Reviewed by the Help Dementia Editorial Team — our editors review every article for accuracy against guidance from the National Institute on Aging, the Alzheimer’s Association, and peer-reviewed sources.
Yes. Wearable sensors can detect changes in how people walk—specifically, slower gait speed, shorter steps, longer step duration, and increased gait variability—that are associated with early dementia and mild cognitive impairment, often appearing before a person or their doctor notice cognitive symptoms. These changes in walking patterns are mediated by the same higher-order cognitive functions that decline in early dementia, making them sensitive digital biomarkers for disease detection.
For example, a person wearing an accelerometer-based wristband or waist patch could show measurable reductions in walking velocity and stride length months or even years before they fail a standard cognitive test or report memory problems to their physician. Research activity on wearable sensors and gait analysis for early dementia detection has remained consistently elevated from 2010 through 2025, with a sharp increase in scientific publications beginning in 2021. This growing body of evidence suggests that what once required expensive laboratory testing and trained neurologists to observe—a patient’s gait—can now be continuously monitored during everyday life using low-cost, wearable devices that measure inertial signals through accelerometers and gyroscopes. The technology has moved beyond proof-of-concept into clinical validation, with international consortiums like the Early Detection of Neurodegenerative Diseases Initiative (EDoN) now running trials to test whether these wearables are acceptable and effective for real-world early dementia screening.
Table of Contents
- How Do Wearable Sensors Measure Gait Changes That Indicate Early Dementia?
- What Types of Wearable Devices Are Used, and What Do They Actually Measure?
- What Research Evidence Shows These Sensors Can Detect Dementia Early?
- How Can People Actually Use Wearable Gait Sensors in Real Life?
- What Are the Key Limitations and Challenges of Wearable Gait Sensors for Dementia Detection?
- What Is the Current State of Clinical Translation and FDA Approval?
- What Does the Future Hold for Wearable Dementia Detection?
- Conclusion
How Do Wearable Sensors Measure Gait Changes That Indicate Early Dementia?
Wearable systems designed for dementia screening typically use tri-axial or three-dimensional accelerometers paired with gyroscopes, often attached to the lumbar vertebra (lower back) or worn as wristbands and waist patches. These sensors measure inertial signals during walking and extract more than ten distinct gait features, which researchers then cluster into four main categories: pace (walking speed), rhythm (step timing), time variability (consistency of timing across steps), and length variability (consistency of stride length). Machine learning models trained on these measurements can distinguish between healthy aging, mild cognitive impairment, and dementia with high accuracy by recognizing the subtle, systematic differences in how people move. In comparison, clinical observation of gait during a typical office visit is crude and subjective.
A physician might note that a patient “walks a bit slowly,” but they lack precise quantification. A wearable system, by contrast, can measure that a person’s gait velocity has dropped from 1.3 meters per second to 1.1 meters per second, their stride length has shortened by 8 centimeters, and the variability between consecutive steps has increased by 15%—all changes that correlate with cognitive decline. Patients with amnestic mild cognitive impairment (aMCI) consistently show reduced walking velocity, shorter stride length, and increased stride time variability compared to cognitively healthy controls of the same age. These gait changes appear even when people are unaware of them and before standard neuropsychological testing detects cognitive impairment, making wearable gait analysis a potential screening tool that could prompt earlier clinical evaluation and intervention.
What Types of Wearable Devices Are Used, and What Do They Actually Measure?
Current wearable systems for dementia research use low-power, multimodal sensors worn as wristbands or waist patches that simultaneously measure inertial signals, heart rate, electrodermal activity (skin conductance), pulse transit time, and even sound for analyzing speech patterns. The most common research devices include Shimmer wearables and Axivity accelerometers, which attach directly to the skin or clothing. These systems are designed to function continuously during everyday activities—not just during formal testing—allowing researchers and clinicians to collect real-world gait data as people move around their homes, neighborhoods, and daily environments over weeks at a time. A limitation of current wearable systems is that most remain in the research phase.
While accelerometer-based technology is proven and increasingly inexpensive, consumer wearables like fitness trackers typically do not extract the specific gait parameters needed for dementia detection. They count steps and estimate distance but do not measure stride time variability, gait speed with clinical precision, or the other quantitative metrics that correlate with cognitive decline. Additionally, interpretation of gait data requires machine learning models trained on dementia patient populations; the same raw accelerometer signal from a healthy person’s irregular walk cannot be directly compared to data from someone with early dementia without proper algorithmic processing. As of April 2026, no wearable devices have received FDA clearance specifically for early dementia detection based on gait analysis, though multiple technologies are currently in clinical validation phase. This means wearables are available for research but not yet recommended as clinical diagnostic tools.
What Research Evidence Shows These Sensors Can Detect Dementia Early?
Recent bibliometric mapping has tracked scientific output on wearable sensor technologies and gait analysis for early dementia diagnosis from 2010 through 2025, documenting consistent research growth and elevated publication activity through 2024 and into 2025. A 2024 landscape analysis revealed a substantial increase in publication volume on digital health technologies for Alzheimer’s disease and related dementias, with accelerated research beginning in 2021. This sustained and growing scientific attention reflects genuine progress: machine learning models trained on actigraphy and motor data from wearables have achieved high accuracy in distinguishing Alzheimer’s disease from other types of dementia and in detecting mild cognitive impairment at earlier stages than traditional cognitive testing. The most concrete example of this research translating into clinical action is the CODEC-II trial, registered on ClinicalTrials.gov in June 2025 (Trial ID: NCT07051408) as part of the Early Detection of Neurodegenerative Diseases initiative (EDoN)—an international consortium of research institutions developing wearable technology toolkits specifically for early dementia detection.
This trial is investigating the acceptability of wearable technology to patients and caregivers, a crucial step before these devices can be integrated into clinical practice. Unlike earlier studies conducted in controlled laboratory settings, CODEC-II aims to understand whether people are willing to wear these sensors during real life and whether the data collected in homes is as reliable as data from research centers. However, high accuracy in research studies does not automatically translate to clinical utility. Sensitivity and specificity vary depending on the population studied, the specific machine learning model used, and how gait data is collected and analyzed. A device that works well in a research cohort of 200 people aged 60–75 may perform differently in a diverse clinical population that includes younger individuals with early-onset dementia, people with mobility limitations from other conditions, or those taking medications that affect gait.
How Can People Actually Use Wearable Gait Sensors in Real Life?
Wearable sensors for dementia screening are designed for continuous, free-living monitoring during everyday activities, ideally collecting at least one hour of movement data per day for two weeks or more to establish baseline patterns and detect meaningful changes over time. Unlike laboratory gait analysis, which requires patients to walk specific distances under controlled conditions, wearable-based assessment happens naturally—during walks, household chores, shopping, and exercise—making the data more representative of how someone actually moves day-to-day. This real-world advantage comes from the sensors’ low power consumption and unobtrusive design. A practical comparison: Traditional gait assessment in a neurology clinic takes 10–20 minutes, involves walking a marked corridor while a clinician observes, and generates subjective impressions.
A wearable sensor assessment requires the patient to wear a device for two weeks during normal life, transmit data to a cloud-based or local analysis system, and receive a quantitative gait report without any additional effort. The tradeoff is that wearable data requires sophisticated software for interpretation and depends on the wearer remembering to keep the device on and secure. Some patients may forget to charge it, wear it inconsistently, or remove it during bathing and swimming, introducing gaps in the data. Additionally, wearables can be affected by other conditions that impair gait—arthritis, Parkinson’s disease, stroke, or even pain from an ingrown toenail—potentially creating false positive signals for cognitive decline. For individuals at higher risk of dementia (family history, subjective cognitive complaints, or age over 65), a periodic wearable screening—say, an annual two-week monitoring period—could provide an objective digital biomarker to flag whether gait changes warrant further cognitive evaluation.
What Are the Key Limitations and Challenges of Wearable Gait Sensors for Dementia Detection?
One significant limitation is distinguishing dementia-related gait changes from gait changes caused by other conditions. Parkinson’s disease, normal pressure hydrocephalus, multiple sclerosis, hip osteoarthritis, and even depression can all alter walking patterns in ways that superficially resemble early dementia. A wearable sensor alone cannot differentiate these causes; it can only flag that gait parameters have shifted. A person with a sprained ankle will show reduced gait speed and stride length, but this reflects acute injury, not cognitive decline. Additionally, medications—including some antidepressants, blood pressure drugs, and pain relievers—can affect balance and gait, potentially triggering false alarms. Another warning: current wearable systems are not standardized.
Different research groups use different accelerometer brands, different attachment locations (lumbar spine, wrist, hip), different data sampling rates, and different machine learning models. This lack of standardization means that a gait analysis from one research center may not be directly comparable to or compatible with another center’s system. Clinical adoption would require agreed-upon standards—similar to how ECG signals or blood pressure measurements are standardized—before wearables could be widely recommended and reimbursed by insurance. Cost-effectiveness remains theoretical rather than proven. While wearable sensors are inexpensive to manufacture, the full system cost includes the device, cloud infrastructure, machine learning analysis, and clinical interpretation. Current research suggests these systems could be cost-effective compared to neuropsychological testing, but specific pricing and reimbursement pathways have not been established as of April 2026. Without insurance coverage or regulatory approval, most people would have to pay out of pocket.
What Is the Current State of Clinical Translation and FDA Approval?
As of April 2026, wearable devices that measure gait for early dementia detection are in the clinical validation phase—meaning researchers are testing whether they work safely and reliably in real patient populations—but none have received FDA clearance for this specific use. The FDA does recognize the potential of sensor-based digital health technology and has issued guidance on developing medical devices that incorporate sensors, but the regulatory pathway for a wearable gait-analysis device marketed for dementia screening is not yet fully established. This represents a chicken-and-egg problem: companies are reluctant to invest in regulatory submissions without evidence that such devices will be approved and reimbursed, while the FDA waits for more evidence before establishing clear approval criteria.
The CODEC-II trial and similar international initiatives are meant to bridge this gap by generating the clinical evidence and real-world usability data that regulators and payers need to make decisions. If these trials show that wearable gait monitoring is acceptable, safe, and can detect dementia earlier than standard care, the regulatory environment could shift rapidly. Analogous to how continuous glucose monitors moved from research into clinical practice over a decade, wearable gait sensors could follow a similar trajectory—first for high-risk populations, eventually potentially for broader screening.
What Does the Future Hold for Wearable Dementia Detection?
The trajectory suggests that wearable sensors will eventually become part of dementia screening, but likely as one piece of a multimodal toolkit rather than as a standalone diagnostic test. Combining gait data with voice analysis, heart rate variability, sleep patterns, and cognitive games or app-based tasks could create a more comprehensive digital biomarker profile than gait alone. The EDoN initiative is already pursuing this multimodal approach, developing integrated wearable toolkits that measure multiple aspects of neurological and cognitive function simultaneously. As machine learning models improve and are trained on larger, more diverse populations, accuracy and generalizability will increase.
The path forward also depends on integration into healthcare systems, reimbursement decisions by insurance companies, and public acceptance. Regulatory approval alone is insufficient; clinicians must be trained to interpret wearable data, and patients must trust that continuous monitoring serves their interests rather than exposing them to privacy risks or unnecessary worry. The CODEC-II trial is explicitly investigating acceptability, recognizing that the best technology is only useful if people are willing to use it. Over the next three to five years, we should see clarity on whether wearable gait sensors become a routine part of dementia screening or remain a research tool.
Conclusion
Wearable sensors can detect gait changes—including slower walking speed, reduced stride length, and increased step-to-step variability—that correlate with early dementia and mild cognitive impairment, often before cognitive symptoms are noticeable. This capability represents a meaningful advance in early detection because gait changes reflect the underlying cognitive dysfunction of dementia and can be monitored continuously during real life without expensive laboratory visits. Current systems use accelerometer-based wearables worn as wristbands or waist patches and employ machine learning models to extract clinically meaningful patterns from raw motion data.
However, wearable gait sensors are not yet approved for clinical use, cannot distinguish dementia-related gait changes from gait changes caused by other conditions without additional clinical information, and currently lack standardization across research systems. If ongoing clinical trials like CODEC-II demonstrate safety, reliability, and patient acceptability, and if regulatory approval and insurance reimbursement pathways become clear, wearables could move into mainstream practice within the next five years. For now, anyone interested in these technologies should follow developments through clinical trial registries and their healthcare provider, recognizing that wearable gait monitoring may soon become a valuable tool for identifying cognitive decline early—when interventions are most likely to be effective.
You Might Also Like
- The Financial Literacy Program for People With Early Stage Dementia That Helps Them Maintain Independence
- The Emergency Medication Kit for Dementia Patients That Paramedics Can Use When Medical Records Are Unavailable
- Why Every Family Should Have a Conversation About Dementia Preferences Before Anyone Shows Symptoms
