Differentiation between non-neural and neural contributors to ankle joint stiffness in cerebral palsy
© de Gooijer-van de Groep et al.; licensee BioMed Central Ltd. 2013
Received: 24 August 2012
Accepted: 14 June 2013
Published: 23 July 2013
Spastic paresis in cerebral palsy (CP) is characterized by increased joint stiffness that may be of neural origin, i.e. improper muscle activation caused by e.g. hyperreflexia or non-neural origin, i.e. altered tissue viscoelastic properties (clinically: “spasticity” vs. “contracture”). Differentiation between these components is hard to achieve by common manual tests. We applied an assessment instrument to obtain quantitative measures of neural and non-neural contributions to ankle joint stiffness in CP.
Twenty-three adolescents with CP and eleven healthy subjects were seated with their foot fixated to an electrically powered single axis footplate. Passive ramp-and-hold rotations were applied over full ankle range of motion (RoM) at low and high velocities. Subject specific tissue stiffness, viscosity and reflexive torque were estimated from ankle angle, torque and triceps surae EMG activity using a neuromuscular model.
In CP, triceps surae reflexive torque was on average 5.7 times larger (p = .002) and tissue stiffness 2.1 times larger (p = .018) compared to controls. High tissue stiffness was associated with reduced RoM (p < .001). Ratio between neural and non-neural contributors varied substantially within adolescents with CP. Significant associations of SPAT (spasticity test) score with both tissue stiffness and reflexive torque show agreement with clinical phenotype.
Using an instrumented and model based approach, increased joint stiffness in CP could be mainly attributed to higher reflexive torque compared to control subjects. Ratios between contributors varied substantially within adolescents with CP. Quantitative differentiation of neural and non-neural stiffness contributors in CP allows for assessment of individual patient characteristics and tailoring of therapy.
KeywordsCerebral palsy Movement disorder Ankle joint Stiffness Spasticity Contracture Neuromechanics System identification Neuromuscular modeling
Cerebral palsy (CP) comprises a variety of non-progressive upper motor neuron (UMN) lesions occurring in the developing fetal or infant brain. The resulting movement and posture disorders are generally characterized by loss of muscle strength, i.e. paresis, improper muscle activation by e.g. increased reflexes and loss of coordination by e.g. flexion synergies. In addition, changes of tissue viscoelastic properties may modulate the characteristics of the primary motor disorders [1, 2]. Spastic CP is the most common type of CP , which is characterized by increased joint stiffness (resistance to movement). Increased joint stiffness in the relaxed condition can be of either neural (hyperreflexia, “spasticity”) or non-neural origin (altered tissue viscoelastic properties “contracture”) . Treatment of spastic CP is generally aimed at diminishment of joint stiffness in order to improve passive and active joint range of motion. In case of suspected neural origin, therapy is aimed at reducing muscle activation and blocking the stretch reflex loop by botulinum toxin , intra thecal baclofen  or selective dorsal rhizotomy . In case of suspected non-neural origin, i.e. changes in viscoelastic properties of muscle and connective tissues, corrective casting, splinting and surgical lengthening can be applied . Current manual tests, like the Ashworth  and Tardieu , are based on the paradigm of increased reflex activity as a result of neural damage, leading to a velocity dependent joint resistance or spasticity . This paradigm is however an oversimplification . Inherently, by manual testing, it is not possible to quantitatively discriminate between underlying neural and non-neural contributors to joint stiffness as each of these contributors may generate a velocity dependent joint resistance. This makes the selection of treatment aiming at the dominant contributor difficult.
De Vlugt et al.  developed an instrumented method to quantify neural and non-neural contributors to joint stiffness for the ankle joint in patients with chronic stroke. The ankle was rotated in a precise and controlled way using a robotic manipulator. Using neuromuscular modeling, the key neural and non-neural contributors to ankle joint stiffness were quantified from recorded ankle torque and EMG of leg (below the knee) muscles. Compared to healthy subjects, patients with stroke showed increased tissue stiffness and to a lesser extent increased reflex activity.
The objective of the present study was to quantify neural and non-neural contributors to ankle joint stiffness in patients with spastic CP and to assess its validity and reliability. A quantitative discrimination between the neural and non-neural components of joint stiffness in CP gives insight in pathophysiological mechanisms and may provide a strong instrument for development of tailored intervention strategies and their follow-up.
Characteristics of study population
Cerebral palsy (n = 23)
Healthy subjects (n = 11)
Age, mean (SD)
Male gender, n (%)
Unilateral, n (%)
GMFCS* I, II, III, n (%)
Ashworth, median (range)
SPAT**, median (range)
Measurements were performed on the most affected ankle of each patient and at the right ankle in case of controls. Maximum plantar and dorsal flexion angles were assessed by a gradually increasing flexion torque from 0 to a maximum value of 15 Nm. RoM was defined as the difference between the maximum plantar and dorsal flexion angle and used as boundary for the subsequent RaH rotations. During the RaH rotations, the ankle was rotated at 4 different angular velocities (15, 30, 60, 120 deg/sec) over the individually assessed RoM, starting in maximal plantar flexion. RaH rotations were started at random time instants. The hold phase lasted 4 seconds after which the ankle was moved back again to the neutral position. Time to cover a complete RaH rotation did not exceed 15 sec. Rest periods of about 30 sec were introduced between each RaH rotation to avoid hysteresis effects . All RaH rotations were performed twice. Thus, the complete experimental procedure consisted of 1 RoM and 2 times 4 RaH rotations. Subjects were asked to remain relaxed during the entire experiment and not actively resist any motion. EMG prior to RaH rotation was offline checked to be between minus and plus 3 times standard deviation from the lowest EMG value over the whole signal as determined by a moving average procedure (window width 1 sec.). RaH rotations not fulfilling this requirement were discarded from further analysis.
Model description and validation
With i the sample time and n the number of data points used for the parameter estimation. Tmeasured,i is the measured ankle reaction torque and Tmodel,i the predicted ankle torque. Those rotations exhibiting a VAF score lower than the mean VAF over all rotations minus 2 times standard deviation were excluded from analysis.
Primary outcome parameters were RoM, tissue stiffness and viscosity and torque from triceps surae (TS) and tibialis anterior (TA) stretch reflexes. As passive tissue stiffness and viscosity strongly depend on joint angle, values at the maximal common dorsal flexion angle of all subjects were calculated for inter-subject analysis. This particular angle was chosen as exhibiting probably the largest contrast between subjects . Model simulations and data analyses were performed in MATLAB (The Mathworks Inc., Natick MA). An extensive validity and reliability analysis of the used method and the estimated model parameters was performed previously .
Difference in RoM between patients with CP and healthy controls was tested using an unpaired t-test. A linear mixed model was used to determine the difference in primary outcome variables between healthy controls and patients with CP (random factor) and to assess the effect of velocity (fixed, repeated factor). Stepwise linear regression procedures and one way ANOVA with Bonferroni correction were applied to assess associations of primary outcome variables with RoM and secondary outcome variables i.e. speed of ankle rotation, age, gender, GMFCS , Ashworth  and spasticity test (SPAT)  scores. Inter-trial variability was assessed using intraclass correlation coefficients (ICC, 2-way mixed model). Statistical analysis was performed using SPSS 17.0 (SPSS Inc.) and GraphPad Prism 5 (Graphpad Software) with a significant level of .05.
One subject (healthy) could not complete the RaH measurement due to insufficient relaxation and in this particular case only the RoM was used. In 2 other healthy subjects 1 RaH rotation had to be excluded due to technical problems. In total, 16 of the 256 RaH rotations from 9 subjects (8 CP), were excluded due to poor model fits (10, 4% of 256 RaH rotations) or insufficient relaxation (6, 2 % of 256 RaH rotations). The VAF of the remaining RaH rotations was above 98.9%, meaning that the model could well describe ankle torque dynamics.
Range of motion
Non-neural contributors to joint stiffness: tissue stiffness and viscosity
Tissue stiffness was independent of velocity (F = 0.35, df = 3, p = .79) and was significantly larger in CP compared to healthy controls (F = 6.28, df = 1, p = .018), see Figure 2 (bottom left). There was a large variation in tissue stiffness within the CP group. Viscosity decreased with angular velocity (F = 9.86, df = 3, p < .001). We found no significant difference between the groups regarding ankle viscosity (F = 1.35, df = 1, p = .254).
Neural contributor to joint stiffness: reflexive torque
TS reflexive torque (Figure 2, top right) was higher in CP than in healthy controls (F = 11.6, df = 1, p = .002) and the difference increased with velocity (F = 4.61, df = 3, p = .009). TS reflexive torque showed a large variation within the group with CP. TA reflexive torque was not significantly different between CP and healthy controls (F = 2.864, df = 1, p = .104) and did not change with velocity (F = 0.602, df = 3, p = .620).
Relation tissue stiffness and range of motion
Relation tissue stiffness and reflexive torque
For patients with CP, tissue stiffness at the lowest ankle rotation speed (15 deg/sec) was on average 4.2 times higher than reflexive torque at the highest ankle rotation speed (120 deg/sec) with a standard deviation of 3.3 indicating a substantially variation between subjects. Total explained variation of reflexive torque (120 deg/sec) by tissue stiffness (15 deg/sec) was 38%. Association between tissue stiffness and reflexive torque was low (ICC: less than .5).
Relation with clinical phenotype
Tissue stiffness showed a good conformity between the two repetitive RaH rotations especially at the lowest ankle rotation speed: ICC .93 at 15 deg/sec. For reflexive torque, inter-trial reliability was especially good at the highest ankle rotation speed: ICC .80 at 120 deg/sec. Reliability was similar for CP and healthy subjects.
Ankle joint stiffness in CP was successfully separated into its neural and non-neural components using an instrumented and model based approach. Compared to healthy subjects, patients with CP showed a smaller RoM, higher TS reflexive torque and higher tissue stiffness. Ratios between contributors varied substantially within the group with CP.
Higher tissue stiffness and smaller range of motion in cerebral palsy
Previously, in larger groups of children with CP, RoM was associated with level of spasticity as expressed by Ashworth score and GMFSC I-II [16–19]. Decreased RoM in CP is explained by increased passive tissue stiffness , likely originating from changes in the mechanical property of fiber bundles and/or fewer sarcomeres (in series) which might result in increased sarcomere length [20–22] or actively, i.e. hypertonia . In vivo measurements in CP show that muscles appeared to undergo much higher stresses with increased muscle length  and torque-angle relationships are much steeper in CP , which is supported in our study by the correlation between tissue stiffness and RoM (Figure 3). RoM measurements at low speed may therefore represent passive tissue stiffness, i.e. “static” contracture . Note that despite the instruction to the subjects to relax, tissue stiffness may be modulated by a constant level of (increased muscle activation) . Separation of these components requires further effort.
Clinical implication: variation in tissue stiffness and reflexive torque
Even in this relatively mild affected group of adolescents with CP, a large inter-subject variation was found for the ratio between TS reflexive torque and tissue stiffness. Association between the two was low. This variation is the rationale for pursuing the development of personalized therapy. A variation in CP may be induced by ageing and the corresponding growth spurt in puberty since we measured adolescents in a wide age range (12–19 years). It was suggested previously that the role of passive stiffness may increase over reflex activity with age in children with spastic diplegic CP  and that the range of dorsiflexion of the ankle joint in CP decreases on average 19 deg during the first 18 years of life . We found an association of reflex activity but not tissue stiffness with age in the present study. The present study was however not designed to study age effects.
Correlation of SPAT score with both tissue stiffness and reflexive torque underlines the fact that it is difficult to split the neural and non-neural component with the manual tests such as SPAT. In future work we will measure more patients and with a wider range of GMFCS and SPAT scores to study this correlation more extensively.
The present and former  studies show that the instrumented approach can be used in different patient groups to quantitatively determine neural and non-neural contributors of ankle joint stiffness.
Non-neural and neural components in cerebral palsy compared to stroke
TS reflexive torque was more dominant in CP than in stroke . For the patients with an Ashworth score of 1, the ratio between TS reflexive torque at the highest ankle rotation speed and tissue stiffness at the lowest ankle rotation speed was for CP three times higher than for stroke (CP ≈ 0.3 and stroke ≈ 0.1). In contrast to stroke, viscosity was not significantly increased in CP. RoM was smaller in stroke compared to CP, reflecting also the higher tissue stiffness component in the stroke group. CP differs from stroke by onset of the disease with respect to age. The main question is whether the differences between stroke and CP may be explained by purely an age effect or whether there might be etiological differences.
Reliability and validity
Tissue stiffness and reflexive torque could both be reliably estimated: tissue stiffness especially at the low ankle rotation speed and reflexive torque at the high speeds. This illustrates the feasibility of the method to distinguish contributors to joint stiffness on an individual basis. As expected, reflexive torque and not tissue stiffness was significantly influenced by ankle rotation speed and especially tissue stiffness was associated with RoM at low speed. Significant associations of SPAT score with both tissue stiffness and reflexive torque show agreement with clinical phenotype.
We selected patients with CP with a relatively high GMFSC score (median of 1) to create a homogenous population. This limits validity of the present study and prevents more extensive elaboration of the relation with clinical phenotype which will be important for goal directed therapy. There were only 3 patients with a GMFCS score higher than 1, so correlations of GMFCS with neuromuscular parameters could not be studied well. The present study did not differentiate contributions of gastrocnemius and soleus and was performed at 1 knee angle. Joint stiffness should also be assessed in relation to functional movements, like walking using static and dynamic measurements . In this study we were able to split neural and non-neural contributors to increased joint stiffness. However, the neural component in this study comprised only the reflex activity and not cross bridge dynamics and background muscle activity during passive (and active) conditions. Achilles tendon stiffness was assumed to be of a magnitude greater than the total (active and passive) muscle stiffness for the current conditions applied in this study  and was taken in our model as infinitely stiff, considering the low torque and passive conditions  applied in the present study.
Future work will comprise the assessment of Achilles tendon stiffness by ultrasound measurements, the assessment of joint stiffness as a function of ankle rotation angle in a more detailed way, measurements at different knee angles and under functional (loaded) conditions.
Using a novel instrumented assessment technique, patients with CP showed a smaller RoM and higher tissue stiffness and reflexive torque compared to control subjects. Good reliability and validity of the assessment technique combined with considerable intra-individual variance are a base for individual tailored therapy.
This research is supported by the Dutch Technology Foundation STW, which is part of the Netherlands Organisation for Scientific Research (NWO) and partly funded by the Ministry of Economic Affairs, Agriculture and Innovation.
- Rosenbaum P, Paneth N, Leviton A, Goldstein M, Damiano D, Dan B, Jacobsson B: A report: the definition and classification of cerebral palsy april 2006. Dev Med Child Neurol Suppl 2007, 109: 8-14.PubMedGoogle Scholar
- Goldstein M: The treatment of cerebral palsy: what we know, what we don't know. J Pediatr 2004, 145: S42-S46. 10.1016/j.jpeds.2004.05.022View ArticlePubMedGoogle Scholar
- SCPE: Prevalence and characteristics of children with cerebral palsy in europe. Dev Med Child Neurol 2002, 44: 633-640.Google Scholar
- Dietz V, Sinkjaer T: Spastic movement disorder: impaired reflex function and altered muscle mechanics. Lancet Neurol 2007, 6: 725-733. 10.1016/S1474-4422(07)70193-XView ArticlePubMedGoogle Scholar
- Sheean G: Botulinum toxin should be first-line treatment for poststroke spasticity. J Neurol Neurosurg Psychiatry 2009, 80: 359.View ArticlePubMedGoogle Scholar
- Albright AL, Barron WB, Fasick MP, Polinko P, Janosky J: Continuous intrathecal baclofen infusion for spasticity of cerebral origin. JAMA 1993, 270: 2475-2477. 10.1001/jama.1993.03510200081036View ArticlePubMedGoogle Scholar
- Mittal S, Farmer JP, Al-Atassi B, Gibis J, Kennedy E, Galli C, Courchesnes G, Poulin C, Cantin MA, Benarouch TE: Long-term functional outcome after selective posterior rhizotomy. J Neurosurg 2002, 97: 315-325. 10.3171/jns.2002.97.2.0315View ArticlePubMedGoogle Scholar
- Thompson AJ, Jarrett L, Lockley L, Marsden J, Stevenson VL: Clinical management of spasticity. J Neurol Neurosurg Psychiatry 2005, 76: 459-463. 10.1136/jnnp.2004.035972PubMed CentralView ArticlePubMedGoogle Scholar
- Ashworth B: Preliminary trial of carisprodol in multiple sclerosis. Practitioner 1964, 192: 540-542.PubMedGoogle Scholar
- Gracies J-M, Marosszeky J, Renton R, Sandaman J, Gandevia S, Burke D: Short-term effects of dynamic lycra splints on upper limb in hemiplegia patients. Arch Phys Med Rehabil 2000, 81: 1547-55. 10.1053/apmr.2000.16346View ArticlePubMedGoogle Scholar
- Lance JW: Symposium synopsis. In Spasticity: disordered motor control. Edited by: Feldman RG, Young RR, Koella WP. Chicago: Year Book Medical Publishers; 1980:485-495.Google Scholar
- de Vlugt E, de Groot JH, Schenkeveld KE, Arendzen JH, van der Helm FC, Meskers CG: The relation between neuromechanical parameters and ashworth score in stroke patients. J Neuroeng Rehabil 2010, 7: 35. 10.1186/1743-0003-7-35PubMed CentralView ArticlePubMedGoogle Scholar
- Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B: Development and reliability of a system to classify gross motor function in children with cerebral palsy. Dev Med Child Neurol 1997, 39: 214-223.View ArticlePubMedGoogle Scholar
- Gajdosik RL, Lentz DJ, McFarley DC, Meyer KM, Riggin TJ: Dynamic elastic and static viscoelastic stress-relaxation properties of the calf muscle-tendon unit of men and women. Isokinetics and Exercise Science 2006, 14: 33-44.Google Scholar
- van den Noort JC, Scholtes VA, Harlaar J: Evaluation of clinical spasticity assessment in cerebral palsy using inertial sensors. Gait Posture 2009, 30: 138-143. 10.1016/j.gaitpost.2009.05.011View ArticlePubMedGoogle Scholar
- Shortland AP, Harris CA, Gough M, Robinson RO: Architecture of the medial gastrocnemius in children with spastic diplegia. Dev Med Child Neurol 2002, 44: 158-163. 10.1017/S0012162201001864View ArticlePubMedGoogle Scholar
- McDowell BC, Salazar-Torres JJ, Kerr C, Cosgrove AP: Passive range of motion in a population-based sample of children with spastic cerebral palsy who walk. Phys Occup Ther Pediatr 2012, 32: 139-50. 10.3109/01942638.2011.644032View ArticlePubMedGoogle Scholar
- Hägglund G, Wagner P: Spasticity of the gastrosoleus muscle is related to the development of reduced passive dorsiflexion of the ankle in children with cerebral palsy: a registry analysis of 2,796 examinations in 355 children. Acta Orthop 2011, 82: 744-8. 10.3109/17453674.2011.618917PubMed CentralView ArticlePubMedGoogle Scholar
- Alhusaini AA, Crosbie J, Shepherd RB, Dean CM, Scheinberg A: Mechanical properties of the plantarflexor musculotendinous unit during passive dorsiflexion in children with cerebral palsy compared with typically developing children. Dev Med Child Neurol 2010, 52: e101-e106. 10.1111/j.1469-8749.2009.03600.xView ArticlePubMedGoogle Scholar
- Smith LR, Lee KS, Ward SR, Chambers HG, Lieber RL: Hamstring contractures in children with spastic cerebral palsy result from a stiffer extracellular matrix and increased in vivo sarcomere length. J Physiol 2011, 589: 2625-2639. 10.1113/jphysiol.2010.203364PubMed CentralView ArticlePubMedGoogle Scholar
- de Vlugt E, de Groot JH, Wisman WH, Meskers CG: Clonus is explained from increased reflex gain and enlarged tissue viscoelasticity. J Biomech 2012, 45: 148-55. 10.1016/j.jbiomech.2011.09.023View ArticlePubMedGoogle Scholar
- Tabary JC, Tardieu C, Tardieu G, Tabary C: Experimental rapid sarcomere loss with concomitant hypoextensibility. Muscle Nerve 1981, 4: 198-203. 10.1002/mus.880040305View ArticlePubMedGoogle Scholar
- Hof AL: Changes in muscles and tendons due to neural motor disorders: implications for therapeutic intervention. Neural Plast 2001, 8: 71-81. 10.1155/NP.2001.71PubMed CentralView ArticlePubMedGoogle Scholar
- Tardieu C, Huet de la Tour E, Bret MD, Tardieu G: Muscle hypoextensibility in children with cerebral palsy: I. Clinical and experimental observations. Arch Phys Med Rehabil 1982, 63: 97-102.PubMedGoogle Scholar
- Burne JA, Carleton VL, O'Dwyer NJ: The spasticity paradox: movement disorder or disorder of resting limbs? J Neurol Neurosurg Psychiatry 2005, 76: 47-54. 10.1136/jnnp.2003.034785PubMed CentralView ArticlePubMedGoogle Scholar
- Pierce SR, Prosser LA, Lauer RT: Relationship between age and spasticity in children with diplegic cerebral palsy. Arch Phys Med Rehabil 2010, 91: 448-451. 10.1016/j.apmr.2009.11.016View ArticlePubMedGoogle Scholar
- Hagglund G, Wagner P: Spasticity of the gastrosoleus muscle is related to the development of reduced passive dorsiflexion of the ankle in children with cerebral palsy. Acta Orthop 2011, 82: 744-748. 10.3109/17453674.2011.618917PubMed CentralView ArticlePubMedGoogle Scholar
- Desloovere K, Molenaers G, Feys H, Huenaerts C, Callewaert B, Van de Walle P: Do dynamic and static clinical measurements correlate with gait analysis parameters in children with cerebral palsy? Gait Posture 2006, 24: 302-313. 10.1016/j.gaitpost.2005.10.008View ArticlePubMedGoogle Scholar
- Kubo K, Kanehisa H, Fukunaga T: Is passive stiffness in human muscles related to the elasticity of tendon structures? Eur J Appl Physiol 2001, 85: 226-232. 10.1007/s004210100463View ArticlePubMedGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.