Walking Under Cognitive Load: Graded Dual-task And Gait Performance In Parkinson's Disease Vs Controls

The dual-task (DT) paradigm reveals impairments in movement and executive functions in pathology and ageing. Individual variability in mental calculation skills and baseline cognition results in an identical mathematical subtraction task causing a different subjective cognitive load: overload may result in DT failure, insufficient load in unmeasurable effects. PURPOSE: Test the feasibility and validity of an innovative graded DT in PD and age-matched controls. METHODS: 37 individuals (17 PD, 69.5 ± 6.7 yrs, 1.8 ± 0.6 H&Y; 30 C, 69.6 ± 5.2 yrs) performed 10-m walking trials at self-selected speed in single task (ST) and back-counting dual task (DT) at 3 levels of difficulty (-1, -3, -7), in randomized order. Gait velocity (GV), step length (SL), and % of double support (%DS) were measured with a photoelectric sensor system and photocells. For DT trials we calculated: i) cost for all parameters = % difference from ST; ii) difficulty index (DI) = 1/[(correct subtractions - errors)/s-1]. Variables, costs and DI were compared across trials and populations by two-way RM-ANOVA. The individual relationship of gait variables with DI was tested with linear fitting: intercept and slope of PD and C were compared by unpaired t-test. RESULTS: GV and SL were lower and %DS was higher in PD vs C (main effect of group p < .001 for all). The cost of DTs increased with task difficulty (main effect of trial p < .001), with no group difference (p > .05) nor interaction (p > .05) for all variables. DI increased from DT-1 to DT-3 to DT-7 (main effect of trial p < .001) with no difference between groups (p > .05) nor interaction (p > .05). The intercept of GV and SL was higher, that of %DS was lower in C vs PD (<.001 for all variables), while the slope of the linear relationship of all variables with DI was not different between groups (>.05 for all variables). CONCLUSION: A linear, dose-response relationship was revealed between cognitive difficulty of DT and gait performance in PD and C. Early-stage PD individuals had a lower initial performance vs C; however, under increasing cognitive load, they showed a comparable, progressive, reduction in GV and SL and increase in %DS. Graded DT appears feasible and provides results that may minimise the effects of subjective cognitive load on gait parameters’ variability, facilitating cross-sectional and longitudinal comparisons. Provincia Autonoma di Bolzano-Ufficio Innovazione e Tecnologia