The choice stepping reaction task: how prefrontal network topology reconfigures in aging and dementia

The ability to initiate stepping promptly and accurately is critical for safe mobility in later life. Choice stepping reaction time (CSRT) is an independent predictor of recurrent falls in older adults and is often prolonged in dementia. Although the prefrontal cortex (PFC) is crucial for rapid response execution, existing studies primarily focus on regional activation, leaving the topological reorganization of prefrontal functional networks poorly understood. In this study, twenty-six healthy older adults (HOAs) and 25 older adults with mild-to-moderate dementia (MMD) completed four blocks each of a simple stepping task and a choice stepping task in randomized order. Prefrontal hemodynamic responses were recorded using a 48-channel functional near-infrared spectroscopy (fNIRS) system. We quantified task-evoked regional activation and derived graph-theoretic network metrics within the PFC. Finally, we investigated their associations with CSRT. Compared with HOAs, the MMD group showed higher right DLPFC activation during simple stepping (p = 0.005), yet the activation decreased from simple to choice stepping (p = 0.032) in MMD. Graph analysis revealed a significantly greater prefrontal clustering coefficient (C_P) in MMD (p = 0.005) compared with HOAs, whereas other global network metrics were comparable between groups. Moreover, higher C_P during choice stepping was associated with shorter CSRT (r = -0.55, p = 0.021) in MMD, suggesting that greater prefrontal network segregation may correlate with faster stepping reaction time. Our findings suggest altered task-evoked prefrontal activation alongside localized topological reconfiguration in older adults with dementia, with greater prefrontal network segregation linked to faster volitional stepping execution.