Characterization and clinical implications of ankle impedance during walking in chronic stroke

Abstract

Individuals post-stroke experience persisting gait deficits due to altered joint mechanics, known clinically as spasticity, hypertonia, and paresis. In engineering, these concepts are described as stiffness and damping, or collectively as joint mechanical impedance, when considered with limb inertia. Typical clinical assessments of these properties are obtained while the patient is at rest using qualitative measures, and the link between the assessments and functional outcomes and mobility is unclear. In this study we quantify ankle mechanical impedance dynamically during walking in individuals post-stroke and in age-speed matched control subjects, and examine the relationships between mechanical impedance and clinical measures of mobility and impairment. Perturbations were applied to the ankle joint during the stance phase of walking, and least-squares system identification techniques were used to estimate mechanical impedance. Stiffness of the paretic ankle was decreased during mid-stance when compared to the non-paretic side; a change independent of muscle activity. Inter-limb differences in ankle joint damping, but not joint stiffness or passive clinical assessments, strongly predicted walking speed and distance. This work provides the first insights into how stroke alters joint mechanical impedance during walking, as well as how these changes relate to existing outcome measures. Our results inform clinical care, suggesting a focus on correcting stance phase mechanics could potentially improve mobility of chronic stroke survivors.

Conflict of interest statement

The authors declare no competing interests.

© 2021. The Author(s).

Figures

Figure 1

Average inter-subject stiffness (A) and damping (B) as a function of stance phase. Ankle impedance estimates during walking of individuals with chronic stroke are indicated in dark green (paretic limb) and light green (non-paretic limb). Dark grey traces indicates impedance estimates of three gait-speed matched older adults without stroke, within a similar age range to participants with chronic stroke. Light grey traces present impedance as a function of stance phase for young healthy adults walking at a faster speed from previous literature. Stiffness for stroke participants was constant across the stance phase of walking and did not demonstrate the stereotypical increase in mid-stance that prepares for forward propulsion for either limb. Stiffness of the non-paretic limb was significantly larger than the paretic limb, and both were increased compared to age and gait-speed matched controls. Older adults walking at a slower pace exhibited a similar pattern of stiffness variation to young healthy adults with a lower peak stiffness in mid-stance. Damping did not vary significantly across stance phase for either limb of stroke participants or age and gait-speed matched controls.

Figure 2

Average normalized EMG of the tibialus anterior (A) and medial gastrocnemius (B) across the stance phase of walking. Trials were normalized to the average peak EMG activity of a muscle throughout stance for each participant.

Figure 3

Stiffness (A) and damping (B) regressed across co-contraction index. Stiffness and CCI were significantly correlated for the paretic limb of individuals with chronic stroke and the young healthy adult; however the displayed opposite correlations. During walking, for the young healthy adult increased co-contraction was associated with lower stiffness, while for the paretic limb increased co-contraction was associated with higher stiffness. The non-paretic ankle stiffness of stroke participants and gait speed matched older adults did not correlate with co-contraction index. Ankle damping was not correlated with CCI.

Figure 4

Stiffness (A–E) and damping (F–J) asymmetry linearly regressed across four clinical measures. Six Minute Walk Test distance was significantly correlated with the difference in damping between the paretic and non-paretic limbs (F), but did not relate to stiffness asymmetry (A). Ten Meter Walk Test speed was significantly correlated with damping asymmetry at the self-selected speed (G) but not the fast speed (H). Stiffness asymetry did not correlate with either 10MWT (B,C). Lower extremity Fugl–Meyer motor score did not significantly correlated with either ankle stiffness asymmetry (D) or damping asymmetry (I). Similarly, modified Ashworth score did not significantly correlate with either damping (J) or stiffness (E) asymmetry.