• Peer-reviewed study of 3,328 adults links subtle movement patterns to 1,887 health measures and finds independent predictive information across nearly every major body system evaluated
  • Newton Tech is translating this science into scalable, camera-based movement biomarkers for clinical trials, capturing dimensions of treatment response, disease progression and compensation that conventional functional tests can miss
  • Markerless, on-site assessment is designed to bring advanced movement analysis out of the confines of limited and hard-to-access gait laboratories and into broad clinical and commercial use

BOSTON and TEL AVIV, Israel , Nov. 06, 2026 (GLOBE NEWSWIRE) — Newton Tech, a developer of AI-powered movement intelligence technology for clinical research, today announced the publication of a new study it co-authored in Nature Medicine showing that artificial intelligence can reveal extensive health information encoded in subtle patterns of human movement. In 3,328 adults, the model identified significant associations with 1,887 of 3,210 health measures spanning 18 body systems, from blood tests and liver imaging to sleep, mental health and clinical diagnoses.

The publication, “A Gait Foundation Model Predicts Multi-System Health Phenotypes from 3D Skeletal Motion,” will be published in the Nov. 12 issue of Nature Medicine, and the online version [https://www.nature.com/articles/s41591-026-04724-9] was published today.

For pharmaceutical and biotechnology companies, the findings have immediate relevance to clinical development. Traditional functional assessments often reduce movement to a single number, such as the time required to walk a set distance. Newton Tech is applying AI-powered movement analysis in clinical trials to measure dozens of dimensions of how a patient moves, from joint-level coordination and balance strategy to mechanical effort and compensation, giving sponsors a more sensitive view of treatment response and disease progression than conventional assessments alone may provide.

Using Newton Tech’s camera-based Movement Lab technology, researchers collected three-dimensional movement data during five standardized motor assessments. The research team analyzed the data using an AI foundation model capable of identifying complex patterns in movement that conventional gait measurements, such as walking speed and stride length, often miss. The study, whose lead author is Adam Gabet, was conducted through the Human Phenotype Project, a large-scale research cohort.

The study identified associations with a wide range of health measures, including cardiovascular and metabolic health, bone density, liver function, sleep, mental health and physical frailty. Importantly, movement data provided additional predictive information beyond what age, body mass index, visceral fat and height already explain.

“Movement contains far more biological information than the handful of numbers conventional tests record,” said Yuval Brodsky, CEO of Newton Tech and co-author of the study. “This paper shows the scale of that signal. Newton Tech has spent years building the datasets, domain-specific models and clinical infrastructure required to translate movement into practical, objective measures for drug development. Our goal is to give biopharma companies a scalable way to see treatment effects, disease progression and patient-to-patient differences that blunt endpoints can miss.”

The findings suggest that human movement could play a much broader role in understanding health and disease than previously recognized.

“This study is an important step toward establishing movement as a multi-system biomarker,” said Paolo Bonato, PhD, Director of the Motion Analysis Laboratory at Spaulding Rehabilitation Hospital and Associate Professor of Physical Medicine and Rehabilitation at Harvard Medical School. “Clinical gait analysis has long focused on a handful of measures tied to specific conditions. These findings show that the way a person moves carries information about health across nearly every body system, and that this information can be captured at scale, as this study did across more than 3,000 people.”

Revealing a New Dimension of Human Health

Newton Tech’s Movement Lab captured and processed the movement data, and Newton Tech team members are co-authors of the study.

Researchers analyzed 17,589 movement sequences from study participants. The models identified statistically significant associations with 1,887 of the 3,210 health measures evaluated. After accounting for age, body mass index, visceral fat and height, the models improved predictions across all 18 body systems evaluated in men and 17 of 18 in women.

Among the most striking findings, researchers discovered that movement analysis could reveal information about aspects of health not traditionally associated with walking or conventional gait analysis, including:

  • Liver health: Movement patterns predicted characteristics of liver function, including liver elasticity, the primary noninvasive measure of liver fibrosis. Adding movement data improved prediction of liver elasticity beyond age and body mass index by as much as, or more than, adding the full panel of standard liver blood tests, in both men and women. Liver health leaves a detectable imprint on how people move.
  • Mental health: Movement data revealed associations with depression and other mental health measures, improving predictions of reported depression beyond established demographic and body composition factors. In a screening analysis, adding movement data increased the number of depression cases identified by 81% in men. The models also predicted antidepressant use, a drug class known to affect movement.
  • Metabolic health: Movement data predicted visceral adipose tissue, the fat surrounding internal organs that is associated with metabolic risk, with a correlation of 0.83, substantially outperforming conventional gait measurements. In the same screening analysis, adding movement data increased the number of diabetes cases identified by 83% in men.

The models are also explainable, showing which parts of the body carry which health information. Lower-body movement was particularly informative for metabolic health and frailty, while torso movement contributed to predictions related to sleep and lifestyle. For drug developers, that explainability helps connect a change in how a patient moves to the physiological systems involved.

Beyond drug development, these insights can help clinicians better understand patients’ functional limitations, inform treatment decisions and track changes in physical function over time. Spaulding Rehabilitation Hospital in Boston, part of Mass General Brigham, is bringing Newton Tech’s Movement Lab into its Motion Analysis Laboratory under a multi-year agreement.

Spaulding’s Motion Analysis Laboratory is the primary clinical gait laboratory for Boston Children’s Hospital’s cerebral palsy team, where instrumented gait analysis guides surgical planning. Newton Tech’s system has already been installed and operated there, producing reports in Spaulding’s clinical format. The program spans pediatric cerebral palsy, stroke, traumatic brain injury and other conditions.

“Bringing this depth of analysis into routine clinical care requires systems that are fast and practical for busy clinics and trial sites,” Spaulding’s Bonato said. “Newton Tech’s markerless approach captures detailed whole-body movement in minutes, without markers or a complex laboratory setup, which is why we are bringing it into our laboratory. It takes clinical-grade movement analysis beyond a handful of specialist gait labs and into everyday clinical care and clinical trials.”

Bringing Advanced Movement Analysis into Clinical Trials Today

Clinical trials frequently rely on timed function tests, functional rating scales and other standardized assessments to evaluate whether patients are improving. These measures compress a rich, whole-body movement into a single number, typically a time or a distance, and discard most of the information it contains. They can fail to capture subtle changes in movement quality or distinguish true functional recovery from compensatory strategies.

For drug developers, more detailed movement analysis can help answer questions that conventional functional endpoints may not fully address: whether patients are experiencing meaningful functional recovery or compensating for impairments, how individual patients differ in their movement patterns, whether observed changes are consistent with a therapy’s intended mechanism of action, and whether a treatment causes movement-related side effects. In the Nature Medicine study, the models predicted which medications participants were taking, including antidepressants, proton pump inhibitors and antihistamines, beyond what age and body composition explain.

The need is especially acute in heterogeneous diseases and areas of high unmet need, including Duchenne and other muscular dystrophies, ALS, Parkinson’s and Alzheimer’s diseases, stroke and traumatic brain injury, where diagnosis, progression and treatment response can be difficult to characterize and established functional endpoints may be variable or insufficiently sensitive. Disease-specific validation will be required, but a richer, objective movement phenotype could help reveal drug effects and progression that are otherwise difficult to separate from natural variability or compensatory behavior.

Newton Tech’s MovementOS platform transforms video recordings of standard movement assessments into hundreds of objective biomechanical markers per patient, including gait symmetry, stability, coordination, joint movement patterns, variability, smoothness and mechanical effort. These insights support patient stratification, exploratory and secondary endpoints, responder identification and treatment-response analyses at the level of the individual patient.

By combining sophisticated AI analysis with accessible video-based technology, Newton Tech helps drug developers keep the standard functional assessments they have relied on for decades and extract far deeper insights from them, measuring hundreds of aspects of movement where these tests have traditionally recorded a single time.

Unlike traditional gait analysis, which often requires an expensive, specialized laboratory, reflective markers and trained technical staff, Newton Tech’s platform is designed for deployment at clinical trial sites using standardized, markerless video capture. It can be installed with minimal site training, and on-site processing converts recordings into non-visually-identifiable skeletal-motion data for Newton’s analysis. The platform can be used prospectively or applied retrospectively to existing clinical trial videos—including recordings made for rater quality control—helping sponsors extract substantially more information from functional assessments already embedded in their programs.

Newton Tech’s platform is already used in pharmaceutical clinical trials. In May 2026, NervGen Pharma reported an exploratory, post-hoc gait analysis from its Phase 1b/2a CONNECT SCI study of NVG-291 in chronic spinal cord injury, built on Newton Tech’s blinded markerless video analysis of the study’s walking-test recordings. In that analysis of 20 participants, a global statistical test combining measures of coordination, mechanical effort and postural stability favored NVG-291 over placebo (p=0.0197). Ten of 10 NVG-291 participants were classified as responders, compared with 1 of 10 on placebo (p=0.0001). The analysis identified treatment-associated movement patterns that were not captured by walking speed alone. This work was conducted separately from the Nature Medicine research and illustrates how detailed movement analysis can provide important additional information in clinical development.

Opening New Possibilities for Disease Screening and Diagnostics

Beyond clinical trials, the ability to extract information related to multiple physiological systems from a simple, noninvasive movement assessment that can be performed in almost any setting raises the possibility of using gait analysis to complement traditional diagnostic tools, help identify individuals who may benefit from further evaluation and monitor changes in health over time. In the study, adding movement data substantially increased the number of cases identified for conditions including depression, diabetes and peptic ulcer disease.

Unlike blood draws, imaging, wearable sensors or other specialized diagnostic procedures, markerless movement assessment can be completed quickly with affordable equipment and minimal site training. On-site processing converts video into non-visually-identifiable skeletal-motion data before analysis, supporting repeatable deployment across trial sites, clinics and community settings.

Screening and diagnostic applications remain investigational. Additional research, including prospective clinical studies and validation in more diverse populations, will be necessary to determine how effectively movement-derived biomarkers can support disease detection, diagnosis and monitoring.

The same multi-system sensitivity may also be relevant to longevity research, where interventions are intended to affect several physiological systems at once and no single measure can establish that aging has been slowed. Longitudinal movement biomarkers could become one objective component of a broader evidence package for evaluating whether a drug, lifestyle program or other intervention is producing meaningful change across systems; that application also remains investigational and will require prospective validation.

About the Study

The study, “A Gait Foundation Model Predicts Multi-System Health Phenotypes from 3D Skeletal Motion,” analyzed movement recordings from 3,328 participants in the Human Phenotype Project, applying deep learning to three-dimensional skeletal motion and its associations with thousands of physiological, clinical and behavioral characteristics.

About Newton Tech

Newton Tech develops AI-powered movement intelligence technology to advance clinical research and drug development. Its MovementOS platform transforms video recordings of standard movement assessments into detailed, objective biomechanical data, helping pharmaceutical and biotechnology companies better understand functional changes, evaluate treatment responses and gain deeper insights from clinical trial results. Newton Tech’s technology supports retrospective and prospective clinical studies across neurological, neuromuscular, musculoskeletal, cardiometabolic, rare and age-related diseases, and other conditions in which movement and physical function are important indicators of health.

For more information, visit https://newton-tech.net/.

Media Contact

Kira Gordon
BrightPoint Communications, LLC
kira@brightpointny.com


Primary Logo

Media gallery

About The Author
Newton Tech