Variability of Leg Kinematics during Overground Walking in Persons with Chronic Incomplete Spinal Cord Injury

Abstract

Incomplete spinal cord injury (iSCI) often leads to partial disruption of spinal pathways that are important for motor control of walking. Persons with iSCI present with deficits in walking ability in part because of inconsistent leg kinematics during stepping. Although kinematic variability is important for normal walking, growing evidence indicates that excessive variability may limit walking ability and increase reliance on assistive devices (AD) after iSCI. The purpose of this study was to assess the effects of iSCI-induced impairments on kinematic variability during overground walking. We hypothesized that iSCI results in greater variability of foot and joint displacement during overground walking compared with controls. We further hypothesized that variability is larger in persons with limited walking speed and greater reliance on ADs. To test these hypotheses, iSCI and control subjects walked overground. Kinematic variability was quantified as step-to-step foot placement variability (end-point), and variability in hip-knee, hip-ankle, and knee-ankle joint space (angular coefficient of correspondence [ACC]). We characterized sensitivity of kinematic variability to cadence, auditory cue, and AD. Supporting our hypothesis, persons with iSCI exhibited greater kinematic variability than controls, which scaled with deficits in overground walking speed (p < 0.01). Significant correlation between ACC and end-point variability, and with walking speed, indicates that both are markers of walking performance. Moreover, hip-knee and hip-ankle ACC discriminated AD use, indicating that ACC may capture AD-specific control strategies. We conclude that increased variability of foot and joint displacement are indicative of motor impairment severity and may serve as therapeutic targets to restore walking after iSCI.

Keywords: AD; SCI; joint coordination; kinematics; variability; walking.

Conflict of interest statement

No competing financial interests exist.

Figures

FIG. 1.

(A) Schematic representation of the end-point trajectory during the swing phase of overground walking. Hip, knee, and ankle joints are indicated with circles. End-point trajectory reflected the metatarsophalangeal joint (MTP) marker trajectory during the swing phase of gait. (B) Schematic depicting the computation of normalized tolerance area of MTP marker trajectory; 95% tolerance ellipses of accumulated trajectory points computed for each bin of the swing phase normalized to the mean foot trajectory length, MTP, fifth metatarsophalangeal joint; AB, able-bodied; SCI, spinal cord injury; EV, end-point variability.

FIG. 2.

(A) End-point foot trajectory during the swing phase in one representative control subject compared with one incomplete spinal cord injury (iSCI) subject. See Methods. X-axis: fore-aft position; zero indicates the reference point (posterior superior iliac spine [PSIS]) in the sagittal plane. Y-axis: vertical elevation in sagittal plane. (B) Comparison of end-point variability during the swing phase for able-bodied (AB) and iSCI groups. Each data point represents the mean end-point variability for each subject averaged across cadence conditions. Lower values for end-point variability indicate less variability. Mean and standard deviation of each condition are listed in Table 2 (column “All”). Bold dots indicate outliers. AB, able-bodied cohort (controls). AB subjects walked without an assistive device. SCI, subjects with spinal cord injury.

FIG. 3.

Left. Representative joint-space traces (cyclograms) from an able-bodied (AB) control and a person with incomplete spinal cord injury (iSCI). Angular coefficient of correspondence (ACC) values for each indicated in the lower right corner. Right. Interjoint coordination consistency (ACC) between hip-knee (HK), knee-ankle (KA), and hip-ankle (HA). An ACC value of 1 indicates perfect consistency of interjoint coordination for two joints across multiple gait cycles. In all three joint combinations, iSCI had significantly lower ACCs than AB. Bold dots indicate outliers.

FIG. 4.

End-point variability versus angular coefficient of correspondence (ACC) across three joint combinations: hip-knee (HK), knee-ankle (KA), and hip-ankle (HA). Line indicates linear regression and gray shaded region delineates 95% confidence interval. Each marker represents one subject, and subjects are shape-coded by type of assistive device. In all three joint combinations of ACC, the relationship between end-point variability and ACC was significant, as indicated by the p values.

FIG. 5.

Relationship between walking speed and variability. (A) End-point variability and angular coefficient of correspondence (ACC) versus walking speed from the 10 Meter Walk Test (10MWT). (B) End-point variability and ACC versus trial walking speed during overground walking (self-selected cadence). Only hip-knee ACC is shown. The relationship between end-point variability and walking speed as well as the relationship between ACC and walking speed were significant. Line indicates linear regression and gray shaded region delineates 95% confidence interval. Shaded ellipse indicates 95% confidence ellipse for able-bodied (AB) data. For the first column, trial walking speed was used for AB ellipses, because the 10MWT was not performed with AB controls. Therefore, for AB data, the ellipses in the left and right columns are equivalent and positioned at the same coordinates for comparison with the data from those with incomplete spinal cord injury (iSCI). Each marker represents one subject and is shape-coded by types of assistive device.

FIG. 6.

Impact of types of assistive devices (AD) on end-point variability, angular coefficient of correspondence hip to knee (ACCHK) and the 10 Meter Walk Test (10MWT) speed in persons with incomplete spinal cord injury (iSCI). Ranges of ACC and speed are separated by the level of AD support. When divided into three groups according to the number of ground contacts per device (e.g., Cane: 1, LS:2, Walker:4), ACC and speed both decreased as the number of ground contacts per device decreases, whereas there was no difference among groups in end-point variability. Note that cane and LS are grouped together because there was only one cane user and the numbers of ground contacts per device were similar.

FIG. 7.

Joint ranges of motion of knee, ankle, and hip for able-bodied (AB) and incomplete spinal cord injury (iSCI) groups during overground walking at a self-selected speed.