AI treadmill detects early Alzheimer’s and Parkinson’s signs

A team led by French neuroscientist Leslie Decker at University of Caen has developed a device capable of detecting early signs of cognitive impairment and neurodegenerative conditions such as Alzheimer’s disease and Parkinson’s disease.

Introduced in 2019 at the university’s virtual reality laboratory, the Présage project is an ambitious academic initiative that merges virtual reality, mathematics, and artificial intelligence.

Treadmill-Like System Installed in Virtual Reality Lab

The French research team designed a system that resembles a treadmill and installed it in a roughly 15-by-9-meter space at CIREVE, the university’s virtual reality lab.

According to Decker, the treadmill assesses the locomotor system and detects health biomarkers. It adapts to the patient’s walking speed and uses two force platforms to measure ground reaction force and dynamic balance data.

Dynamic Treadmill Integrates Virtual Environment for Cognitive Testing

He explains that the treadmill can tilt in the participant’s chosen direction and side-to-side, requiring greater cognitive effort to maintain balance. The team also fully integrates the device with a virtual environment.

During the assessment, the researchers expose the patient to cognitive stimuli while they walk—first at a steady speed, then at varying speeds for each leg.

The French research team then applies mathematical metrics to analyze and characterize the patient’s movements, assessing both cognitive and motor risk factors. When researchers identify abnormalities, the likelihood of developing severe neurocognitive disorders increases threefold.

As Leslie Decker explains, the team aims to determine whether they can recognize patients at risk of developing these conditions at a very early stage.

Around 100 patients between the ages of 55 and 87 have already used the tool, and 20 of them showed signs of motoric cognitive risk syndrome (MCR), a condition marked by slowed walking speed and self-reported cognitive concerns.

To define a locomotor profile for the syndrome, the team used AI models and compared results with healthy participants to establish benchmarks, according to Baptiste Perthuy.

Gait Analysis Reveals Early Neurodegenerative Risk

Perthuy explains that this approach helps identify individuals who may be at risk of developing neurodegenerative diseases. He notes that gait forms a locomotor profile reflecting a person’s physical condition, offering insights into pathology, overall health, and even emotions.

Another team member, Julien Rossato, says that when movement and cognition are affected, the test can measure cognitive reserve—the brain’s ability to adapt to aging. This reserve may decline with disease onset or naturally diminish over time.

Rossato explains that one of the project’s main focuses is evaluating performance in the dual tasks of walking and responding to stimuli. To achieve this, sensor-like electrode markers are attached to participants, while surrounding cameras track their position in space.

Measuring Movement and Cognitive Response

This setup allows researchers to measure variables such as joint angles and the time required to lift a leg. At the same time, cognitive performance is assessed through voice recordings and reaction-time analysis.

The team then applies mathematical models and advanced algorithms to interpret the data and design personalized prevention strategies. According to Rossato, the project’s final phase is to connect these performance indicators with neurocognitive assessments and social questionnaires.

The team is adapting the device currently used in the university laboratory for use in medical clinics. The system, created by the startup a-gO, relies on three iPhone devices to record a patient’s movements as they walk on a treadmill for five minutes.

Using these recordings, AI creates a 3D model of gait to detect motoric cognitive risk syndrome, a condition linked to slow walking and cognitive issues that can precede neurodegenerative diseases, according to Alexandre Dalibot.

Early Risk Profiling for Personalized Treatment

The objective is to build profiles of patients who either have the syndrome or display warning signs that warrant closer monitoring, making it possible to introduce preventive actions or more tailored treatments.

Dalibot explains that a-gO aims to develop a tool that identifies people at risk of neurodegenerative diseases early, while neuronal function and cognitive reserve are still largely intact.

---

Read the original article on:g1.globo

Read more:Can hyper-realistic virtual worlds improve our well-being?