Relationship of lower-limb muscle properties with motor control, walking and balance in stroke survivors

Abstract

Stroke survivors exhibit muscular characteristics that can influence motor function, mobility, and balance outcomes. This study aimed to quantify the muscle strength (MST) and stiffness of ankle dorsiflexors (ADF) and plantar-flexors (APF) and compare them between stroke survivors and healthy older adults, and to identify the correlation between the muscle properties of ankle muscles and stroke-related outcomes. This cross-sectional study assessed outcomes using outcome tools/measures, including a handheld dynamometer and MyotonPRO for assessing the muscle properties of the tibialis anterior (TA) and medial gastrocnemius (MG), Fugl-Meyer motor assessment of lower extremity (FMA-LE), Berg balance scale (BBS), timed up and go test (TUG), 10 m walk test (10mWT), limit of stability (LOS) test, and Oxford participation and activities questionnaire (OxPAQ). Total of 65 chronic stroke survivors and 31 healthy controls participated. The MST of the ADF demonstrated a significant positive correlation with that of APF. Additionally, the MST of the ADF and APF demonstrated significant positive correlations with motor control and certain aspects of postural stability. Conversely, it exhibited a significant negative correlation with functional mobility and walking capacity. The TA and MG stiffness was significantly negatively correlated with various aspects of postural stability. Significant differences in the ankle MST were observed between healthy individuals and stroke survivors.

Introduction

Hypertension, obesity, and dyslipidemia are the main risk factors for stroke, which is the leading cause of mortality in China with escalating burdens¹. Stroke survivors experience long- term mobility limitations, prompting assistance from other people in daily life activities². Musculoskeletal issues are among the common post-stroke consequences, with muscles becoming weak or spastic³. Changes in peripheral skeletal muscles are one of the post-stroke functional consequences⁴. Stroke survivors face difficulties with balance, community walking, and fear of falls, considerably contributed by foot and ankle impairments, such as pain and weakness⁵.

Muscle weakness, a frequent post-stroke consequence⁶, is not solely a deficiency in strength; rather, it encompasses a spectrum of characteristics⁷. Motor impairment is a significant factor contributing to the reduced performance of activities of daily living (ADL) in stroke survivors⁸. Additionally, post-stroke isometric lower extremity motor deficits have been reported to be mainly caused by weakness⁹. These impairments can increase the effort and fatigue, thereby reducing voluntary force generation.

Stiffness is a multifaceted phenomenon characterised by both “reflex-mediated” and “nonreflex-mediated” resistance to passive movements⁴. The prevalence of spasticity among stroke survivors ranges from 30% to 80%¹⁰, and is a prevalent condition in the muscles of the ankle and foot¹¹. Hence, it is crucial to examine muscle properties, including strength and stiffness, in stroke survivors. In the context of stroke rehabilitation, muscle strength and quality are correlated with the recovery of activities of daily living (ADL)¹².

The control of ankle-foot complex is crucial for sustaining balance and ensuring gait stability, with gait and balance improvements linked to increased ankle strength and flexibility¹³. Chronic foot deformities can impede ambulation in stroke survivors¹⁴. According to Ghafari et al. (2009)¹⁵. ankle plantarflexor contributes approximately 70% of the joint work while walking. Key factors in determining walking speed include the strength and velocity of plantar flexion¹⁶. Insufficient ankle dorsiflexion may cause toe dragging, reducing step length and walking velocity while elevating fall risk¹⁷. This underscores the crucial role of ankle muscles in maintaining balance and enhancing walking performance.

A major goal of stroke rehabilitation is to optimise the performance of motor tasks, including balance and walking. However, muscle properties, including muscle weakness and stiffness of the ankle dorsiflexor and plantarflexor, have not been systematically investigated in people after stroke. Additionally, no published study has systematically quantified the correlation between the muscle properties of ankle muscles and stroke-specific impairment assessments. Therefore, the objectives of this study were as follows: (1) to quantify the muscle strength and stiffness of the ankle dorsiflexor and plantarflexor in stroke survivors; (2) to compare the muscle strength and muscle stiffness between stroke survivors and healthy older adults; and (3) to ascertain the correlation between the properties of ankle muscles and their relationship with motor control of the lower limb, balance, walking performance, and participation.

Materials and methods

Study design and sample estimation

This cross-sectional study was approved by the Ethics Committee of The Hong Kong Polytechnic University (No. HSEARS20230807001). The study procedures were based on the guidelines of the Declaration of Helsinki. Informed consent was obtained from all participants, who were fully informed of the study objectives and procedures. The sample size in this study was estimated based on the correlation coefficient (r) obtained from a previous study¹⁸. A significant correlation was previously found between the peak plantar flexor torque of the paretic leg and TUG scores (r=-0.37)¹⁸. The same correlation was assumed in the current study when calculating the sample size using G*Power 3¹⁹. Thus, a sample size of 52 was required to obtain a significance level of 0.05 and a power of 80%. To enhance the robustness of the findings, the sample size was increased to 65 participants.

Subjects

Sixty-five stroke survivors were recruited from a local stroke self-help group through poster advertisements, with the following inclusion criteria: (i) aged 55–80 years; (ii) diagnosed with stroke by computed tomography or magnetic resonance imaging more than 1 year previously (i.e. chronic stroke); (iii) able to walk independently for at least 10 m; (iv) have an Abbreviated Mental Test score > 7²⁰; and (v) able to sign a written consent form. The exclusion criteria for stroke survivors were other neuromuscular (e.g. Parkinson’s disease, multiple sclerosis), cardiac, musculoskeletal (e.g. knee osteoarthritis, recent fractures), or systemic immunological (e.g. rheumatoid arthritis) conditions. Thirty-one healthy subjects with stable health conditions were also recruited. The inclusion and exclusion criteria for healthy subjects were the same as those for the stroke subjects except that they had no stroke history. All participants provided written consent after being informed of the goals and content of the study. Common conditions (e.g. diabetes mellitus, hypertension, hyperlipidaemia) were recorded for each participant.

Outcome measurements

Ankle muscle strength

Isometric muscle strength (kg) of the TA and MG of both the paretic and non-paretic legs was measured using a handheld dynamometer (Lafayette Instrument Corp., USA). The make tests were performed. The subjects were instructed to push with maximal exertion and maintain the contraction for  5 seconds, followed by a 30-second rest. The positioning of the patient was supine with 90° ankle dorsiflexion. The handheld dynamometer was placed on the dorsal/plantar surface of the first metatarsal head when conducting the ankle dorsiflexion/plantarflexion. The make test was performed for 5 seconds for each muscle in each trial. A practice trial was conducted, and then performed 2 trials, the average of the 2  trials was used to enhance the precision of the result. Peak torque has demonstrated excellent test-retest intra-observer reliability (intraclass correlation coefficient [ICC] = 0.95–0.99) for assessing lower limb strength in chronic stroke survivors²¹. The assessment was conducted by one of the five raters, all of whom were well trained, thereby ensuring consistent instructions to the participants and minimising inter-rater variability.

Ankle muscle stiffness

The muscle stiffness of the TA and MG of the subjects was measured on both sides of the legs using a MyotonPRO device (Myoton AS, Estonia and Myoton Ltd, London, United Kingdom). The test sites were located in the proximal one-third on the muscle belly from the muscle origin with the participants instructed to relax. The detailed measurement procedure was in line with previous studies^22,24. Measurements were recorded by the accelerometer at the probe²³. The MyotonPRO device has demonstrated satisfactory test-retest reliability in assessing muscle stiffness for TA and MG of chronic stroke survivors (intraclass correlation coefficient [ICC] = 0.79, 0.77)²⁴. Measurements were taken at one site/point for each muscle. For localisation of measurements sites: For the TA, the measurement site is located anteriorly, 30% of the proximal distance from the lateral intercondylar space to the lateral malleolus, with the patient in supine position²⁴. For the MG, the measurement site is located posteriorly, 30% of the proximal distance from the popliteal crease to the superior calcaneus border, with the patient in prone position²⁴. During the test, the probe delivered 5 consecutive mechanical impulses (1 Hz) to the skin. The mean stiffness value of the 5 impulses was used. The assessment was performed by one of the five raters randomly selected by a lucky draw. All the raters were well trained, thus they could provide consistent instructions to the participants and minimise inter-rater variability.

Fugl-Meyer assessment of lower extremity (FMA-LE)

The FMA-LE was used to measure the lower limb motor function of stroke survivors. It includes 17 items with a total score of 34, each rated from 0 to 2, with higher scores indicating better control recovery. It has excellent intra- and inter-rater reliability early post-stroke²⁵.

Berg balance scale (BBS)

The BBS was used to assess balance ability and fall risk in older adults. It consists of 14 items, each rated on a 5-point scale (0–4), with a maximum score of 56. The BBS demonstrates excellent internal consistency, inter-rater reliability, and intra-rater reliability²⁶. It also exhibited excellent construct validity with the Barthel Index and FIM score, along with good known-group construct validity.

Timed up and go test (TUG)

The TUG test measures the functional mobility skills of older adults. The participants, who were allowed to use walking aids, were asked to stand up from a chair with a backrest, walk 3 m, turn around, return to the chair, and sit down again at a comfortable speed. The time taken to complete the task was recorded. This test demonstrates excellent inter-rater and intra-rater reliability and shows strong construct validity with the 6-minute walk test²⁷.

10 m walk test (10mWT)

The 10mWT assesses the walking capacity of patients with various conditions. The participants were instructed to walk 10 m at their usual comfortable walking speed and maximum speed, and the means of three trials were used. Walking speed (m/s) was calculated excluding the acceleration and deceleration phases. This test has demonstrated good inter-rater reliability²⁸, with a better agreement level in the later post-stroke stages. In both TUG and 10Mwt walking aids were used for all subjects to prevent accidental falls. It was assessed by one of the five raters randomly by lucky draw. All the raters were well trained, thus they provided consistent instructions to the participants.

Limit of stability (LOS)

The LOS test used a computerised dynamic posturography machine with dual force plates and a video screen to assess the postural stability. It identified target positions based on the sway angle, recording the centre of pressure (COP) displacement to represent the theoretical LOS, while participants leaned their bodies and maintained specified targets²⁹. Five aspects of LOS were recorded: direction control (DCL) (%), indicating the smoothness of COP displacement; endpoint excursion (EPE) (%), the distance of COP movement toward the target on the first attempt; movement velocity (MVL) (m/s), the average velocity during 5% to 95% of the distance; maximum excursion (MXE) (%), the maximum distance travelled; and reaction time (RT) (s), the interval between the indication to move and the start of voluntary movement. The test shows moderate to high test-retest reliability^30,31 and correlates with BBS for concurrent validity³⁰ in people with mild stroke. The inventory information of the system is the Bertec^® Computerised Dynamic Posturography Plate (Bertec Corporation, Columbus, OH). It was assessed by 1 of the 5 raters randomly by lucky draw. All the raters were well trained and provided consistent instructions to the participants. The participants were instructed to stand barefoot on the force plate wearing harnesses and move their COP as far as possible by leaning at the ankles in eight directions (forward, backward, left, right, forward-left, forward-right, backward-left, and backward-right), one time for each direction. All participants were assessed using the same directional targets. No adjustments were made based on the paretic side in stroke participants.

Oxford participation and activities questionnaire (OxPAQ)

The OxPAQ is a patient-reported outcome measure that assesses participation and activity levels in patients with various health conditions. It includes sections on routine activities (RA) (14 items), emotional well-being (EW) (5 items), and social engagement (SE) (4 items). Higher scores indicate inferior functioning. This questionnaire demonstrated good-to-excellent internal reliability and excellent external reliability. The OxPAQ also demonstrates strong concurrent validity with MOS SF-36 and EQ-5D-5 L³². The OxPAQ is completed with assistance from the raters. All the raters were well-trained, thus they can provide consistent instruction to the participant.

Assessment procedure

The performance of the participants was measured using tests, assessment tools, and outcome measures in a university-based rehabilitation laboratory. Demographic data, muscle strength, and stiffness of the tibialis anterior (TA) and medial gastrocnemius (MG) on both legs and stroke-specific outcome measures (FMA-LE, BBS, TUG, 10mWT, LOS, and OxPAQ scores) of stroke subjects were collected on one day. In healthy subjects, only the muscle properties of both legs were examined. To prevent any learning effects, the outcome measures were collected in a random order, using a lucky draw for randomisation. The procedure was performed at the Hong Kong Polytechnic University. Five assessors/raters were involved in the data collection process, and they were provided with sufficient training.

Statistical analysis

IBM Statistical Package for the Social Sciences (SPSS) version 29 was used for data analysis. The level of significance was set at ⍺ = 0.05. The demographic characteristics, muscle strength and stiffness of the TA and MG, and other stroke-specific outcome measures of the participants were analysed using descriptive statistics. The Kolmogorov-Smirnov and Levene’s tests were used to evaluate the normality and homogeneity of data variance, respectively. Parametric and non-parametric data are presented as mean (standard deviation) and median (interquartile range), respectively. The differences in muscle properties between stroke and healthy subjects were analysed using independent t-tests and Mann-Whitney U tests, for parametric and non-parametric data respectively. The within-group differences in muscle properties between two sides of the participants were calculated using paired t-tests and Wilcoxon signed-rank tests for parametric and non-parametric data respectively. The correlation between muscle properties and other stroke-specific outcomes and the correlation among muscle properties were expressed using Spearman’s rho correlation coefficient.

Results

Demographic characteristics of all participants and stroke-related outcomes

The average age of the stroke subjects was 66.91(6.40) years, with an average post-stroke duration of 9.45(4.99) years. No significant differences were observed in any of the demographic characteristics between the stroke survivors and healthy controls. Table 1 shows the demographics of the participants and stroke-related outcome measurements.

Table 1.

Demographic characteristics and performance of stroke related outcome measures of the stroke subjects and the healthy controls.

Characteristics	Stroke (n = 65) mean (SD)	Healthy (n = 31) mean (SD)	p-value
Age, years old	66.91 (6.40)	64.84 (7.85)	0.083
Gender, male/female, n	35/30	17/14	0.927
Height, cm	162.30 (7.93)	164.93 (8.79)	0.149
Body weight, kg	63.25 (10.37)	64.63 (11.70)	0.613
Body mass index, kg/m2	24.03 (3.38)	23.77 (4.09)	0.373
Type of stroke, ischemic/haemorrhagic/mixed	40 / 24 / 1	N/A	N/A
Paretic leg, left/right, n	31 / 34	N/A	N/A
Post-stroke duration, years	9.45 (4.99)	N/A	N/A
FMA-LE	24.06 (5.07)	/
BBS	47.08 (6.04)	/
TUG, s	18.22 (10.07)	/
10mWT – usual speed, m/s	18.20 (11.57)	/
10mWT – maximum speed, m/s	14.00 (8.60)	/
LOS – DCL, %	56.71 (15.88)	/
LOS – EPE, %	43.63 (14.52)	/
LOS – MVL, deg/s	2.50 (1.01)	/
LOS – MXE, %	55.91 (17.44)	/
LOS – RT, s	1.27 (0.28)	/
OxPAQ - EW	18.36 (20.26)	/
OxPAQ - RA	19.22 (19.21)	/
OxPAQ - SE	11.52 (17.30)	/

SD: standard deviation, n: sample size, cm: centimetre, kg: kilogram, FMA-LE: Fugl-Meyer assessment of lower extremity, BBS: Berg balance scale, TUG: timed up and go test, 10mWT: 10 m walk test, LOS: limit of stability, N/A: not applicable, DCL, direction control, EPE: end-point excursion, MVL: movement velocity, MXE: maximum excursion, EW: emotional well-being, RT: reaction time, RA: routine activities, SE: social engagement.

Comparison of muscle properties between Paretic and non-paretic leg of stroke subjects

The mean ankle dorsiflexor strength of the paretic leg (9.26 ± 4.62 kg) was significantly weaker than that of the nonparetic leg (14.64 ± 4.25 kg) (p < 0.001). The mean ankle plantarflexor strength of the paretic leg (13.40 ± 6.16 kg) was also significantly weaker than that of the non-paretic leg (17.50 ± 5.16 kg) (p < 0.001). Table 2 compares the muscle properties of paretic and non-paretic legs of stroke survivors.

Table 2.

Comparison of muscle properties between Paretic and non-paretic leg of stroke subjects.

Muscle properties	Stroke (paretic) mean (SD)	Stroke (non- paretic) mean (SD)	p-value
Muscle strength (kg): ankle dorsiflexion - average	9.26 (4.62)	14.64 (4.25)	< 0.001*
Muscle strength (kg): ankle plantarflexion - average	13.40 (6.16)	17.50 (5.16)	< 0.001*
Muscle stiffness (N/m): tibialis anterior - Stiffness	366.05 (87.37)	376.88 (65.74)	0.272
Muscle stiffness (N/m): medial gastrocnemius - Stiffness	233.31 (47.56)	223.02 (32.92)	0.080

*Indicates a significant difference at the p ≤ 0.05 level of confidence.

Comparison of muscle properties between stroke subjects (paretic leg) and healthy controls (non-dominant side)

The mean ankle dorsiflexor (9.26 ± 4.62 kg) and plantarflexor (13.40 ± 6.16 kg) muscle strength of stroke survivors was significantly weaker than that of healthy older adults (p < 0.001). Table 3 shows the comparison of muscle properties for the paretic leg of stroke subjects and the non-dominant leg of healthy older adults.

Table 3.

Comparison of muscle properties between stroke subjects (paretic leg) and healthy controls (non-dominant side).

Muscle properties	Stroke (paretic) mean (SD)	Healthy (non-dominant) mean (SD)	p-value
Muscle strength (kg): ankle dorsiflexion - average	9.26 (4.62)	16.52 (5.84)	< 0.001*
Muscle strength (kg): ankle plantarflexion - average	13.40 (6.16)	20.20 (5.95)	< 0.001*
Muscle stiffness (N/m): tibialis anterior - Stiffness	366.05 (87.37)	384.66 (84.35)	0.338
Muscle stiffness (N/m): medial gastrocnemius - Stiffness	233.31 (47.56)	235.79 (38.16)	0.805

*Indicates a significant difference at the p ≤ 0.05 level of confidence.

Comparison of muscle properties between healthy controls (non-dominant leg) and stroke subjects (non-paretic leg)

The muscle strength of the plantar flexor was significantly different between stroke and healthy subjects (p = 0.015). Table 4 compares the non-dominant leg of healthy subjects and the non-paretic leg of stroke survivors.

Table 4.

Comparison of muscle properties between healthy controls (non-dominant leg) and stroke subjects (non-paretic leg).

Muscle properties	Stroke (non-paretic) mean (SD)	Healthy (non-dominant) mean (SD)	p-value
Muscle strength (kg): ankle dorsiflexion - average	14.64 (4.25)	16.52 (5.84)	0.081
Muscle strength (kg): ankle plantarflexion - average	17.50 (5.16)	20.20 (5.95)	0.015*
Muscle stiffness (N/m): tibialis anterior - Stiffness	376.88 (65.74)	384.66 (84.35)	0.740
Muscle stiffness (N/m): medial gastrocnemius - Stiffness	223.02 (32.92)	235.79 (38.16)	0.102

*Indicates a significant difference at the p ≤ 0.05 level of confidence.

Correlation within muscle properties of Paretic leg of stroke subjects

The results showed that the muscle strength of the dorsiflexor was significantly positively correlated with that of the plantar flexor (0.623, p < 0.001) (Table 5).

Table 5.

Correlation within muscle properties of Paretic leg of stroke subjects.

	Strength-DF	Strength-PF	TA-stiffness	MG-stiffness
Strength-DF	1	/	/	/
Strength-PF	0.623 p < 0.001*	1	/	/
TA-Stiffness	0.055 p = 0.666	0.123 p = 0.334	1	/
MG-Stiffness	-0.218 p = 0.083	-0.152 p = 0.231	0.121 p = 0.339	1

*Indicates a significant difference at the p ≤ 0.05 level of confidence.

Correlation of muscle properties with stroke-specific outcome

The muscle strength of the ankle dorsiflexor was positively and significantly correlated with the FMA-LE (p < 0.001), BBS, LOS-DCL, LOS-EPE and LOS-MXE scores but negatively and significantly correlated with the TUG (p < 0.001) and 10mWT (p < 0.001) scores. The muscle strength of the ankle plantarflexor was positively and significantly correlated with the FMA-LE and LOS-DCL but negatively and significantly correlated with the TUG and 10mWT. The stiffness of the TA was positively and significantly correlated with LOS-RT, and the stiffness of the MG was negatively and significantly correlated with LOS-EPE, and LOS-MXE. Table 6 shows the correlations between muscle properties and stroke-related outcome measures.

Table 6.

Correlation of muscle properties with stroke-specific outcome.

	Strength-DF	Strength-PF	TA-stiffness	MG-stiffness
FMA-LE	0.672 p < 0.001*	0.372 p = 0.002*	-0.025 p = 0.847	-0.182 p = 0.149
BBS	0.286 p = 0.022*	0.100 p = 0.431	-0.169 p = 0.181	-0.144 p = 0.256
TUG	-0.421 p < 0.001*	-0.322 p = 0.010*	0.024 p = 0.853	0.115 p = 0.364
10mWT - usual	-0.478 p < 0.001*	-0.313 p = 0.012*	-0.014 p = 0.912	0.068 p = 0.595
10mWT - max	-0.514 p < 0.001*	-0.348 p = 0.005*	0.047 p = 0.712	0.110 p = 0.387
LOS-DCL	0.333 p = 0.007*	0.252 p = 0.045*	-0.099 p = 0.436	-0.123 p = 0.335
LOS-EPE	0.314 p = 0.012*	0.213 p = 0.091	-0.196 p = 0.121	-0.287 p = 0.022*
LOS-MVL	0.214 p = 0.090	0.154 p = 0.223	-0.184 p = 0.146	-0.214 p = 0.089
LOS-MXE	0.334 p = 0.007*	0.197 p = 0.118	-0.240 p = 0.056	-0.261 p = 0.037*
LOS-RT	-0.224 p = 0.076	-0.216 p = 0.087	0.265 p = 0.035*	0.081 p = 0.523
OxPAQ-EW	0.050 p = 0.693	0.074 p = 0.560	0.064 p = 0.615	0.006 p = 0.964
OxPAQ-RA	-0.111 p = 0.385	0.069 p = 0.586	-0.040 p = 0.753	-0.113 p = 0.373
OxPAQ-SE	-0.029 p = 0.819	0.192 p = 0.129	-0.029 p = 0.818	-0.041 p = 0.749

*Indicates a significant difference at the p ≤ 0.05 level of confidence.

SD: standard deviation, n: sample size, cm: centimetre, kg: kilogram, FMA-LE: Fugl-Meyer assessment of lower extremity, BBS: Berg balance scale, TUG: timed up and go test, 10mWT: 10 m walk test, LOS: limit of stability, N/A: not applicable, DCL, direction control, EPE: end-point excursion, MVL: movement velocity, MXE: maximum excursion, EW: emotional well-being, RT: reaction time, RA: routine activities, SE: social engagement,

Discussion

This study investigated lower limb muscle properties and their correlations with functional outcomes in stroke survivors. The main findings of the study were as follows: (1) the muscle strengths of the TA and MG in the paretic leg of stroke survivors were weaker than those in the non-paretic leg and the healthy controls; (2) there were no significant differences in the muscle stiffness of the TA and MG in the paretic leg compared to the non-paretic leg and healthy controls; (3) no significant correlation was found between muscle strength and stiffness; (4) the muscle strengths of the TA and MG muscles in the paretic leg were significantly correlated with FMA-LE, TUG, 10mWT, and LOS-DCL in stroke survivors, but not with Ox-PAQ; and (5) the muscle stiffness of the TA and MG in the paretic leg did not significantly correlate with most outcome measures for stroke survivors, except for some LOS parameters with moderate correlations.

Muscle strength of TA and MG in stroke survivors

The strength of the paretic leg was weaker than the non-paretic leg and the healthy controls

Stroke survivors exhibited significantly lower ankle plantar flexion and dorsiflexion strengths in the paretic leg than in the non-paretic leg and the healthy controls. This result aligns with previous findings that paretic lower limbs of stroke survivors exhibit significantly weaker strength across all muscle groups compared to healthy controls³³. The reduction in muscle strength can be attributed to multiple factors. These include stroke, which causes an upper motor neuron lesion by damaging the motor cortex and corticospinal tract. Another reason could be neuronal apoptosis which reduces the number of neurones available for muscle fiber innervation, diminishing the descending input and motor unit firing rate. Paretic muscle neuromuscular junctions show increased polyaxonal innervation, suggesting that reduced motor unit numbers post-stroke may reflect unit overlap rather than actual losses³⁴.

In addition to causing direct neuronal damage, disuse-induced structural alterations in the muscles can indirectly affect muscle strength. Neural deficits resulting from stroke significantly modify both the input and utilisation of individual muscles³⁵. The primary factors contributing to muscle atrophy after stroke include disuse, nerve regeneration, inflammation, protein metabolism changes, fiber-type transformation, mitochondrial alterations, and nutrient supply³⁶. The observed decrease in serial sarcomeres within human muscles, particularly in individuals with neural impairments that lead to disuse and consistently maintain the muscle in a shortened position, is evident as reported in a previous study³⁵. Furthermore, an inverse relationship exists between muscle spasticity and isokinetic strength, suggesting that spastic muscles have reduced strength and require therapeutic intervention³⁷. Post-stroke spasticity and contracture continue to be prevalent issues, resulting in considerable functional impairment, pain, disability, and diminished quality of life³⁸. During contracture, the muscle loses its ability to shorten effectively, requiring greater strength for contractions, thereby diminishing force production efficiency.

The strength of the non-paretic leg was also weaker than the healthy controls

The muscle strength in the non-paretic leg of the stroke group was also weaker than that of the healthy controls, although to a lesser extent than the paretic leg. This finding was consistent with that of a previous study that reported that the intact lower limb of stroke survivors exhibited significantly reduced strength compared to that of healthy controls across nearly all muscle groups³³. The reduction in strength in the non-paretic leg could be attributed to neurological factors. About 85–90% of corticospinal tract fibres decussate at the pyramidal decussation, projecting to the contralateral side, while 10–15% of the fibres that do not decussate remain ipsilateral³⁹. Thus, if few damaged corticospinal projections persist after stroke, pathways from the nonlesioned hemisphere are likely the only resource for regaining lower limb motor control⁴⁰. However, Madhavan et al. (2010) found that higher ipsilateral conductivity from the non-lesioned hemisphere to the paretic limb would show a greater decline in tracking accuracy⁴⁰. This may explain the reduced muscle strength in the non-paretic leg compared to healthy controls.

Muscle stiffness of TA and MG in stroke survivors

Surprisingly, muscle stiffness in both the paretic and non-paretic legs did not significantly differ from that in healthy subjects. However, our results differed from those of Fröhlich-Zwahlen et al. (2014)²⁴, who reported significantly higher MG stiffness in stroke survivors than in healthy controls with MyotonPRO. Our result also contrast with those of a systematic review that reported that paretic limbs exhibit greater stiffness than non-paretic limbs or healthy control⁴¹ and in line with a study that found no significant difference in gastrocnemius stiffness between chronic stroke and control groups⁴². Several factors may explain the lack of significant differences in the TA and MG stiffness between stroke survivors and healthy controls. Our recruited subjects were community-dwelling stroke survivors who were active and independent in ADL because they were active members in the stroke self-help group. Thus, an active lifestyle may alleviate disuse atrophy and reduce the likelihood of muscle stiffness. In addition, the results measured by ultrasound elastography are not directly comparable because of different technological assumptions and models employed by these devices for calculating muscle stiffness. Moreover, assessing muscle stiffness in the resting position may produce different outcomes compared to studies that evaluate muscle stiffness at varying ankle angles or during muscle contraction. Furthermore, this study recruited participants at least one-year post-stroke. Modifying the joint angle consequently affects muscle stiffness, which is reduced following rehabilitation or pharmacological intervention in stroke survivors⁴¹.

Correlation between muscle strength and muscle stiffness

In our study, no correlation was found between muscle strength and stiffness. These findings contrast with those of previous studies that reported linear relationships between stiffness and muscle load⁴³ as well as between stiffness and force output⁴⁴. Myotonometric parameters, including stiffness and compliance, have consistently shown robust links with force production and muscle activation⁴⁵. Additionally, Ando and Suzuki (2019)⁴⁶ reported that increased passive muscle stiffness at a specific joint angle may facilitate rapid force production. A previous study also demonstrated non-significant correlation between MG shear modulus and maximal voluntary contraction torque⁴⁶, while muscle-tendon unit stiffness showed no independent relationship with voluntary explosive force production or evoked explosive force capacity⁴⁷. A study suggested that the stiffness of a material affects the speed of force transmission, with extended tissues, including muscles and tendons, exhibiting increased compliance, which may slow force transmission⁴⁸. As the current study measured isometric muscle strength using a handheld dynamometer, this may explain the lack of significant a correlation between muscle strength and the stiffness.

Correlation of muscle properties of stroke survivors with outcome measures

First, TA/MG muscle strength and stiffness were significantly correlated with FMA-LE, highlighting the correlation between impaired ankle physiology and muscle strength. The FMA-LE evaluates motor function, balance, sensation and joint function in stroke survivors. It involves significant ankle dorsiflexion to assess flexor synergy. Additionally, during the single-leg stance, both the dorsiflexor and plantar flexor were recruited to maintain a stable ankle in contact with the ground. Therefore, greater muscle strength in the TA and MG was associated with better FMA-LE performance. Importantly, it should be recognised that previous finding reported the ankle joints’ resistive stiffness correlated moderately with spasticity in dorsiflexion and plantarflexion muscles, and inversely with volitional movement with synergy in the FMA-LE among stroke survivors⁴⁹.

Second, TA/MG muscle strength and stiffness were significantly correlated with anticipatory postural adjustment (i.e. components in LOS and BBS). The ankle strategy is the primary approach for maintaining balance during body sway. For instance, when people sway or lean forward, their bodies act as inverted pendulum. The gastrocnemius is activated to restore the trunk to a vertical position. The TA is engaged similarly when swaying backward. Notably, muscles are activated in a triphasic pattern during postural control: agonists initiate movement, antagonists decelerate it and control torque, and agonists reactivate to counteract the braking torques and passive forces for stabilisation⁵⁰. Stroke survivors showed higher muscle activation and agonist-antagonist coactivation than healthy controls, with lower mean power frequency values, indicating abnormal gait and obstacle-crossing patterns⁵¹. Crucially, it has been reported that the activity of the TA is less than that of the gastrocnemius, with minimal activity after tilting initiation, suggesting that predictive postural control is primarily executed by the gastrocnemius⁵².

Regarding LOS, both dorsiflexor and plantarflexor strengths correlated particularly with DCL. It has been reported that the distinction between EPE and MXE could suggest the strategy and confidence of participants during the displacement of the COG⁵³. Only dorsiflexor strength correlated with EPE and MXE. Notably, MG stiffness was negatively correlated with EPE and MXE. One hypothesis is that the determinants of the distance travelled are ankle range of motion (ROM) and muscle strength. If the MG is stiff and spastic, the ROM is limited, thereby reducing the distance travelled. Thus, ROM becomes the limiting factor, overriding muscle strength.

For the BBS, sufficient TA and MG strength are required for anticipatory postural adjustment and dynamic balancing tasks. A previous study showed that ankle dorsiflexion strength and 1-Repetition Maximum/body weight collectively explained 39% of the variance in BBS for stroke survivors in geriatric rehabilitation⁵⁴, which aligns with our findings. However, MG strength did not correlate with BBS score, possibly due to the small sample size.

Third, muscle strength was correlated with walking and dynamic stability in the TUG test. The gait cycle can be divided into two phases: the stance phase (60% of the cycle, from heel strike to toe-off) and the swing phase (40% of the cycle)⁵⁵. The ankle plays a crucial role in effectively modulating the trajectory of the swing foot to ensure safe ground clearance⁵⁶. Individuals with restricted ankle dorsiflexion demonstrate a range of changes in the kinematic and dynamic patterns across the pelvis, hip, knee, and foot during walking and jogging⁵⁷. Weak ankle dorsiflexion can make it difficult to lift the foot and keep the knee stable, causing walking problems. Ankle plantarflexion is important for pushing off the ground during walking. If the front leg drags on the ground, it slows down the forward movement. The back leg then pushes off to make up for this loss. The ankle plantar flexor muscles are integral to human locomotion, and impairments in these muscles are believed to contribute to gait abnormalities⁵⁸. Hence, weak ankle plantar flexors can reduce walking velocity during push-off. In addition, findings of a previous study suggest that diminished muscle strength in the lower extremities has a poor impact on dynamic balance in stroke survivors⁵⁹.

Interestingly, no correlation was found between muscle stiffness and the TUG test. Compensatory movements, such as hip hiking and leg circumduction, may be utilized to facilitate walking in stroke survivors with fair muscle strength. Finally, no significant correlation was found between muscle properties and OxPAQ scores. It is a patient-reported outcome measure that assesses participation and activity levels in patients with various health conditions, whereas muscle properties reflect motor functions. The involvement of psychosocial components in the OxPAQ may explain the insignificant correlation between muscle properties and the OxPAQ score.

Limitations and further studies

This study had several limitations. First, the stroke survivors were community-dwelling older adults. Generally, this group of stroke survivors is relatively more active and healthier, implying better performance in terms of balance and walking ability. This may obscure the actual differences between the overall population of stroke survivors and healthy individuals. Future studies should consider recruiting more participants with different degrees of motor capabilities. Second, a handheld dynamometer was used to measure the isometric muscle strength of the TA and MG, which might not accurately reflect muscle strength during dynamic functions, such as walking and balance. Assessing isokinetic muscle strength could provide a better understanding of muscle strength throughout the entire ankle ROM phase. Third, MyotonPRO was used to measure stiffness, which might yield different results compared to previous studies using ultrasound elastography because of different techniques used to assess muscle stiffness. Further research is needed to compare the results of MyotonPRO and ultrasound elastography. The varying thickness of subcutaneous fat in the legs of different particiaonts may have contributed to the differences between the measurements of muscle stiffness. Finally, comorbidities were not included in the analysis; therefore, future studies should incorporate these factors.

Justification for clinical relevance

The outcome of this study would go a long way in guiding physiotherapists to focus more on ankle joint movements, muscles involved, the properties of the muscles and associated motor function, gait, and balance parameters during stroke rehabilitation. These findings also highlight the importance of assessing muscle properties including both strength and stiffness, to design effective interventions aimed at improving balance and walking performance in stroke survivors in future studies. This finding will also inform further studies on multifactorial interventions for stroke survivors, instead of solely focusing on muscle strength and coordination.

Conclusions

Our results showed that the muscle strengths of the TA and MG in the paretic leg of stroke survivors were weaker than the non-paretic leg and healthy controls, but the muscle stiffness of the affected TA and MG did not show significant differences with the non-paretic leg and the healthy controls. Additionally, the muscle strengths of the affected TA and MG muscles in stroke survivors were significantly correlated with the FMA-LE, TUG, 10mWT, LOS-DCL scores, but not with Ox-PAQ scores. Muscle stiffness of the affected TA and MG in stroke survivors was not significantly correlated with most outcome measures.

Author contributions

SSMN, PC: conception, design, data analysis, interpretation of data, drafting the work and revising it critically for important intellectual content. PHC, KKC, SHL, CYL, TKM: Data acquisition, data analysis, interpretation of data and drafting the work. JSU, CLH, KL and MMYT: data analysis, interpretation of data, drafting the work and revising it critically for important intellectual content.

Funding

This research was funded by a grant from the General Research Fund (Reference Number: 15101023) from the Research Grants Council of the Hong Kong SAR, China, awarded to Prof. Shamay S.M. Ng and her team.

Data availability

Data will be made available by the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.