Pendulum test parameters are useful for detecting knee muscle hypertonia and quantifying response to an intrathecal baclofen bolus injection.

Our objective was to determine which pendulum test parameters are useful for detecting hypertonia in the knee muscles and assessing the group and individual responses to intrathecal baclofen (ITB) bolus injection among prospective pump recipients. We included 15 neurological patients with lower limb hypertonia (mainly spinal cord injury, n  = 7) and collected data the day before (baseline), and 2.5 and 5.0 h after the 50-µg ITB bolus injection. For comparison, data were collected in 15 healthy controls. The average over six test repetitions was obtained for the number of oscillations, swing time (SwingT), amplitudes of the first flexion and extension, maximum angular velocities of the first flexion (F1V) and extension (E1V), relaxation index, and damping coefficient (DampC). Across the patient group, all pendulum parameters indicated a significant decrease in hypertonia from baseline to postinjection (analysis of variance P  ≤ 0.004), except DampC. On the basis of the cutoffs from the receiver operating characteristic curve, all parameters were good or excellent discriminators of hypertonia in patients from normotonia in controls (area under the curve ≥0.85), with the highest sensitivity for SwingT and E1V (≥93%). Furthermore, all parameters except F1V revealed a significant shift from preinjection hypertonia to postinjection normotonia among patients (McNamar test P  ≤ 0.002, DampC excluded due to missing data), with the greatest responsiveness for E1V and relaxation index (≥73%). The results confirm the overall usefulness of pendulum test parameters in this patient population and indicate that some parameters are better at detecting hypertonia (SwingT, E1V) whereas others (E1V, relaxation index) are more responsive to the ITB injection.

The contribution of walking speed versus recent stroke to temporospatial gait variability.

BACKGROUND: Inconsistent results have been reported for temporospatial gait variability after stroke. Given the large differences in gait speed across stroke subjects and relative to healthy controls, it is not clear which changes in temporospatial gait variability can be ascribed to the walking speed during gait evaluation versus the consequences of stroke. RESEARCH QUESTIONS: Does temporospatial gait variability differ between (1) stroke subjects grouped in clinically-relevant functional ambulation classes, (2) the paretic and non-paretic legs within each class, and (3) stroke and healthy subjects after controlling for gait speed? METHODS: Stroke subjects were evaluated at their comfortable speed < 2 months post-onset and classified into the household (<40 cm/s, n = 38), limited-community (40-80 cm/s, n = 35), and full-community (>80 cm/s, n = 14) walkers. Coefficients of variation (CVs) for paretic and non-paretic stance, initial double-support, and single-support times, step length, step cadence, and step width were compared across the stroke ambulation classes and between the two legs. For the parameters with significantly different CVs between stroke subjects and 33 age-matched controls walking at very-slow and free speeds, a 1-way ANCOVA was used with the gait speed as a covariate. RESULTS: For most step parameters, CVs were greater in slower stroke ambulation classes except for the smaller step width CV. The differences between the paretic and non-paretic legs emerged in slower walkers only. After controlling for the gait speed, CVs of stroke subjects no longer significantly differed from controls walking at very-slow speed. With controls walking at free speed, however, CVs for the paretic and non-paretic single-support times and the non-paretic step time remained significantly different. SIGNIFICANCE: Gait is more variable at slower speeds both in stroke subjects and healthy controls. After accounting for the free gait speed, the increased variability of only a few temporal parameters may be attributed to a recent stroke.

Reply to Daly, J.J. Comment on "Chow, J.W.; Stokic, D.S. Longitudinal Changes in Temporospatial Gait Characteristics during the First Year Post-Stroke. Brain Sci. 2021, 11, 1648".

This commentary [...].

Correction: Chow, J.W. and Stokic, D.S. Longitudinal Changes in Temporospatial Gait Characteristics during the First Year Post-Stroke. Brain Sci. 2021, 11, 1648.

The authors have made some changes in response to the reviewer's comments after the manuscript was published in December 2021 [...].

Pitching shoulder passive flexibility: torque-angle analysis for external rotation and internal rotation.

In this study, a custom device was developed to analyse the pitching shoulder's external rotation (ER) and internal rotation (IR) passive flexibility. We analysed three novel measures: the resistance onset angle (ROA = angle where the shoulder begins stretching), rotational stiffness, and torque at the end range of motion (ROM). The purpose was to conduct a bilateral analysis to determine if there are significant differences between the throwing and non-throwing shoulder. Participants were 30 upper level pitchers (13 division I, 17 minor league). During testing, pitchers laid supine on a treatment table and the arm was secured to a rotational wheel with the shoulder abducted 90° and elbow flexed 90°. Dependent t-tests revealed significant (p < 0.01) and relatively extreme bilateral differences for all three variables. The throwing shoulder had: increased ER ROA (9°), decreased IR ROA (5.3°), increased ER stiffness (17%), increased IR stiffness (34%), increased ER torque (21%), and increased IR torque (30%). Secondary correlation analysis was completed to determine if the torque-angle variables were good predictors of the end ROM. Stiffness correlations were weak for ER (r = 0.35, p = 0.048) and IR (r = 0.42, p = 0.017) but ROA correlations were strong for ER (r = 0.85, p < 0.001) and IR (r = 0.86, p < 0.001).

Longitudinal Changes in Temporospatial Gait Characteristics during the First Year Post-Stroke.

Given the paucity of longitudinal data in gait recovery after stroke, we compared temporospatial gait characteristics of stroke patients during subacute (<2 months post-onset, T0) and at approximately 6 and 12 months post-onset (T1 and T2, respectively) and explored the relationship between gait characteristics at T0 and the changes in gait speed from T0 to T1. Forty-six participants were assessed at T0 and a subsample of twenty-four participants were assessed at T2. Outcome measures included Fugl-Meyer lower-extremity motor score, 14 temporospatial gait parameters, and symmetry indices of 5 step parameters. Except for step width, all temporospatial parameters improved from T0 to T1 (p ≤ 0.0001). Additionally, significant improvements in symmetry were found for the initial double-support time and single-support time (p ≤ 0.0001). As a group, no significant differences were found between T1 and T2 in any of the temporospatial measures. However, the individual analysis revealed that 42% (10/24) of the subsample showed a significant increase in gait speed (Welch's t-test, p ≤ 0.002). Yet, only 5/24 (21%) of the participants improved speed from T1 to T2 according to speed-based minimum detectable change criteria. The increase in gait speed from T0 to T1 was negatively correlated with gait speed and stride length and positively correlated with the symmetry indices of stance and single-support times at T0 (p ≤ 0.002). Temporospatial gait parameters and stance time symmetry improved over the first 6 months after stroke with an apparent plateau thereafter. A greater increase in gait speed during the first 6 months post-stroke is associated with initially slower walking, shorter stride length, and more pronounced asymmetry in stance and single-support times. The improvement in lower-extremity motor function and bilateral improvements in step parameters collectively suggest that gait changes over the first 6 months after stroke are likely due to a combination of neurological recovery, compensatory strategies, and physical therapy received during that time.

Characteristics of rectus femoris activation and rectus femoris-hamstrings coactivation during force-matching isometric knee extension in subacute stroke.

The spectral properties of surface electromyographic (EMG) signal in the rectus femoris (RF) and the coactivation in the medial hamstrings (MH) were investigated in 45 stroke subjects (22 ± 12 days post-onset) and 30 age-matched healthy controls who performed unilateral knee extensions at maximum effort (100% MVC) and during 5-s force-matching tasks (10, 30, 50% MVC). The spectral properties were obtained through a power spectrum analysis based on Fast Fourier Transform. The coactivation was measured as the MH amplitude (%max) and MH/RF amplitude ratio. Force variability was expressed as the coefficient of variation. Both knee extensors and flexors were weaker in the paretic leg than the non-paretic and control legs (p < 0.001). A significantly higher relative power in the 5-13 and 13-30 Hz bands was found in the paretic than the non-paretic leg across all force levels (p ≤ 0.001) without changes in the 30-60 and 60-100 Hz bands or the mean and median frequencies. Regarding the antagonist coactivation, MH amplitude in the paretic leg was higher than in the non-paretic leg (submaximal levels, p < 0.0001) and the control leg (all force levels, p = 0.0005) with no differences between legs in the MH/RF ratio. The steadiness of the knee extension force was not related to the spectral properties of the agonist EMG or antagonistic coactivation. Greater coactivation was associated with weaker paretic knee flexors (p ≤ 0.0002). The overall results suggest variably altered agonist activation and antagonistic coactivation over the range of isometric knee extension contractions in subacute stroke.

Relations between knee and ankle muscle coactivation and temporospatial gait measures in patients without hypertonia early after stroke.

It is unclear whether muscle coactivation during gait is altered early after stroke and among which muscles. We sought to characterize muscle coactivation during gait in subacute stroke subjects without hypertonia and explore the relationship with temporospatial parameters. In 70 stroke (23 ± 12 days post-onset) and 29 age-matched healthy subjects, surface electromyography signals were used to calculate coactivation magnitude and duration between rectus femoris and medial hamstring (knee antagonistic coactivation), tibialis anterior and medial gastrocnemius (ankle antagonistic coactivation), and rectus femoris and medial gastrocnemius (extensor synergistic coactivation) during early double-support (DS1), early single-support (SS1), late single-support (SS2), late double-support (DS2), and swing (SW). Compared to both free and very-slow speeds of controls, stroke subjects had bilaterally decreased ankle coactivation magnitude in SS2 and duration in SS1 and SS2 as well as increased extensor coactivation magnitude in DS2 and SW. Both non-paretic knee and ankle coactivation magnitudes in SS2 moderately correlated with most temporospatial parameters (|r| ≥ 0.40). Antagonistic and synergistic coactivation patterns of the knee and ankle muscles during gait are altered bilaterally in subacute stroke subjects without lower limb hypertonia suggesting impairments in motor control. Greater coactivation magnitudes in the non-paretic knee and both ankles during the terminal stance (SS2) are associated with the overall worse gait performance. Unlike previously reported excessive coactivation or no change in chronic stroke, bilaterally decreased and increased coactivation patterns are present in subacute stroke. These findings warrant longitudinal studies to examine the evolution of changes in muscle coactivation from subacute to chronic stroke.

Knee Muscle Stretch Reflex Responses After an Intrathecal Baclofen Bolus in Neurological Patients With Moderate-to-Severe Hypertonia.

OBJECTIVES: To examine the prevalence, onset threshold, and response magnitude of stretch reflex response (SRR) in the knee extensors and flexors before and after an intrathecal baclofen (ITB) bolus injection in patients with moderate-to-severe hypertonia. MATERIALS AND METHODS: SRRs were elicited by reciprocal passive knee extension/flexion movements at preset angular velocities of 5, 60, 120, 180, 240, and 300°/s using an isokinetic dynamometer and recorded with surface electromyographic (EMG) electrodes placed over the knee extensors and flexors in 53 neurologic patients before and at 2.5 and 5 hours after an ITB injection via lumbar puncture. Outcome measures included the number of patients with presence/absence of SRRs, the number of SRRs per session, SRR onset threshold angle and velocity, and response magnitudes (peak EMG and area under the EMG curve) for each muscle. Pre-post comparisons were completed using the Fisher's exact and Wilcoxon signed rank tests. RESULTS: For both knee extensors and flexors, the proportion of patients with present SRRs (p < 0.0001) and the number of SRRs per session (p ≤ 0.027) decreased from pre- to post-ITB. The threshold velocity significantly increased post-injection in both muscles (p ≤ 0.001) without significant changes in the threshold angle. The response magnitudes significantly decreased in the knee extensors (p ≤ 0.016) but not the knee flexors after the injection. CONCLUSIONS: The prevalence and threshold velocity of SRR emerged as the most robust and practical parameters for assessing hyperreflexia during ITB bolus trial that can complement clinical assessment of muscle hypertonia.

Gait Impairments in Patients Without Lower Limb Hypertonia Early Poststroke Are Related to Weakness of Paretic Knee Flexors.

OBJECTIVE: To describe gait characteristics of patients without clinical evidence of lower limb hypertonia within 2 months of stroke and explore the relationship between gait and residual motor function. DESIGN: Cohort study. SETTING: Motion analysis laboratory in a tertiary-care free-standing rehabilitation hospital. PARTICIPANTS: Consecutive sample of 73 eligible inpatients (first-known stroke <2 months postonset, walking independently, modified Ashworth score of 0 in the paretic lower limb) and 27 healthy controls (N=100). INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: Gait speed, stride and step lengths and cadences, stance time, single-support and double-support times, and associated symmetry measures in patients at self-selected normal speed and controls at very slow speed (51.1±32.6 cm/s and 61.9±21.8 cm/s, respectively, P=.115); Fugl-Meyer lower extremity motor score (FM-LE) and isometric knee flexion and extension strength in patients. RESULTS: Except the stride/step cadence, all temporospatial parameters significantly differed between the stroke and control participants. Furthermore, significantly greater asymmetries were found in the patients for the overall stance time, initial double-support and single-support times, and step cadence, reflecting smaller values in the paretic than nonparetic limb. Most temporospatial parameters moderately to strongly correlated with the gait speed (|r|: .72-.94, P<.0001), FM-LE (|r|: .42-.62, P≤.0005), and paretic knee flexor strength (|r|: .47-.57, P≤.0004). CONCLUSIONS: Gait of patients without clinical evidence of lower limb hypertonia within 2 months of stroke is characterized by many temporospatial deviations and asymmetries. The self-selected normal gait speed, FM-LE, and paretic knee flexor strength can discriminate gait impairments in these patients shortly before inpatient discharge. It remains to be determined whether the observed relationships between paretic knee flexor strength and gait measures warrant the development of interventions for strengthening of the paretic knee flexors in order to improve gait early poststroke.

Developing reliable measures of the passive torque-angle relationship for shoulder internal and external rotation: Implications for overhead athletics.

OBJECTIVES: 1) Thoroughly assess shoulder flexibility by establishing the passive torque-angle relationship for internal and external rotation with the arm in an overhead athletics position (abducted 90°) and 2) test the reliability of four passive torque-angle measures. DESIGN: Reliability study. SETTING: Data were collected in a university biomechanics laboratory. PARTICIPANTS: Bilateral shoulder flexibility of 15 male college students (20.7 ±â€¯1.1 y) was evaluated twice in two sessions over 7-10 days. MAIN OUTCOME MEASURES: For both ER and IR, reliability was assessed bilaterally (intra-session, inter-session, and inter-tester) for the traditional range of motion measure and three novel kinetic measures: torque at end ROM, resistance onset angle, rotational stiffness. This resulted in 48 total assessments. RESULTS: Thirty-four assessments had good to excellent reliability (ICC ≥ 0.8), 10 had fair reliability (0.7 ≤ ICC < 0.8), and 4 had poor reliability (ICC< 0.7). Three of the four flexibility measures had a good overall ICC score: ROM (0.83), torque at end ROM (0.84), and resistance onset angle (0.81). The fourth, stiffness, had a fair overall reliability score (0.74). CONCLUSIONS: The passive torque-angle measures should be assimilated into clinical and research settings to determine the relevance to injury, rehabilitation, and performance.

Improvements in force variability and structure from vision- to memory-guided submaximal isometric knee extension in subacute stroke.

We examined changes in variability, accuracy, frequency composition, and temporal regularity of force signal from vision-guided to memory-guided force-matching tasks in 17 subacute stroke and 17 age-matched healthy subjects. Subjects performed a unilateral isometric knee extension at 10, 30, and 50% of peak torque [maximum voluntary contraction (MVC)] for 10 s (3 trials each). Visual feedback was removed at the 5-s mark in the first two trials (feedback withdrawal), and 30 s after the second trial the subjects were asked to produce the target force without visual feedback (force recall). The coefficient of variation and constant error were used to quantify force variability and accuracy. Force structure was assessed by the median frequency, relative spectral power in the 0-3-Hz band, and sample entropy of the force signal. At 10% MVC, the force signal in subacute stroke subjects became steadier, more broadband, and temporally more irregular after the withdrawal of visual feedback, with progressively larger error at higher contraction levels. Also, the lack of modulation in the spectral frequency at higher force levels with visual feedback persisted in both the withdrawal and recall conditions. In terms of changes from the visual feedback condition, the feedback withdrawal produced a greater difference between the paretic, nonparetic, and control legs than the force recall. The overall results suggest improvements in force variability and structure from vision- to memory-guided force control in subacute stroke despite decreased accuracy. Different sensory-motor memory retrieval mechanisms seem to be involved in the feedback withdrawal and force recall conditions, which deserves further study. NEW & NOTEWORTHY We demonstrate that in the subacute phase of stroke, force signals during a low-level isometric knee extension become steadier, more broadband in spectral power, and more complex after removal of visual feedback. Larger force errors are produced when recalling target forces than immediately after withdrawing visual feedback. Although visual feedback offers better accuracy, it worsens force variability and structure in subacute stroke. The feedback withdrawal and force recall conditions seem to involve different memory retrieval mechanisms.

Increased alertness, better than posture prioritization, explains dual-task performance in prosthesis users and controls under increasing postural and cognitive challenge.

Sensorimotor impairments after limb amputation impose a threat to stability. Commonly described strategies for maintaining stability are the posture first strategy (prioritization of balance) and posture second strategy (prioritization of concurrent tasks). The existence of these strategies was examined in 13 below-knee prosthesis users and 15 controls during dual-task standing under increasing postural and cognitive challenge by evaluating path length, 95% sway area, and anterior-posterior and medial-lateral amplitudes of the center of pressure. The subjects stood on two force platforms under usual (hard surface/eyes open) and difficult (soft surface/eyes closed) conditions, first alone and while performing a cognitive task without and then with instruction on cognitive prioritization. During standing alone, sway was not significantly different between groups. After adding the cognitive task without prioritization instruction, prosthesis users increased sway more under the dual-task than single-task standing (p ≤ 0.028) during both usual and difficult conditions, favoring the posture second strategy. Controls, however, reduced dual-task sway under a greater postural challenge (p ≤ 0.017), suggesting the posture first strategy. With prioritization of the cognitive task, sway was unchanged or reduced in prosthesis users, suggesting departure from the posture second strategy, whereas controls maintained the posture first strategy. Individual analysis of dual tasking revealed that greater postural demand in controls and greater cognitive challenge in prosthesis users led to both reduced sway and improved cognitive performance, suggesting cognitive-motor facilitation. Thus, activation of additional resources through increased alertness, rather than posture prioritization, may explain dual-task performance in both prosthesis users and controls under increasing postural and cognitive challenge.

Intrathecal baclofen bolus reduces exaggerated extensor coactivation during pre-swing and early-swing of gait after acquired brain injury.

OBJECTIVE: To characterize the concurrent activation of rectus femoris (RF) and medial gastrocnemius (MG) muscles (extensor coactivation) during gait in subjects with pronounced resting hypertonia after acquired brain injury (ABI) and examine changes after intrathecal baclofen (ITB) bolus injection. METHODS: Magnitude and duration of extensor coactivation during different phases of gait were assessed by recording gait kinematics and activity in bilateral RF and MG muscles in 18 controls and 18 ABI subjects before and at 2, 4, and 6h after a 50-µg ITB injection. RESULTS: Compared to controls, the magnitude of extensor coactivation was significantly increased in all phases of gait except the single support (p≤0.005), while the duration was significantly prolonged throughout (p≤0.001) in both legs of ABI subjects. After ITB bolus, only the duration of extensor coactivation significantly shortened in the more-affected leg during the late double-support and early swing (p≤0.026). CONCLUSIONS: Extensor coactivation is bilaterally exaggerated during gait in ABI subjects. ITB bolus effectively shortens the extensor coactivation in the more-affected leg during the pre-swing and early swing phases of gait. SIGNIFICANCE: Shortening of the prolonged extensor coactivation during gait may serve as an index of neurophysiological response to ITB bolus injection in subjects with ABI.

Variability, frequency composition, and temporal regularity of submaximal isometric elbow flexion force in subacute stroke.

We compared variability, frequency composition, and temporal regularity of submaximal isometric elbow flexion force at 10, 20, 35, and 50 % of peak torque between 34 stroke subjects (5-48 days post-onset, both arms) and 24 age-matched controls (dominant arm), and related the findings in the paretic arm to motor impairment. Force variability was quantified by the coefficient of variation (CV), frequency composition by the median frequency and relative power in 0-3-, 4-6-, and 8-12-Hz bands, and regularity by the sample entropy (SampEn). The paretic elbow flexors showed significantly increased CV and relative power in 0-3-Hz band, decreased power in 4-6- and 8-12-Hz bands, and decreased SampEn compared to both the non-paretic and control elbow flexors (P ≤ 0.0002), with no differences between the latter two (P ≥ 0.012). With increasing contraction intensity, the relative power in different frequency bands was insufficiently modulated and SampEn excessively decreased in the paretic elbow flexors. Also, CV in the paretic elbow flexors was non-linearly related to the relative power in different frequency bands and SampEn across contraction intensities (rectangular hyperbolic fit, 0.21 ≤ R 2 ≤ 0.55, P ≤ 0.006), whereas no force parameter correlated with arm motor impairment. These results largely extend our previous findings in the paretic knee extensors to the elbow flexors in subacute stroke, except that here force variability was increased only in the paretic elbow flexors and modulation of force regularity with increasing contraction intensity showed the opposite, decreasing pattern, which was considerably exaggerated in the paretic muscles.

Intersegmental coordination scales with gait speed similarly in men and women.

We applied principal component analysis (PCA) to thigh, shank, and foot elevation angles to examine the impact of speed on intra-limb coordination during gait. The specific aims were to (1) determine speed-related changes in segment loadings on three principal components (PCs) and (2) examine differences between men and women. The subjects (26 women, 21 men) walked overground at five self-selected paces (very slow, slow, normal, fast, very fast). PCA yielded percent variation (PV) explained by each PC and thigh, shank, and foot loadings on PC1-PC3. These parameters were regressed against the speed normalized to body height (BH/s) to derive individual and aggregate slopes and P values, separately for men and women. PV1 increased with speed, whereas PV2 and PV3 decreased (all P < 0.001). The loadings of thigh and foot segments on PC1 increased with speed (0.14 and 0.04 per BH/s, P < 0.001, respectively), and the loading of shank decreased (-0.10, P < 0.001). Compared to PC1, the changes in segment loadings on PC3 were the opposite (thigh -0.18, shank 0.09, foot -0.04 per BH/s, P < 0.001). The changes in segment loadings on PC2 were inconsistent and generally small. The only significance (P = 0.006), albeit a minor difference between men and women, was in the slope of thigh loading on PC2 (-0.005 ± 0.019 and 0.015 ± 0.026 per BH/s, respectively). We conclude that intersegmental coordination during gait scales with speed, with the greatest impact on the thigh segment, but no differently between men and women.

Effect of intrathecal baclofen bolus injection on ankle muscle activation during gait in patients with acquired brain injury.

BACKGROUND: Intrathecal baclofen (ITB) bolus injection effectively decreases spinal excitability but the impact on lower limb muscle activation during gait has not been thoroughly investigated. OBJECTIVE: Examine activation of medial gastrocnemius (MG) and tibialis anterior (TA) muscles during gait before and after ITB bolus injection in patients with resting hypertonia after acquired brain injury. METHODS: Lower extremity Ashworth score, temporospatial gait parameters, characteristics of the linear relationship between electromyogram (EMG) and lengthening velocity (LV) in MG during stance, and the duration and magnitude of TA-MG coactivation were assessed before and at 2, 4, and 6 hours after a 50-µg ITB injection via lumbar puncture in 8 hemorrhagic stroke and 11 traumatic brain injury subjects. RESULTS: Temporospatial gait parameters did not significantly differ across the evaluation points (P ≥ .170). However, Ashworth score (P < .001), frequency and gain of significant positive EMG-LV slope (P ≤ .020), and duration of TA-MG coactivation (P ≤ .013) significantly decreased in the more-affected leg after ITB bolus. EMG changes were not significantly different between patients who did (n = 10) and did not (n = 9) increase gait speed after the injection. The timing of the largest decrease in Ashworth score and the largest decrease in EMG parameters coincided in 36% of cases, on average. CONCLUSIONS: ITB bolus injection alters the activation of MG and TA during gait. However, the changes in muscle activation are not closely related to the changes in gait speed or resting muscle hypertonia. The analysis of ankle muscle activation during gait better characterizes the response to ITB bolus injection than gait kinematics.

Intersegmental coordination of gait after hemorrhagic stroke.

We compared gait using the planar law of intersegmental coordination between 14 hemorrhagic stroke subjects walking at a self-selected normal speed (56 ± 21 cm/s) and 15 age-matched healthy controls walking at a very slow speed (56 ± 19 cm/s). Sagittal plane elevation angles of the thigh, shank, and foot segments were submitted to principal component analysis. Additional outcome measures included the range of elevation angle and timing of peak elevation angle of the thigh, shank, and foot segments. The range of elevation angles at the shank and foot was significantly smaller in the paretic leg than non-paretic and control legs. Also, the peak elevation angle at the thigh occurred significantly later in the gait cycle in the paretic than control leg. Gait of both stroke and control subjects followed the planar law with the first two principal components explaining approximately 99% of the variance. However, the three-dimensional trajectory of elevation angles (gait loop) in stroke subjects deviated from the typical teardrop shape bilaterally, which was more exaggerated in the paretic leg. Compared to the non-paretic and control legs, the paretic leg showed significantly increased absolute loading of the thigh elevation angle and decreased absolute loadings of the shank and foot elevation angles on the first principal component, whereas the opposite was observed for the second principal component. Despite following the planar law, the gait of chronic stroke subjects is characterized by atypical timing of the thigh motion and disrupted intersegmental coordination of both legs.