Using immersive virtual reality and nonlinear dynamics to reveal motor signatures and adaptive strategies from successful aging to early neurocognitive vulnerability

The ProfilAGE project (Identifying new cognitive-locomotor PROFILes of AGEing using a data-driven approach) builds on the rich multimodal dataset collected during the PRESAGE study (Multimodal PREdictive study of motoric cognitive risk Syndrome and underlying executive mechanisms in AGEing), which involved 100 participants aged 55 and over. These participants spanned a continuum from successful aging to motoric cognitive risk (MCR) syndrome.

ProfilAGE aims to deepen our understanding of aging trajectories by identifying distinct cognitive-locomotor profiles using an unsupervised, data-driven approach. The project leverages immersive virtual reality to recreate realistic walking environments, combined with progressively challenging cognitive tasks. This design enables researchers to explore how attention, motor control, and cognitive reserve interact across the aging spectrum.

By combining model-free and model-based analyses with advanced dimensionality reduction techniques, ProfilAGE seeks to detect early and subtle changes in performance that may signal emerging neurocognitive vulnerabilities. Ultimately, the project aims to improve early identification of at-risk profiles and contribute to the development of personalized prevention strategies that support autonomy and promote healthy aging.

The scientific rationale for this research is rooted in the rapid and ongoing demographic aging of the population. In France, for example, projections indicate that individuals over the age of 65 will represent one-third of the population by 2070—up to 32% in Normandy, compared to 21% in 2018. This demographic shift is accompanied by a higher prevalence of age-related conditions and an increased risk of dependency. Aging is a known risk factor for serious events such as falls, neurodegenerative diseases, and various geriatric syndromes. The depletion of functional reserves with age contributes to the transition from preserved function to frailty, which increases the risk of adverse health outcomes.

However, aging is a highly heterogeneous process. While some individuals experience significant functional decline, others maintain high levels of cognitive and physical functioning, as well as active social participation. This phenomenon is often referred to as “successful aging.” This variability underscores the necessity of supplementing existing diagnostic frameworks with more detailed, personalized characterizations of aging trajectories. In this context, the concept of frailty, understood as a transitional state between robust health and overt disease, is useful. Frailty delineates a critical window for early intervention, during which personalized prevention strategies can reduce the risk of neurocognitive disorders, limit the incidence of falls, and postpone the loss of autonomy.

It is essential to identify early markers of frailty. Slower gait speed and increased gait variability are well-established predictors that are widely used in clinical assessments. Walking is not just a motor activity; it also requires attention and cognitive resources. Evaluating gait under dual-task conditions, in which a cognitive task is performed simultaneously, increases the sensitivity of these indicators due to competition for cognitive resources. Difficulties in these contexts are often associated with cognitive reserve, which reflects an individual’s ability to adapt to aging or pathology.

ProfilAGE’s data-driven approach is specifically designed to exploit these observations. Its scientific objectives are structured in three steps: (1) identifying the strategies older adults use to manage increasingly complex dual-task situations, (2) clustering individuals into subgroups that share similar strategies, and (3) characterizing these profiles based on neuropsychological functioning and cognitive reserve. By moving beyond standard clinical classifications, this approach aims to better capture the diversity of aging processes and improve the early identification of vulnerable profiles.

To achieve this, ProfilAGE draws on data from the PRESAGE study, which generated a multimodal dataset of 100 participants aged 55 and older. This dataset spans a continuum from successful aging to motor-cognitive risk (MCR) syndrome. The PRESAGE protocol included two visits. During the first visit at the CIREVE facility, participants walked on an instrumented M-Gait treadmill while immersed in a virtual urban environment. During this task, participants performed an auditory Stroop-like cognitive paradigm with three levels of difficulty: low (fully congruent trials), medium (mixed congruent/incongruent), and high (predominantly incongruent). This paradigm was designed to progressively increase executive load. This setup allowed for an in-depth analysis of spatiotemporal and kinematic gait parameters under varying cognitive loads, capturing motor and cognitive adaptability in dual-task conditions.

During the second visit at the Cyceron imaging center, participants completed a comprehensive neuropsychological battery to assess key cognitive domains, including processing speed, episodic and working memory, sustained and divided attention, and executive functioning. Participants also completed validated questionnaires evaluating their sociodemographic and educational background (e.g., years of formal education and occupational complexity) and current and retrospective engagement in leisure activities across cultural (e.g., reading and attending museums), physical (e.g., regular exercise and walking), and social (e.g., club membership and volunteering) domains. Perceived social support was evaluated using indicators such as network size, frequency of social interactions, and subjective satisfaction with received support. These variables were then integrated into a multidimensional cognitive reserve index in line with theoretical frameworks that posit lifelong intellectual, social, and physical engagement buffers age-related neural decline and enhances resilience to pathology.

The data analysis pipeline in ProfilAGE will unfold in multiple stages. Dual-task performance will be quantified using model-free metrics, such as error rates, response times, and spatiotemporal and kinematic gait parameters, as well as model-based metrics derived from cognitive and locomotor performance models. These variables will undergo a principal component analysis (PCA) to identify components that reflect the various strategies that participants adopt to manage dual-task conditions, which may be effective or ineffective depending on task difficulty. A bootstrap analysis will then be used to determine if there are statistically significant differences in strategy across difficulty levels. Participants will then be clustered based on their expression of these components. This will allow us to identify subgroups that share similar dual-task adaptation strategies. Finally, we will characterize each subgroup by its neuropsychological performance and cognitive reserve index. Our underlying hypothesis is that individuals who express the least efficient strategies will also display lower cognitive reserve and/or reduced neuropsychological functioning. This will lead to a more precise classification of individuals susceptible to cognitive-motor decline.

ProfilAGE has both clinical and societal ambitions. The project aims to refine our understanding of aging trajectories and generate actionable insights for clinical decision-making by enabling the early and precise identification of individuals presenting specific cognitive-motor vulnerability patterns. These profiles will serve as the foundation for developing standardized, scalable tools that support the early detection of at-risk individuals across diverse older adult populations.