision could detect those subtle micro-signals that are usually noticed by the trained eye of a medical professional.

During laboratory testing, participants were asked to perform five physical activity tasks: walking on a treadmill at a slow pace of 3 kilometers per hour, walking at their usual pace, walking in place, standing up from a chair, and standing still to assess balance.

A special camera equipped with a depth sensor tracked the position of the joints in three-dimensional space. Subsequently, the engineers trained the neural network using the “fill-in-the-blanks” principle: they intentionally cut out certain segments from the video recordings, forcing the program to independently reconstruct the missing motion trajectory.

To reconstruct these fragments as accurately as possible, the algorithm had to learn to understand complex biomechanics—specifically, how different segments coordinate and work together.

The resulting biomechanical profiles were converted into a numerical data set and compared with the participants’ medical records, which included blood tests, body measurements, chronic conditions, and daily habits.

Among the 3,210 health parameters examined, the computer model found a statistically significant correlation with 1,887 indicators. Moreover, the analysis of whole-body motor function proved to be significantly more comprehensive than traditional measurements.

One of the most telling examples was the liver. In medical practice, the degree of liver fibrosis or scarring is assessed based on tissue stiffness. As it turned out, digital movement patterns allow this indicator to be predicted with an accuracy that rivals that of specialized biochemical blood tests.

The connection to sleep proved to be no less surprising: it turned out that the nature of torso movements—in the area between the shoulder girdle and the pelvis—reflects how actively a person exercises and how well they sleep.

“The discovery that particularly struck me was just how much sleep quality affects our bodies,” Gabet admitted.

Artificial intelligence successfully identified signs characteristic of obstructive sleep apnea. In addition, the system detected the effects of insomnia and even took into account how many hours a participant had slept the previous night. This proves that sleep disturbances directly affect the body’s biomechanics.

When indicators of movement patterns were added to the basic predictive models—which already took into account age, height, body mass index, and visceral fat volume—were expanded to include indicators of movement patterns, the accuracy of detecting diabetes, depressive disorders, insomnia, and even medication use increased significantly.

The project’s authors expect that in the future, this type of screening will become a routine tool in doctors’ offices. It is a rapid test that does not require blood tests or complex procedures.

There are caveats to this diagnostic method

At the same time, scientists caution against jumping to conclusions. The study was cross-sectional in nature, meaning that the volunteers’ condition was assessed at a single point in time rather than over time; therefore, it is impossible to say with certainty which comes first—the disease or the change in gait.

Another limitation was the homogeneity of the sample: most participants were generally healthy Israelis of Ashkenazi descent, so the results still need to be validated in other ethnic and age groups.

“Before implementation in medical practice, clinical studies will be needed to demonstrate a clear distinction between indicators with high specificity,” the researcher concluded.

Going forward, the team plans to determine whether a person’s gait begins to change even in the preclinical stage, before the first noticeable symptoms appear.

Latest news