Ageing affects the neural mechanisms of reflexive step initiation

Introduction: Step initiation is usually preceded by mediolateral anticipatory postural adjustments (ML-APAs), with recruitment of the stepping-leg gluteus medius acting to direct the body’s center of mass (CoM) towards the stance side. Yet, under specific circumstances, even earlier activation (~ 100 ms) may occur in the stance-leg gluteus medius (GM) that drives the CoM directly towards the intended stepping location, even in the absence of an ML-APA (Billen et al., 2023). These so-called “express visuomotor responses” (EVRs) are presumably subcortically mediated (Corneil et al., 2004; Pruszynski et al., 2010) and facilitate step reaction times when speed is critical (Billen et al., 2023; Giesbers et al., 2026). Yet, EVRs are typically suppressed when the impending postural demands increase, for example by increasing task difficulty (i.e., stepping anterolaterally versus anteromedially) (Billen et al., 2023) or when balance capacity is reduced (Nonnekes et al., 2010). In such scenarios, occasional EVRs may potentially be detrimental for balance if not compensated by an adequate ML-APA (Giesbers et al., 2026). As ageing is known to compromise postural control and increase fall risk, older people may prioritize postural control at the expense of fast, reflexive step initiation. Here we tested the hypothesis that ageing affects rapid step initiation, by an overall reduced EVR- and increased ML-APA expression, with longer step reaction times for older compared to young participants.Methods: We used an emerging target paradigm where 18 healthy young adults (25.2 ± 2.5 years) and 19 older adults (67.9 ± 6.8 years) stepped to visual targets presented in front of either the left or right foot at varying step eccentricities. EMG from bilateral GM and vertical ground-reaction forces (vGRF) were recorded concurrently. For each trial, we determined whether an ML-APA or EVR was present, including their onset times and times-to-peak, as well as the step reaction time. Next, we averaged the time-series GM activity and vGRF on EVR+ and ML-APA+ trials, respectively. All outcome measures were analyzed for the effects of age group, step eccentricity, and their interaction.Results: Older adults showed a general delay in step reaction times, that was most pronounced during anterolateral steps. In anteromedial steps, both groups showed similarly frequent ML-APAs (90 %) and relatively infrequent EVRs (~30 %). Yet, between stepping conditions, the older group showed a smaller increment in EVR+ trials (group x step eccentricity: p = .029) and smaller decrements in ML-APA+ trials during anterolateral compared to anteromedial stepping. EVRs in the older group were initiated marginally later in the older compared to the young group (109 vs. 104 ms, group: p = .002), and peaked lower (group: p < .001) and later (137 vs. 125 ms, group: p < .001). ML-APA onset latencies were similar between groups (189 vs. 183 ms) but lasted longer in the older group (APA time-to-peak 283 vs. 304 ms: group: p = .025) with greater step-side vGRF compared to the young group, specifically during anterolateral stepping (group x step eccentricity: p = .002).Conclusions: In line with our hypothesis, the present findings point towards an age-related decline in rapid step initiation, as shown by fewer EVRs (with marginal latency differences) and more frequent (and prolonged ML-APAs) during anterolateral steps. While these observed age-related changes may reflect a more cautious step strategy, we also observed an altered shape of the EVR, as characterized by a later and lower EVR peak in the older compared to the young group. This result may reflect a structural decline of the subcortical EVR network itself, which could directly limit the potential facilitating effects on rapid step initiation. Both mechanisms, in turn, may contribute to the observed delay in step reaction times.