Cognitive Predictors of Responsiveness to Reactive Step Training in People with Parkinson's Disease at Fall-Risk

Reactive stepping can be improved in people with Parkinson's Disease (PwPD). However, there is variability in the responsiveness to such training. This study examined if cognition could predict the responsiveness of PwPD to a two-week reactive step training intervention. 25 PwPD (70.52 years ± 7.15; Hoehn & Yahr range 1-3) at risk for falls completed a multiple baseline, open-label, uncontrolled pre-post intervention study. Reactive stepping was trained through a two-week (six-session) intervention with repeated support surface translations. Stepping performance was measured at two baseline assessments (B1 and B2), immediately after the intervention (P1), and two months after training (P2). Primary stepping outcomes were anterior-posterior margin of stability (MOS), step length, and step latency during backward steps. The primary aim assessed whether global cognition (Scales for Outcomes in Parkinson's Disease-Cognition - SCOPA-COG, & Montreal Cognitive Assessment - MoCA) was related to two-month retention of improvements in reactive stepping after practice. The secondary aim explored whether specific cognitive domains predicted retained stepping improvements, including attention/working memory, executive function, language, memory, and visuospatial function. Greater baseline global cognition was related to better two-month retention of step length improvements (SCOPA-COG: p = 0.002, f² = 0.31; MoCA: p = 0.002, f² = 0.38). However, only SCOPA-COG retained statistical significance after p-value adjustment for multiple comparisons (p = 0.04). Optimal cut-point analysis revealed that a SCOPA-COG threshold of 31 or higher was optimal for identifying individuals likely to retain improvement. Specific cognitive domains did not predict changes in reactive stepping outcomes. Participants with greater baseline global cognition, particularly as measured by SCOPA-COG, demonstrated greater retention of improvements in reactive stepping. In this cohort, a SCOPA-COG threshold of 31 could predict individuals likely to benefit from the intervention. These findings highlight the potential of cognitive screening to identify people more or less likely to benefit from reactive balance training.