Characteristics of thigh and shank affected side acceleration parameter during stroke hemiplegic gait

Abstract

Background

Despite its clinical relevance, there is a relative lack of research examining flexibility and stability based on the acceleration or speed of localized limb segments, such as the shank and thigh.

Objective

This study aimed to evaluate gait characteristics based on acceleration in the thigh and shank to identify differences between the affected and unaffected sides in stroke hemiplegic patients.

Method

Forty individuals with stroke-induced hemiplegia were assessed during a 5-meter walk using a 3D motion analysis system and Inertial Measurement Units (IMUs). Spatial-temporal and acceleration parameters were calculated.

Results

Significant differences were observed between the affected and unaffected sides in stance time, swing time, swing phase, and stance phase. In terms of acceleration, the mean acceleration in the anterior-posterior (AP) direction of the thigh and the mean value of the center of mass (CoM) in the AP direction differed significantly. These spatial-temporal findings were consistent with known characteristics of hemiplegic gait. A notable posterior shift of the thigh CoM on the affected side was identified, likely reflecting impaired propulsion and reduced stability.

Conclusion

The posterior displacement of the thigh CoM on the affected side may represent a compensatory mechanism to maintain balance during gait. Clinically, this posterior CoM shift could serve as a meaningful indicator of hemiplegic gait and a potential target for rehabilitation interventions aimed at restoring gait symmetry and improving functional mobility.

Keywords

Stroke gait spatial-temporal parameter inertial measurement unit acceleration lower-limb

1. Introduction

Stroke occurs when a blood vessel in the brain is blocked or ruptures, and it is one of the leading causes of death.¹ The incidence of stroke has increased, especially as society has aged. Each year, approximately 15 million people suffer from stroke, and about 5 million of them experience disabilities.² When a stroke occurs, damage to peripheral nerves and reduced or abnormal signal transmission from the brain lead to impaired muscle activation.³ Particularly, strokes caused by thrombosis or hemorrhage result in damage to parts of the brain, causing paralysis on one side of the body. The severity and pattern of hemiplegia vary depending on which part of the brain is affected. The main symptoms of hemiplegia include paralysis on one side of the body, decreased sensation, impaired motor ability, and sometimes speech and cognitive dysfunctions. Stroke also significantly impacts gait, one of the most essential functions for daily life, as the paralyzed leg cannot move properly, leading to difficulties in performing normal walking patterns.⁴ This is due to abnormal muscle contractions or the occurrence of spasticity.⁵ Typically, the characteristics of hemiplegic gait include reduced walking speed and step length, as well as the occurrence of circumduction gait and foot drop due to abnormal gait event ratios.⁶ The characteristics of gait on the affected side differ significantly from the un-affected side, and the affected side also greatly influences the un-affected side. Therefore, an accurate diagnosis of the severity and characteristics of the affected side is crucial for gait pattern improvement and muscle strength enhancement through gait training during rehabilitation. Various previous studies have identified the characteristics of the affected side in hemiplegic gait. Some study found that during the swing phase of gait, the knee on the affected side was lifted higher due to compensatory movements, as measured by a force plate.⁷ Another study confirmed asymmetry in the affected side of hemiplegic patients, showing reduced step length, walking speed, and ground contact time.⁸

The method of evaluating affected side characteristics using spatial-temporal parameters is commonly used.⁹ Additionally, various approaches such as comparing muscle activation through electromyography measurements and utilizing kinetic analysis with force plates have also been proposed.^10,11 Studies comparing the gait of the un-affected and affected sides after stroke play a crucial role in understanding the impact of neurological damage on gait and in setting rehabilitation goals.¹² These studies analyze the differences in muscle strength, gait patterns, and balance between the affected and un-affected sides, identifying the asymmetrical gait characteristics of stroke patients and suggesting methods to improve them. The parameters used for kinematic analysis are based on gait events, defining repetitive sections of the gait cycle to calculate spatial-temporal variables. For such analyses, optical 3D motion analysis camera systems based on infrared cameras or gait analysis systems based on force plates are commonly employed. Most measurement systems are relatively expensive and have spatial limitations, making them somewhat difficult to use in general hospitals and similar settings. To address these drawbacks, relatively simple IMU sensors have been used more recently. These sensors are attached to the limbs based on the joints of the lower extremities, and the gyroscope calculates the tilting angles or derives gait events.¹³ While traditional gait measurement systems can calculate overall spatial-temporal variables of gait, using IMU sensors allows not only the extraction of events and angle calculations but also the direct or indirect calculation of the speed and acceleration of the attached body parts.¹⁴ This enables the identification of dynamic characteristics at each segment. There have been numerous previous studies using IMU sensors attached to the trunk to examine tilting angles and left-right asymmetry.^15,16 However, there is a relative lack of research evaluating flexibility and stability based on the speed or acceleration of localized limb segments such as the shank and thigh. Measuring the acceleration of shank movement during gait plays an important role in the gait analysis of stroke patients.¹⁷ The acceleration of the shank is controlled by the multi-joints of the knee and ankle and is adjusted by the coordinated activation of muscles such as the tibialis anterior and hamstrings.^18,19 Additionally, the movement of the thigh is related to the body's balance and stability, and can be influenced by the gluteus muscles, iliopsoas, and quadriceps muscles, which affect the pelvic girdle and tilting. By understanding and predicting the rate of segmental speed change per unit time due to these factors, it can be used to assess the efficiency and safety of movement.

In this study, we aim to evaluate gait characteristics based on acceleration in the thigh and shank of stroke hemiplegic patients to identify the differences between the un-affected and affected sides at local positions. The purpose of this research is to derive new gait evaluation indicators based on the results and to facilitate the development of new gait rehabilitation strategies.

2. Materials and method

2.1. Participants

Outpatients or inpatients with hemiparesis due to stroke from Wonkwang University Gwangju Hospital (WKUGH) were selected. The inclusion criteria for participants were: 1) hemiplegia diagnosed with ischemic stroke within the last 10 years, 2) a functional ambulation category of level 3 or higher, and 3) low skin sensitivity to reduce the likelihood of issues arising from sensor attachment. A total of 40 participants were selected (Table 1). All participants were fully informed about the procedures prior to the experiment and signed an informed consent form. The study was conducted in accordance with the procedures approved by the Institutional Review Board (IRB: WKIRB 2022-07, June 29, 2022).

2.2. Data acquisition

Before the gait experiment, participants were thoroughly informed about the test protocol and practiced sufficiently to ensure they could walk as naturally as possible. The gait experiment was performed three times over a 5-meter distance at each participant's preferred walking speed in a straight line.²⁰ To collect kinematic data, an 8-camera optical infrared system for 3D motion analysis (BTS SMART DX, BTS Bioengineering Corp., Milanese, MI, Italy) was synchronized with a 4-channel inertial sensor system (Delsys Trigno Avanti wireless system, Natick, MA, USA). The sampling frequencies for the motion cameras and IMUs were set at 120 Hz and 1200 Hz, respectively, and the IMU's accelerometer and gyroscope data were resampled and synchronized with the motion data. A total of 18 reflective markers were attached according to the Helen Hayes lower limb set for the 3D motion analysis system, and 4 IMU sensors were attached to the front of the shank and thigh.²¹

2.3. Gait measures

The average value of data measured three times while walking 5-meter in a straight line at the preferred walking speed was used. Motion analysis software was used to perform the calculations from 3D motion data (BTS Motion Analysis Lab., BTS Bioengineering Corp., Milanese, MI, Italy). The acceleration values of the shank and thigh on the un-affected side and the affected side measured with the IMU. The MATLAB R2023a software (Mathworks Inc., Natick, MA, USA) was used for noise removal and parameter calculation. The acceleration value in the measured three-axis acceleration values, Medial (+)-lateral (-) (ML), superior (+)-inferior (-) (SI), and Anterior (+)-Posterior (-) (AP), were used in the analysis. To remove noise from all acceleration values, a fourth-order zero-lag Butterworth bandpass filter was applied, and the cut-off was set to 0.5 to 5 Hz to preserve the typical gait signal that exists between 0.5 and 5 Hz.^22,23 The identification of gait events and phases (stance and swing) was based on heel-strike and toe-off timings obtained from the 3D motion capture data. Specifically, the stance phase was defined from heel-strike to toe-off, and the swing phase from toe-off to the subsequent heel-strike. These events were automatically detected using the motion analysis software, which identifies gait cycle events based on the position and velocity of heel and toe markers, in line with established gait analysis protocols. Nine spatial-temporal parameters were derived from the data measured by the 3D motion analysis system: stride, stance and swing time, and gait events such as stance and swing phase ratio, mean velocity, cadence, and stride and step length. Variables derived from the IMU are basic parameters and acceleration movement parameters. The basic parameters are mean acceleration, which is the average of the measured acceleration values, maximum acceleration, which is the maximum value, and range acceleration, which is the range between the maximum and minimum values. The acceleration movement parameter is the average value representing the change in center of mass (CoM) calculated from the acceleration movement, which is the center of mass of the acceleration value itself, and the distance value obtained by integrating the acceleration value twice. The three basic parameters and the two acceleration movement parameters, total five variables are divided into ML, SI, and AP of the three axes, and there are 15 types, and since they are values measured from each thigh and shank, there are a total of 30 variables. In addition, in order to calculate the ellipse area for the CoM in the three-axis direction, the mean and covariance matrix of the data were calculated to calculate the eigenvector, and from this, the parameter settings and the ellipsoid based on the average value were expressed in the figure. This was calculated to confirm the relative tendency difference between the affected and un-affected side CoM in the thigh and shank.

The gait data of 40 hemiplegic patients were divided into two groups, un-affected and affected, based on the clinical diagnosis confirmed in advance. Paired T-test was performed for statistical comparison of all variables between un-affected and affected sides. In addition, Cohen's d was calculated to confirm the effect size.

3. Results

As a result of gait measurement, the mean velocity of all subjects was 0.78 ± 0.35 m/s, and the cadence was 98.78 ± 17.02 setps/min (mean ± standard deviation). In the results of spatial-temporal parameters, there were statistical differences in stance time, swing time, swing phase, and stance phase (Table 2).

Table 1.
Characteristics of subjects.

Characteristics Mean (SD)

Age (years) 66.85 (11.21)

Height (cm) 157.25 (9.14)

Gender (number) male: 13, female: 27

Weight (kg) 61.08 (10.48)

FAC scale (score) 4.40 (0.67)

Diagnosis date (year) 4.38 (2.07)

Characteristics	Mean (SD)
Age (years)	66.85 (11.21)
Height (cm)	157.25 (9.14)
Gender (number)	male: 13, female: 27
Weight (kg)	61.08 (10.48)
FAC scale (score)	4.40 (0.67)
Diagnosis date (year)	4.38 (2.07)

Table 2.

Spatial-temporal parameter.

Spatial-temporal parameter	Affected	Un-affected	p-value	Effect size
Stride time (s)	1.34 (0.36)	1.31 (0.36)	0.319	0.167
Stance time (s)	0.85 (0.23)	0.93 (0.36)	0.002**	1.513
Swing time (s)	0.49 (0.15)	0.41 (0.05)	0.004**	1.160
Stance phase (%)	63.40 (3.97)	67.80 (7.64)	0.002**	8.479
Swing phase (%)	36.50 (3.07)	32.23 (7.05)	0.001**	7.420
Stride length (m)	0.94 (0.27)	0.93 (0.27)	0.254	0.183
Step length (m)	0.47 (0.14)	0.46 (0.14)	0.484	0.112

Paired T-test: **p < 0.01; Effect size: Cohen's d; value: Mean (Standard deviation)

Among the acceleration basic parameters of the thigh and shank, the mean acceleration and maximum acceleration, the variable in which statistical differences were confirmed between the un-affected and affected sides was the mean acceleration of AP-thigh. The acceleration in the posterior direction was higher on the affected side than on the un-affected side (Table 3).

Table 3.

Acceleration basic parameter.

Acc. basic parameter	Affected	Un-affected	p-value	Effect size
Mean acceleration
ML-thigh (m/s²)	−0.023 (0.092)	−0.025 (0.100)	0.925	−0.295
SI-thigh (m/s²)	−1.034 (0.034)	−1.028 (0.030)	0.290	−0.289
AP-thigh (m/s²)	−0.067 (0.078)	−0.038 (0.080)	0.022*	−0.481
ML-shank (m/s²)	−0.004 (0.166)	−0.002 (0.172)	0.960	−0.471
SI- shank (m/s²)	−0.980 (0.032)	−0.983 (0.027)	0.622	−0.695
AP- shank (m/s²)	−0.277 (0.117)	−0.277 (0.068)	0.986	−0.682
Max acceleration
ML-thigh (m/s²)	0.093 (0.091)	0.091 (0.101)	0.954	−0.301
SI-thigh (m/s²)	−0.899 (0.057)	−0.896 (0.049)	0.706	−0.295
AP-thigh (m/s²)	0.152 (0.093)	0.171 (0.114)	0.280	−0.370
ML-shank (m/s²)	0.176 (0.186)	0.184 (0.187)	0.868	−0.363
SI- shank (m/s²)	−0.863 (0.060)	−0.856 (0.055)	0.498	−0.485
AP- shank (m/s²)	−0.085 (0.127)	−0.102 (0.082)	0.379	−0.475
Range acceleration
ML-thigh (m/s²)	0.235 (0.063)	0.231 (0.069)	0.630	−0.234
SI-thigh (m/s²)	0.264 (0.082)	0.247 (0.070)	0.107	−0.230
AP-thigh (m/s²)	0.421 (0.120)	0.399 (0.100)	0.234	−0.056
ML-shank (m/s²)	0.359 (0.092)	0.383 (0.101)	0.158	−0.055
SI- shank (m/s²)	0.238 (0.094)	0.244 (0.096)	0.647	−0.123
AP- shank (m/s²)	0.383 (0.112)	0.359 (0.116)	0.076	−0.120

Acc.: Acceleration; AP, anterior-posterior direction; ML, medial-lateral direction; SI, superior-inferior direction; Paired T-test: *p < 0.05; Effect size: Cohen's d; value: Mean (Standard deviation)

Table 4.

Acceleration movement parameter.

Acc. movement parameter	Affected	Un-affected	p-value	Effect size
Acceleration movement
ML-thigh (m/s²)	0.219 (0.405)	0.080 (0.492)	0.201	0.206
SI-thigh (m/s²)	0.363 (0.086)	0.363 (0.089)	0.826	0.202
AP-thigh (m/s²)	0.290 (0.206)	0.265 (0.315)	0.693	0.035
ML-shank (m/s²)	0.074 (0.417)	0.182 (0.392)	0.309	0.034
SI- shank (m/s²)	0.367 (0.088)	0.369 (0.090)	0.455	0.063
AP- shank (m/s²)	0.358 (0.138)	0.379 (0.191)	0.468	0.062
CoM
ML-thigh (m)	0.001 (0.010)	0.002 (0.009)	0.367	−0.144
SI-thigh (m)	0.098 (0.042)	0.097 (0.042)	0.582	0.088
AP-thigh (m)	−0.006 (0.007)	−0.003 (0.006)	0.033*	−0.349
ML-shank (m)	0.007 (0.018)	0.004 (0.016)	0.343	0.152
SI- shank (m)	0.093 (0.040)	0.094 (0.041)	0.354	−0.148
AP- shank (m)	−0.025 (0.014)	−0.025 (0.012)	0.816	0.037

AP, anterior-posterior direction; ML, medial-lateral direction; SI, superior-inferior direction; Paired T-test: *p < 0.05; Effect size: Cohen's d; value: Mean (Standard deviation), CoM: Center of Mass

In Acceleration movement, which is the center of mass of the acceleration value itself, which is the acceleration center parameter of the thigh and shank, there was no statistical difference between the un-affected and affected sides. In CoM, which is calculated from the distance value obtained by integrating the acceleration value twice, there was a statistical difference in AP-thigh between the affected and un-affected sides (Table 4).

Figure 1 is the result of fitting the ellipse area with the CoM value calculated from the distance value, which is the integral value of the velocity calculated from the acceleration. Based on the transverse plane viewed from above the head, the area of the affected side of the thigh was larger than that of the un-affected side, and the CoM increased in the posterior and lateral directions. The difference in the area of the shank was not large, and the CoM of the affected side showed a tendency to slightly increase in the medial direction.

Figure 1.

Ellipse area of affected and un-affected side center of mass.

4. Discussion

In this study, in addition to the spatial-temporal parameters, which are commonly used variables in comparing the un-affected and affected sides of stroke hemiplegia gait, we conducted a basic study to derive parameters that can be evaluated at the local location of the patient's lower limb from acceleration data measured using a simplified sensor, IMU.

First, in the spatial-temporal parameters, which are common gait evaluation variables, the variables that could confirm significant differences between the un-affected and affected sides were stance and swing time, and stance and swing phase ratio. We confirmed the same trend as the shortening of the stance phase interval and time, which are characteristics of the affected side of stroke hemiplegia patients in general. This is the same as the result of gait asymmetry that occurs when the time to support the weight with the paralyzed leg becomes shorter due to muscle weakness and balance instability, as suggested in various previous studies.^24,25 Since the subjects participating in this experiment showed the characteristics of common hemiplegic patients, it can be interpreted as the result of appropriate subject selection.

Second, the result of calculating the mean, max, and range of the acceleration raw data, which is the Acceleration basic parameter, a significant difference was confirmed between the affected side and the un-affected side only in the mean value of the Thigh in the AP direction. The affected side showed a larger acceleration value on average in the posterior direction. The section where the Thigh increases to a negative value in the opposite direction of the gait progression is the section from mid stance to terminal stance. It is thought that the characteristics of the affected side tend to be that the phase ratio and time of the mid stance tend to be shorter, and that it tends to end the mid stance with a larger acceleration on average.²⁶ However, since there was no statistical difference in the variable indicating the range of the maximum and minimum values of acceleration, additional study is necessary.

Third, the statistical comparison result of the acceleration movement parameter, the center of mass value defined as the average value of the distance calculated from the acceleration also showed a statistically significant difference between the affected side and the un-affected side only in the AP direction value of the Thigh, similar to the mean acceleration. The center of mass refers to the center of mass calculated from the distance value. When the center point for the three-axis value was considered, it increased in the posterior direction. This is related to the weakened muscle strength and difficulty maintaining balance due to paralysis.²⁷ Due to the influence of the weakened extensor muscle, the propulsive force to move the body forward on the affected side is insufficient, causing the center of mass to shift backward.²⁸ In addition, foot drop, etc. occur due to the decline in dorsiflexion function, which can also cause the center of mass to shift posteriorly. This can be explained by the results of a previous study.²⁹

Fourth, Figure 1 confirms the characteristics of the thigh and shank by the size of the ellipse area. Since the absolute numerical results of each subject are different depending on the characteristics such as severity and body length, the tendency of area movement was confirmed for relative comparison, and the thigh-shank coordination.³⁰ This is thought to be because the center of mass of the body is located in the pelvis during normal walking, so the change in the center of mass is more evident in the thigh, which is closer to it.³¹ In addition, it is thought that the thigh, which induces the hip and knee angle, has a greater influence on movement than the shank, which is bound to the ankle.

The limitations of this study are as follows. First, the measured acceleration data was not verified based on the section set as a gait event. This is because the purpose of the result derivation method in this study was to compare the overall un-affected and affected sides using acceleration in the entire 5-m walking section rather than gait events. In a follow-up study, we plan to verify which walking phase the results actually occur in based on the data set as gait events. Second, there is a possibility of error in the values used by integrating the acceleration data of the IMU sensor. This is a shortcoming of most IMU sensors. However, it can be compensated to some extent by applying various filters. In future studies, it is thought that the shortcomings in data post-processing can be compensated for by applying more sophisticated filters. Additionally, given the possibility of Type I error, it is important to review the results by applying the Bonferroni correction. However, given the relatively small number of participants in this study, we believe that the results of the current statistical analysis should be interpreted at an exploratory level rather than generalizable.

5. Conclusion

In this study, rather than directly using acceleration data to identify the characteristics of the affected side of hemiplegic patients, we confirmed that gait evaluation is possible using the CoM value calculated from the distance value. In particular, the increase in the average CoM value in the posterior direction of the thigh, which reflects the characteristics of the affected side, is a gait characteristic of hemiplegic stroke patients, so it can be used as a criterion for gait rehabilitation treatment that aims to eliminate the difference between the un-affected side and the affected side. In future studies, we aim to develop more generalizable diagnostic criteria using the results targeting a larger number of stroke patients.

Footnotes

Acknowledgements

This research was supported by a grant of the Korea Health Technology R&D Project through the Korea Health Industry Development Institute (KHIDI), funded by the Ministry of Health & Welfare, Republic of Korea (grant number : RS-2020-KH088006)

ORCID iDs

Jeong-Woo Seo

Sangkwan Lee

Ethical approval

The studies involving human participants were reviewed and approved by the Ethics Committee of Wonkwang University Gwangju Hospital (WKUGH) (IRB No.: WKIRB 2022-07, June 29, 2022). The patients/participants provided their written informed consent to participate in this study.

Author contributions

SEO JS: writing—original draft, formal analysis, investigation, Lee S: writing—review and editing, and visualization, conceptualization and resources, supervision, and funding acquisition. All authors have read and agreed to the published version of the manuscript.

Funding

The author disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Korea Health Industry Development Institute, (grant number RS-2020-KH088006).

Conflicts of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Data availability statement

All relevant data is contained within the article: The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author/s.

References

Capirossi

Laiso

Renieri

, et al. Epidemiology, organization, diagnosis and treatment of acute ischemic stroke. Eur J Radiol Open 2023 Oct 10; 11: 100527.

Feigin

Brainin

Norrving

, et al. World stroke organization (WSO): global stroke fact sheet 2022. Int J Stroke 2022 Jan; 17: 18–29.

Lee

Triandafilou

Lock

, et al. Impairment in task-specific modulation of muscle coordination correlates with the severity of hand impairment following stroke. PLoS One 2013 Jul 16; 8: e68745.

Woolley

. Characteristics of gait in hemiplegia. Top Stroke Rehabil 2001 Winter; 7: 1–18.

Moore

Hickey

Lord

, et al. Comprehensive measurement of stroke gait characteristics with a single accelerometer in the laboratory and community: a feasibility, validity and reliability study. J Neuroeng Rehabil 2017 Dec 29; 14: 30.

Mohan

Khandoker

Wasti

, et al. Assessment methods of post-stroke gait: a scoping review of technology-driven approaches to gait characterization and analysis. Front Neurol 2021 Jun 8; 12: 650024.

Bensoussan

Mesure

Viton

, et al. Kinematic and kinetic asymmetries in hemiplegic patients’ gait initiation patterns. J Rehabil Med 2006 Sep; 38: 287–294.

Høyer

Opheim

Strand

, et al. Temporal and spatial gait parameters in patients dependent on walking assistance after stroke: reliability and agreement between simple and advanced methods of assessment. Gait Posture 2014; 40: 101–106. Epub 2014 Mar 12.

van Uden

Besser

. Test-retest reliability of temporal and spatial gait characteristics measured with an instrumented walkway system (GAITRite). BMC Musculoskelet Disord 2004 May 17; 5: 13.

10.

Klein

, et al. Editorial: electromyography (EMG) techniques for the assessment and rehabilitation of motor impairment following stroke. Front Neurol 2018 Dec 18; 9: 1122.

11.

Fotiadou

Aggeloussis

Gourgoulis

, et al. Reproducibility of gait kinematics and kinetics in chronic stroke patients. NeuroRehabilitation 2018; 42: 53–61.

12.

Rastegarpanah

Scone

Saadat

, et al. Targeting effect on gait parameters in healthy individuals and post-stroke hemiparetic individuals. J Rehabil Assist Technol Eng 2018 May 23; 5: 2055668318766710.

13.

Yang

Zhang

Novak

, et al. Estimation of spatio-temporal parameters for post-stroke hemiparetic gait using inertial sensors. Gait Posture 2013 Mar; 37: 354–358. Epub 2012 Sep 20.

14.

Simonetti

Bergamini

Vannozzi

, et al. Estimation of 3D body center of mass acceleration and instantaneous velocity from a wearable inertial sensor network in transfemoral amputee gait: a case study. Sensors (Basel) 2021 Apr 30; 21: 3129.

15.

Seo

Kim

, et al. Principal characteristics of affected and unaffected Side trunk movement and gait event parameters during hemiplegic stroke gait with IMU sensor. Sensors (Basel) 2020 Dec 21; 20: 7338.

16.

Trojaniello

Ravaschio

Hausdorff

, et al. Comparative assessment of different methods for the estimation of gait temporal parameters using a single inertial sensor: application to elderly, post-stroke, Parkinson's disease and Huntington's disease subjects. Gait Posture 2015 Sep; 42: 310–316.

17.

Pieper

Lewek

Franz

. Can shank acceleration provide a clinically feasible surrogate for individual limb propulsion during walking? J Biomech 2020 Jan 2; 98: 109449.

18.

Lewek

Schmit

Hornby

, et al. Hip joint position modulates volitional knee extensor muscle activity after stroke. Muscle Nerve 2006 Dec; 34: 767–774.

19.

Sánchez

Acosta

Lopez-Rosado

, et al. Lower extremity motor impairments in ambulatory chronic hemiparetic stroke: evidence for lower extremity weakness and abnormal muscle and joint torque coupling patterns. Neurorehabil Neural Repair 2017 Sep; 31: 814–826.

20.

Cheng

Dagenais

Alsbury-Nealy

, et al. Distance-limited walk tests post-stroke: a systematic review of measurement properties. NeuroRehabilitation 2021; 48: 413–439.

21.

Seo

Kang

Kim

, et al. Characteristics of gait event and muscle activation parameters of the lower limb on the affected Side in patients with hemiplegia after stroke: a pilot study. Arch Rehabil Res Clin Transl 2023 Aug 18; 5: 100274.

22.

Yuwono

Moulton

, et al. Gait cycle spectrogram analysis using a torso-attached inertial sensor. Annu Int Conf IEEE Eng Med Biol Soc 2012; 2012: 6539–6542.

23.

Charry

Lai

Begg

, et al. A study on band-pass filtering for calculating foot displacements from accelerometer and gyroscope sensors. Annu Int Conf IEEE Eng Med Biol Soc 2009; 2009: 4826–4827.

24.

Patterson

Gage

Brooks

, et al. Evaluation of gait symmetry after stroke: a comparison of current methods and recommendations for standardization. Gait Posture 2010 Feb; 31: 241–246.

25.

Olney

Richards

. Hemiparetic gait following stroke. Part I: characteristics. Gait Posture 1996 April; 4: 136–148.

26.

Wang

Yue

, et al. Associations between lower-limb muscle activation and knee flexion in post-stroke individuals: a study on the stance-to-swing phases of gait. PLoS One 2017 Sep 8; 12: e0183865.

27.

Francisco

Zhou

. Post-stroke hemiplegic gait: new perspective and insights. Front Physiol 2018 Aug 2; 9: 1021.

28.

Devetak

Bohrer

RCD

Rodacki

ALF

, et al. Center of mass in analysis of dynamic stability during gait following stroke: a systematic review. Gait Posture 2019 Jul; 72: 154–166.

29.

do Carmo

Kleiner

Barros

. Alteration in the center of mass trajectory of patients after stroke. Top Stroke Rehabil 2015 Oct; 22: 349–356.

30.

Hutin

Pradon

Barbier

, et al. Lower limb coordination patterns in hemiparetic gait: factors of knee flexion impairment. Clin Biomech (Bristol) 2011 Mar; 26: 304–311.

31.

Gard

Miff

Kuo

. Comparison of kinematic and kinetic methods for computing the vertical motion of the body center of mass during walking. Hum Mov Sci 2004; 22: 597–610.