Unilaterally Applied Resistance to Swing Leg Shows a Different Adaptation Pattern Compared to Split-Belt Treadmill in Patients with Stroke

Persons with chronic stroke (PwCS) have a decreased ability to ambulate and walk independently. We aimed to investigate the differences between the motor adaptation process for two different perturbation methods: split-belt treadmill walking and unilaterally applied resistance to the swing leg during walking. Twenty-two PwCS undergo split-belt treadmill walking and unilaterally applied resistance to the swing leg during walking, each one week apart. The test included three phases: the baseline period, the early-adaptation period and the late-adaptation period, as well as the early-de-adaptation period and the late-de-adaptation period. The average step length, swing duration, double-limb support duration, and coefficient of variance (CV) of these parameters were measured. During the split-belt treadmill walking, PwCS showed an adaptation of double-limb support duration symmetry (p = 0.004), specifically a trend between baseline versus early-adaptation (p = 0.07) and an after-effect (late-adaptation compare to early-de-adaptation, p = 0.09). In unilaterally applied resistance to the swing leg during walking, PwCS showed lower swing phase duration CV, in the adaptation period (baseline compare to adaptation, p = 0.006), and a trend toward increased variability of gait in the de-adaptation period compare to the adaptation periods (p = 0.099). The rate of adaptation and de-adaptation were alike between the two perturbation methods. Our findings show that the learning process happening in the central nervous system of PwCS may be dependent on the nature of the perturbation (mechanical resistance vs. split-belt) and that PwCS are able to adapt to two types of errors.