Archives of Physical Medicine and Rehabilitation
Volume 89, Issue 4 , Pages 700-710, April 2008

Separate Quantification of Reflex and Nonreflex Components of Spastic Hypertonia in Chronic Hemiparesis

Presented in part to the American Academy of Physical Medicine and Rehabilitation, October 21–23, 2002, Orlando, FL.

  • Sun G. Chung, MD, PhD

      Affiliations

    • Rehabilitation Institute of Chicago, Chicago, IL
    • Department of Physical Medicine & Rehabilitation, Northwestern University, Chicago, IL
    • Department of Rehabilitation Medicine, Seoul National University, Seoul, South Korea.
    • ,
  • Elton van Rey, PT

      Affiliations

    • Rehabilitation Institute of Chicago, Chicago, IL
    • Department of Physical Medicine & Rehabilitation, Northwestern University, Chicago, IL
    • ,
  • Zhiqiang Bai

      Affiliations

    • Rehabilitation Institute of Chicago, Chicago, IL
    • Department of Physical Medicine & Rehabilitation, Northwestern University, Chicago, IL
    • ,
  • W. Zev Rymer, MD, PhD

      Affiliations

    • Rehabilitation Institute of Chicago, Chicago, IL
    • Department of Physical Medicine & Rehabilitation, Northwestern University, Chicago, IL
    • Department of Biomedical Engineering, Northwestern University, Chicago, IL
    • Department of Physiology, Northwestern University, Chicago, IL
    • ,
  • Elliot J. Roth, MD

      Affiliations

    • Rehabilitation Institute of Chicago, Chicago, IL
    • Department of Physical Medicine & Rehabilitation, Northwestern University, Chicago, IL
    • ,
  • Li-Qun Zhang, PhD

      Affiliations

    • Rehabilitation Institute of Chicago, Chicago, IL
    • Department of Physical Medicine & Rehabilitation, Northwestern University, Chicago, IL
    • Department of Orthopedic Surgery, Northwestern University, Chicago, IL
    • Department of Biomedical Engineering, Northwestern University, Chicago, IL
    • Corresponding Author InformationReprint requests to Li-Qun Zhang, PhD, Sensory Motor Performance Program, Rehabilitation Institute of Chicago, 345 E Superior St, Ste 1406, Chicago, IL 60611

Article Outline

Abstract 

Chung SG, van Rey E, Bai Z, Rymer WZ, Roth EJ, Zhang L-Q. Separate quantification of reflex and nonreflex components of spastic hypertonia in chronic hemiparesis.

Objectives

To isolate and quantify reflex and nonreflex components of the spastic ankle plantarflexors in hemiplegia poststroke and to correlate them with clinical measures of spasticity, which may involve hyperactive stretch reflex and/or increased joint stiffness.

Design

To investigate reflex and nonreflex properties associated with spasticity in a case-control manner.

Setting

Research laboratory in a rehabilitation hospital.

Participants

Hemiplegic patients (n=17) and the same number of healthy subjects.

Interventions

Not applicable.

Main Outcome Measures

Tendon reflexes of spastic muscles were evaluated under an isometric condition, which essentially eliminated passive viscoelastic contributions associated with limb movement. Nonreflex components of spasticity were evaluated by moving the ankle joint slowly, which minimized reflex actions. The reflex and nonreflex measures were investigated and correlated with each other and with clinical measures.

Results

Compared with healthy subjects, patients with stroke showed a lower reflex threshold, higher electromyographic gains, and torque reflex gains, indicating hyperactive reflexes. For nonreflex properties, ankles of stroke patients showed higher stiffness, reduced range of motion (ROM), and larger resistant torque at comparable positions, reflecting peripheral soft-tissue changes at the ankle of the chronic stroke patients. Furthermore, the clinical reflex score correlated with all of the quantitative reflex measures but not with the nonreflex measures, whereas the dorsiflexion ROM showed a significant correlation with a nonreflex measure. The Modified Ashworth Scale was correlated with all of the reflex measures and 1 of the nonreflex measures.

Conclusions

Comprehensive and convenient evaluation of spasticity should be performed quantitatively with the separate measures of reflex and nonreflex components, especially in chronic conditions. With proper simplifications, the current method of separate quantification can potentially be used for convenient clinical evaluations of spasticity.

Key Words: Ankle, Contracture, Hemiplegia, Muscle spasticity, Rehabilitation

 

SPASTICITY IS A MOTOR disorder characterized by a velocity-dependent increase in tonic stretch reflexes (muscle tone) with exaggerated tendon jerks, resulting from hyperexcitability of the stretch reflex.1 A more recent definition focused on hypertonia in children and emphasized the velocity and direction dependence of the resistance increase and the dependence of the resistance on the threshold in speed and joint angle.2 Spasticity is a major source of disability in many patients suffering from brain or spinal cord injury (SCI) of many different etiologies, including traumatic SCI, multiple sclerosis, and stroke.3 Many therapeutic paradigms, such as antispastic medication,4 physical modalities,5 botulinum toxin injection,6 intrathecal baclofen pumps,7 and novel surgical techniques,8 were developed and applied for the purpose of reducing spasticity and improving function. Although such techniques can benefit patients, their effect on spasticity has not been well quantified because there are scarce accurate measures that can be used in clinical setting conveniently.9, 10, 11, 12

One of the reasons of this paucity of quantification methods suitable for clinical practice could be attributed to the fact that many precise and sophisticated quantification techniques based on biomechanic or neurophysiologic principles did not correlate well with the widely used clinical scales.13, 14 These discrepancies are considered to originate from the unclear clinical definition of spasticity15 and/or the quantifications done only for some specific aspects of spasticity.16 It is commonly agreed that spastic muscle hypertonus is due primarily to an increase in stretch reflex activity. However, some investigators17 argued that increased stiffness in a spastic limb could not be directly interpreted as the evidence of a hyperactive reflex because of the fact that the increased resistance to a stretch could be attributed to passive tissue stiffness (from connective tissue, tendons, ligaments, and passive muscle properties), intrinsic stiffness of contracting muscle fibers, and reflex-mediated stiffness. It suggests that spasticity is a complex phenomenon consisting of the reflex- and nonreflex-mediated stiffness or resistance to a passive limb movement, each of which needs to be quantified uniquely and related with its corresponding clinical facet to explain and understand the multifaceted clinical features of spasticity and to evaluate its mechanisms reliably.18

Several studies have been performed to separate the reflex- and nonreflex-mediated contributions to muscle force or joint torque by comparing mechanical contributions of muscles before and after altering reflex pathways, including transection of dorsal roots,19 deep anesthesia to suppress reflex pathway,20 or by stimulating the relevant muscle nerve to override reflex effects.21 Only the last approach of electrically stimulating an appropriate nerve has been used to study human subjects,17 which still has some limitations because exactly matched joint position and muscle force between the reflex-active and reflex-suppressed states need to be achieved for reliable conclusions. A few in vivo human experiments22, 23, 24, 25, 26, 27 have been performed to separate and quantify both the reflex and nonreflex components of spasticity. In separate studies, we evaluated the reflex components11, 28, 29 through system identification of tendon reflexes and the nonreflex components30 through controlled movement.

The purpose of this study was to separately quantify each of the reflex and nonreflex components of hemiplegic spasticity by using convenient methods potentially applicable in clinics (with further simplifications), including an instrumented Achilles’ tendon reflex test and motorized passive movement test to compare the quantitative parameters between the stroke patients and healthy populations and also correlate the quantified variables and clinical measures with each other.

Back to Article Outline

Methods 

Participant Selection 

Seventeen stroke patients (11 men, 6 women; mean ± standard deviation [SD], 57.2±10.4y) and the same number of healthy subjects (11 men, 6 women; age, 52.3±19.4y) participated in the study. All of the patients had supratentorial, cortical, or subcortical stroke at least a year before the experiments, which resulted in chronic spastic hemiplegia. Each patient was examined initially for spasticity by using the clinical deep tendon reflex (DTR) scale (range, 0–4) and Modified Ashworth Scale (MAS) from 0 to 4 assuming the 1+ grade as 1.5.31 Dorsiflexion range of motion (ROM) was measured with a manual goniometer, placing the fulcrum of the goniometer at the center of the lateral malleolus and aligning the proximal and distal arms with the lateral midline of the fibular and the long axis of the sole, respectively.32 Zero degrees of dorsiflexion ROM was defined as the ankle position where the foot and leg were perpendicular to each other. The ankle joint angle dorsiflexed from the 0° of dorsiflexion ROM was denoted as positive, whereas the plantarflexed ankle angle was expressed in a negative value.

All of the examinations were performed while the patient was in the sitting position on a chair with the knee flexed at 60° from full extension, which was the same position used in the quantification experiments. About 3 trials were conducted for each of the clinical measures before the examiner wrote down the values. If the patients had a zero MAS score and a DTR score not higher than the average normative response (≤2), they were considered not spastic and would not be included in the study. Other physical impairments that could be related to the measurements such as ambulatory status were recorded. The use of antihypertonia medications was also documented. Seventeen healthy subjects with no history of neurologic disorders or musculoskeletal injuries were used as controls. Subjects who had previous ankle injury, surgery, or any kind of neurolytic injections on the ankle muscles were excluded from either the patient or control group. All subjects gave informed consent. The study was approved by the Institutional Review Board of Northwestern University.

Experimental Setup 

The evaluation was performed by using a custom-designed joint driving device (fig 1). The joint driving device moved the ankle at a well-controlled slow speed, and it slowed down as resistance torque increased. In this way, the joint muscles were moved under a controlled load, and reflex-mediated responses were minimized.30, 33 The subject was seated upright with the foot held firmly onto a footplate by a premolded plastic cast and some clamps. The footplate was mounted onto a motor shaft through a 6-axis force sensora that measured torques at the ankle joint. The seat and leg support were adjusted to align the ankle flexion axis with the axis of the torque sensor, and the knee and hip joints were positioned at 60° and 85° flexion from straight, respectively. The subject maintained his/her head and trunk in midline throughout the whole experimental procedure. To place the ankle at an initial position for further experimental procedures, the footplate was fixed at the neutral position or at a position as close to neutral as possible without stretching the potentially stiff ankle joints. The neutral ankle joint position was determined by positioning the sole of the foot at 90° with respect to the long axis of the lower leg. While the patient was being asked to relax, the initial offset torque of the ankle joint was measured at the initial position. After the initial offset torque was recorded, the torque sensor was adjusted to have 0 values at the initial position with the patient relaxed. The initial offset torque was combined to the resistant torque during postprocessing. Surface electrodesb were placed on the tibialis anterior, soleus, and the medial and lateral heads of the gastrocnemius muscles to monitor electromyographic activities during passive movement by the joint driving device and to record the reflex-mediated electromyographic response from the soleus muscle. To minimize the effect of different levels of skin impedance among subjects, the skin was cleaned by mild abrasion with a scrubbing pad. The positions of the electrodes were standardized by placing the electrodes on the medial and lateral mass of the calf at 1 handbreadth below the popliteal crease for medial and lateral heads of the gastrocnemius, respectively. For the soleus, the electrode was placed just distal to the belly of the gastrocnemius muscle, medial and anterior to the Achilles’ tendon.34 The tendon tapping force, electromyography signals, and ankle joint torque were sampled by a computer at 500Hz after low-pass filtering at 230Hz with eighth-order Butterworth filters. The motor was controlled to lock the footplate at the initial position for the instrumented Achilles’ tendon reflex test to quantify the reflex properties of spastic hypertonia and to move the ankle joint passively for the quantification of the nonreflex properties.

  • View full-size image.
  • Fig 1. 

    Experimental setup for instrumented Achilles’ tendon reflex test and motorized passive movement test. (A) The leg was strapped to the leg support at 60° of knee flexion. The thigh and trunk were strapped to the seat and backrest, respectively, with the hip at 85° of flexion. The foot and cast were clamped and strapped to the attachment with appropriate alignment. The footplate was mounted onto a torque sensor measuring the ankle flexion torque, which was connected to a servo motor. (B) The instrumented tendon hammer with a force sensor mounted at its head and the rubber pad are shown. Abbreviations: DSP, digital signal processor, LED, light-emitting diode, PC, personal computer.

Instrumented Achilles’ Tendon Reflex Test 

To avoid the changes of reflex properties potentially induced by stretching during motorized passive movement for quantification of the nonreflex properties, we always performed the instrumented Achilles’ tendon reflex test before the motorized passive movement procedure to quantify the reflex components of spasticity. The ankle joint was positioned at the initial position and restricted to an isometric condition to minimize contributions of the passive and intrinsic components associated with joint movement. A hemisphere-shaped rubber pad (diameter, 1cm) with adhesive interface was mounted onto the Achilles’ tendon at a spot where the strongest reflex response was evoked. An instrumented tendon hammer with a force sensorc mounted at its head was used to tap the rubber pad (see fig 1B). The pad arrangement transmitted the tapping force evenly and accurately to the Achilles’ tendon and reduced the coefficient of variation of the tendon reflex parameters.28 Before and during the tapping, the participants were asked to relax and not to expect and/or anticipate the tapping, which was monitored by real-time electromyographic and torque signals. The tapping force was gradually increased until a triceps surae muscle contraction was evoked. The Achilles’ tendon was then tapped at approximately that level for about 7 times during a trial, with a random interstimulus interval of about 3 seconds. Three trial sequences were collected at the neutral ankle joint position.

Motorized Passive Movement Test 

The ankle joint was moved passively in both dorsiflexion and plantarflexion directions starting from the initial position by the joint driving device. While the ankle joint was being moved, the angular position of the ankle joint and the torque generated by the ankle muscles were continuously recorded, and, at the same time, the data were input to the computer to control the motor. For safety purposes, torque limits were set for the dorsiflexion direction of passive movement at 10Nm after the torque at the initial position was adjusted to 0, and position limits were also set to prevent the device from driving the footplate to positions too extreme in both the dorsiflexion and plantarflexion directions. The joint driving device moved the ankle passively and repeatedly in both directions in trials of 90 seconds long. The passive movement was controlled in such a way that the movement velocity decreased inversely proportional to the resistance torque, which helped avoid inducing reflex activation of the muscles involved. Subjects were asked to relax as much as they could, and electromyographic signals and torque were monitored in real time for obvious muscle activations during the passive movement. Data with voluntary or reflexic muscular contraction at any ranges of ankle position showing obvious electromyographic signal were excluded from further analysis.

Data Analysis 

The sampled tendon tapping force, surface electromyographic signals, and ankle plantarflexion torque signals were low-pass filtered. Surface electromyographic signals were full-wave rectified and low-pass filtered with 20-Hz cutoff frequency to extract the electromyographic linear envelope. The force and torque signals measured from the 6-axis force sensors were transformed into anatomic joint torques with the passive resistance torque generated by ankle plantarflexors as positive.

Quantification of reflex properties by instrumented Achilles’ tendon reflex test 

To quantify reflex properties of triceps surae muscles, Achilles’ tendon reflex was viewed as a system that related the system input (the tapping force; figs 2A, 2B) to system outputs (the reflex-mediated electromyographic response and torque, figs 2C−F). From the relationship between the input and outputs, the tendon reflex was characterized by system gains and the reflex threshold in tapping force. As shown in figures 2A and B, the threshold in tendon tapping force (fth) for evoking reflex responses was determined by the averaged peak tapping force because the tendon was tapped repeatedly in the experiment just above the threshold.28 Two gains were estimated to quantify the Achilles’ tendon reflex system by using the impulse responses (figs 2G−J) of reflex-mediated electromyographic response, hEMGf(t), and torque, hMf(t). Because the tapping force was rather brief, it could be approximated as a pulse. Therefore, the impulse response hEMGf(t) and hMf(t) could be easily approximated as the reflex-mediated electromyographic response—EMG(t)—and torque response—M(t)—scaled by the area of the tapping force pulse—f(τ)11, 28, 29:

(1)
(2)

  • View full-size image.
  • Fig 2. 

    (A, B) The tendon tapping force, (C, D) reflex-mediated ankle plantarflexion electromyographic (EMG) response, (E, F) reflexive ankle plantarflexion torque (Torq), (G, H) impulse responses (IR) of the reflex-mediated electromyographic response, and (I, J) torque. The 2 columns correspond to representative results from a spastic stroke patient (left column) and healthy subject (right column). The peak values of tapping force and 2 impulse responses correspond to the threshold in tapping force (fth), electromyographic reflex gain (Ger), and torque reflex gain (Gtr). The solid and dashed lines give the mean and mean ± SD, respectively.

The Gtr was estimated from the peak value of the impulse response, which represents the system gain from the tendon tapping force to the reflex-mediated torque (see figs 2I, J), and the peak value of the impulse response of reflex-mediated electromyographic response was measured as the electromyographic reflex gain (Ger) (see figs 2G, H).

Quantification of nonreflex properties by motorized passive movement test 

The initial torque offset recorded at the beginning of the experiment was added to the measured joint torque in the postprocessing analysis because the torque sensor was calibrated to 0 after the recording of the initial offset. The gravitational force of the foot and the footplate was calculated and compensated at each degree of ankle position within the ROM. The weight and center of mass of the foot were calculated from the anthropometric data including the body weight, foot length, the width, and height of the malleolus measured from the subject.35 The anatomic joint torque and angle were plotted to get torque-ankle curves (hysteresis loops, figs 3A−D). The number of hysteresis loops during each passive movement trial ranged from 4 to 6, depending on the ROM of the subject (see figs 3A, 3B). Multiple loops were reduced into a single representative hysteresis loop (averaged torque-angle curve; see figs 3C, 3D) for each subject as described in the previous article.30 The nonreflex properties of the triceps surae muscles were characterized with the dorsiflexion ankle angle at 10Nm of passive resistant torque (A10Nm) (see fig 3C), the passive resistant torque (T10df), and the stiffness (K10df) at 10° of dorsiflexion angle (see fig 3D). The stiffness of ankle plantarflexors was assessed as K=ΔT/Δθ, where K is the quasistatic stiffness (spring-like property characterized by the elastic stiffness of the spring) and ΔT is an increment of the passive torque during a certain amount of ankle angular movement (Δθ). As Δθ becomes infinitely small, the quasistatic stiffness approaches the slope of a tangential line of the torque-angle curve at a specific ankle position.36 The stiffness was calculated at the point of 10° ankle angle in dorsiflexion direction by taking the slope of the regression curve to fit 7 data points of 7° to 13° ankle angle (see fig 3D).

  • View full-size image.
  • Fig 3. 

    Representative signals during passive movement trials on a stroke patient (left column) with a spastic ankle and a control subject (right column). (A, B) The top and the (C, D) second rows correspond to the ankle joint angle (positive for dorsiflexion) and ankle joint torque (positive for plantarflexor resistance torque), respectively. (E, F) The electromyography signals from the soleus (sol) muscles are shown. (G, H) Plotting the ankle joint torque (ordinate) at each corresponding ankle joint angle (abscissa) generates multiple hysteresis curves, moving clockwise as time progresses as indicated by the arrows. (I, J) The multiple curves are averaged for further measurements of the nonreflex properties. For simplicity, (I) the dorsiflexion angle at 10Nm of passive resistant torque (A10Nm) is shown while (J) the passive resistant torque (T10df) and the stiffness (K10df) at 10° of dorsiflexion angle are shown. The slope of the thick oblique segment in panel J gives the stiffness (K10df).

Statistical Analysis 

To investigate whether the reflex and nonreflex properties of ankle plantarflexors were changed in hemiplegic spasticity, 3 parameters of each component (fth, Ger, and Gtr for the reflex component; A10Nm, T10df, and K10df for the nonreflex component) were compared between the spastic and control groups by using multivariate analysis of variances (MANOVA) and subsequent univariate analysis of variance (ANOVA) procedures.d Because the multivariate normality and homogeneity of covariance were not strictly assumed because of the relatively small sample sizes, the Pillai trace, which is known to be the most robust MANOVA statistic for violation of the assumptions, was chosen to accept the statistical significance at a P value of less than .05.37

The relationships among the clinical, reflex, and nonreflex parameters were analyzed by calculating correlation coefficients between each pair of the variables over all subjects. Because the MAS and DTR were ordinal variables, Spearman ρ coefficients were calculated between all the measures.38 However, because most of the variables in this study were continuous and Spearman ρ correlation coefficients only measured monotonicity instead of a linear relationship, Pearson product-moment correlation coefficients were also calculated between each pair of the variables.38 The critical values to determine the correlations for the Spearman ρ and Pearson product-moment coefficients were set for P equal to .01.

Back to Article Outline

Results 

General Features of the Normal and Spastic Subjects 

Fifteen of the 17 stroke patients showed clinically enhanced Achilles’ tendon reflexes (DTR>2), and all of the stroke subjects had increased triceps muscle tone to passive manual dorsiflexion (MAS score >0). The 2 stroke subjects who had normal Achilles’ tendon reflexes (DTR=2) showed increased triceps surae muscle tone (MAS scores, 3 and 1). The average MAS and DTR scores of the stroke patients were 2.3 and 3.2, respectively. The mean duration of stroke ± SD was 6.7±2.9 years (range, 1−14y). The height, weight, and age of the 2 populations were not significantly different (table 1).

Table 1. General Features of the Stroke and Control Subjects
CharacteristicsStrokeControl
Age (y)57.2±10.452.3±19.4
Sex (male/female)6/116/11
Hemiplegic side (right/left)7/10NA
Height (cm)169.7±11.2172.1±9.5
Body weight (kg)80.1±14.674.5±11.4
MAS score⁎2.3±0.90
Tendon reflex scale score (0–4)3.2±0.71.9±0.3
Dorsiflexion ROM (deg)⁎−0.4±8.113.2±7.7
Years since stroke onset (y)6.7±2.9NA
Use of AFO (yes/no)13/4NA
Medication (yes/no)†2/15NA
Ambulatory (yes/no)‡14/3All

NOTE: Values are in mean ± SD.

Abbreviations: AFO, ankle foot orthosis; NA, not applicable.

⁎Significant differences between the groups at P<.01 by an independent t test (dorsiflexion ROM) or Mann-Whitney U tests (MAS scores and reflex).

†Whether the subject was taking antihypertonia medication (baclofen) at the time.

‡Yes includes household ambulatory level or higher.

Comparison of the Reflex Measures 

As shown in table 2, 1-way MANOVA tested on the reflex properties (fth, Ger, Gtr) revealed a significant multivariate effect for the stroke and control groups (Pillai trace, .284; F=3.31, P=.036). Subsequent univariate ANOVAs showed that the mean tapping force threshold was significantly lower in the spastic limbs than that in normal controls (P<.005). As system parameters characterizing the input and output variables simultaneously, Ger and Gtr were significantly higher in the spastic hemiplegic patients than in normal controls (P=.012, P=.026, respectively) (see table 2). Typical raw data representing the comparison of the reflex components between the 2 groups were shown in figure 2.

Table 2. Comparisons of the Reflex and Nonreflex Properties Between 2 Groups
ComponentParameterStrokeControlMultivariate TestUnivariate Test (P)
Reflexfth (N)13.63±4.2519.08±5.39Pillai trace, .284;.005
Ger (uV/N)5.13±4.791.56±1.39F=3.310,.012
Gtr (cm)6.72±5.433.10±1.95P=.036.026
NonreflexA10Nm (deg)11.40±6.6920.53±8.06Pillai trace, .309;.001
T10df (Nm)10.15±4.585.70±3.09F=4.466,.002
K10df (Nm/deg)0.54±0.300.31±0.18P=.010.011

NOTE: Values are in mean ± SD.

Abbreviations: A10Nm, ankle joint angle at 10Nm passive resistant torque; fth, tapping force threshold; Ger, electromyographic reflex gain; Gtr, torque reflex gain; K10df, stiffness at 10° of ankle joint angle; T10df, passive resistant torque at 10° of ankle joint angle.

Comparison of the Quantified Nonreflex Components 

The multivariate effect for the parameters of the nonreflex component was tested to give a Pillai trace of .309 (F=4.446, P=.01), indicating a significant difference between the 2 groups in those parameters (see table 2). A10Nm of the stroke group (11.40°±6.69°) was far smaller than that of the control group (20.53°±8.06°) (P=.001), and T10df was larger in the spastic patients than the controls (P=.002). Consequently, the stiffness K10df was higher in the stroke group (.54±.30Nm/deg) than in the controls (.31±.18Nm/deg) (P=.011). Figure 3 shows the typical data of the nonreflex components from both groups.

Correlations Among the Reflex, Nonreflex, and Clinical Variables 

As shown in the left lower part (Spearman coefficients) of table 3, significant relations with the DTR were found in all the reflex measures (fth, Ger, Gtr) (fig 4A), but none of the nonreflex measures showed significant correlations. The manually measured dorsiflexion ROM had a meaningful correlation with one of the nonreflex parameters (A10Nm, fig 4D). It is of note that all of the reflex and one (A10Nm) of the nonreflex measures correlated well with the MAS score as shown in table 3 and figures 4E and F. No correlation was found between the reflex and the nonreflex properties (see table 3).

Table 3. Correlation Coefficients Among the Quantitative Parameters and Clinical Measures
Parameters and MeasuresMASDTRDorsiflexion ROMfthGerGtrA10NmT10dfK10df
MAS1.0000.836−0.587−0.4760.6200.514−0.3960.3640.415
DTR0.8521.000−0.566−0.6660.6600.557−0.3090.1900.063
Dorsiflexion ROM−0.691−0.6181.0000.306−0.432−0.3260.453−0.446−0.332
fth−0.487−0.6420.3691.000−0.648−0.6440.105−0.0390.100
Ger0.5710.683−0.450−0.7681.0000.802−0.1690.1070.086
Gtr0.5290.558−0.293−0.7440.8231.000−0.2860.2430.193
A10Nm−0.461−0.3540.5480.054−0.200−0.1881.000−0.917−0.582
T10df0.3860.238−0.423−0.0400.1230.179−0.9591.0000.642
K10df0.4260.186−0.4260.0430.0310.184−0.6130.5761.000

NOTE. The Spearman ρ (lower-left half) and Pearson product-moment correlation coefficients (upper-right half), calculated with data from all 34 subjects. A Pearson coefficient greater than .449 and Spearman ρ greater than .455 indicate significant correlation with P<.01. The boldface indicates the significant correlations between the variables.

  • View full-size image.
  • Fig 4. 

    Scattergraphs of the quantitative data as a function of the clinical measures in 17 normal (open circles) and 17 stroke (filled squares) subjects. (A, B) The peak reflex gain (Gtr) and ankle joint angle at 10Nm of passive resistant torque are plotted versus the tendon reflex scale (DTR) versus (C, D) the manual dorsiflexion ROM, and (E, F) the MAS. The ρ indicates the Spearman correlation coefficient, and r indicates the Pearson correlation coefficient. *The correlation coefficients higher than the critical values (ρ>.449 or r>.455 with P<.01).

Back to Article Outline

Discussion 

The reflex and nonreflex properties of the ankle plantarflexors accompanied with spasticity were quantified separately and compared with those of the control group, showing decreased reflex threshold, increased electromyographic and torque reflex gains, and higher stiffness accompanied with smaller ROM and larger passive resistant torque in the spastic triceps surae than in the healthy control muscles. The separately quantified parameters had meaningful relationships with their corresponding clinical measures. MAS had good relations with both the reflex and nonreflex components.

Quantification of the Reflex Components 

As in the instrumented Achilles’ tendon reflex test, use of the isometric condition largely eliminated the mechanical contributions of joint stiffness, viscosity, and limb inertia because there was essentially no limb motion. Therefore, the reflex contribution was readily separated from intrinsic and passive contributions to the joint torque, which was otherwise difficult to separate when limbs were allowed to move.

The quantification of the reflex components was obviously beneficial because Gtr provided a level of precision not readily matched by the 5-point (0 to 4) tendon reflex scale widely used in clinical practice.39, 40 One stroke patient who participated in the study showed very strong Achilles’ tendon reflexes, and his Gtr was about 7 times greater than the average Gtr of the normal subjects. This shows the insufficiency of the 0 to 4 tendon reflex scale and the potential clinical utility of the more sensitive quantitative measures, such as the fth, Ger, and Gtr.

From a system point of view, the system gain is defined by the area of the impulse response. In this study, we used the peak of the impulse response as the tendon reflex gain. In fact, we calculated the system gains of tendon reflexes by using both the peak and the area of the impulse response, and the results on the 2 gain measures were similar to each other, considering the shape of the impulse response is similar among the subjects.11 Practically, the area calculation might involve larger relative error in subjects with very low reflex responses, and it was affected by the tapping impact-induced artifact at the beginning of the impulse response. In contrast, the peak of the impulse response was more easily distinguished from the background activity and was separated clearly in time from the impact-induced artifact and, thus, provided a more robust and simpler measure.

Quantification of the Nonreflex Components 

Although both reflex and nonreflex changes in ankles with spasticity can substantially affect functional performance of stroke patients, several studies41, 42, 43 suggested that nonreflex changes had more profound and consistent effects than reflex changes. Moreover, some argued that changes in ankle passive biomechanic properties could contribute to the internal ankle joint torque in functional movement depending on the severity of spasticity.43, 44, 45 In response to the need for more precise evaluation and comprehensive understanding of the passive biomechanic changes in hemiplegic ankles, we were able to quantify the passive properties of the spastic plantarflexors isolated from the reflex properties by moving ankle joints slowly under precise control, minimizing velocity-dependent, phasic reflex activation without provoking considerable reflex-mediated electromyographic responses as in our previous study.30 The phasic reflex activity might not have been eliminated completely but rather minimized even though we asked the subjects to relax as much as they could and, if necessary, waited until subjects settled down. In a recent study, Burne et al23 showed that patients poststroke may have slightly elevated electromyographic activities, which may be part of the pathologic changes after stroke and may have contributed to the increased tendon reflex excitability.

The nonreflex parameters, such as the passive ROM, measured in this study may appear larger than those in clinical settings partly because the ankles were moved by a motor until a controlled peak resistance torque was reached. When the motor moved the ankles of the stroke patients, the ankle resistance torque was about twice as high as that in the healthy ankles for the same range of dorsiflexion (see table 2). Our results were comparable with previous studies46, 47 that used motors and measured dorsiflexion ROM at 10Nm of resistant torque. Harlaar et al46 reported 11.3° of the averaged ankle dorsiflexion ROM of 8 hemiplegic limbs with extended knee (≈0°) position and 20.3° with flexed knee (≈90°) position. In 13 patients with brain injury, Singer et al47 reported 10.2° of ankle dorsiflexion ROM at 10Nm resistance torque with the knee at 90° of flexion.

Correlations Among the Quantitative Parameters and Clinical Measures 

As expected, the DTR and dorsiflexion ROM showed significant correlations with each corresponding quantitative variables (fth, Ger, and Gtr for DTR; A10Nm for dorsiflexion ROM). However, weaker relationships were found between the quantitative parameters of 1 component and the clinical scales considered to represent the other component of spasticity. The fth, Gtr, and Ger, as parameters for the reflex properties, had weak Pearson correlation coefficients (r=−.306, r=−.432, r=−.326, respectively) with the manual dorsiflexion ROM, which might characterize the nonreflex properties. On the other hand, none of the nonreflexic parameters (A10Nm, T10df, K10df) showed any significant correlations with the clinical parameter of the reflex properties, the DTR scores. Because there was a lack of hemiplegic subjects with no spasticity, the range of scores for calculating correlations was narrow and did not include the full range of clinical scores. However, the scatter grams in figure 4 show that the data from the patient and control groups are not dichotomous but continuous, showing intermediate zones that include both stroke and normal subjects together. An interesting finding was that the MAS correlated well with the nonreflex component and the reflex component, suggesting that the scoring of MAS may be dependent on both of the reflex and nonreflex components.

There have been conflicting reports regarding the relationships of the MAS and quantified parameters of spastic hypertonia. Lin and Sabbahi48 observed a strong and consistent correlation between spasticity and reflexive electromyographic activities of the stretch reflex, whereas Pandyan et al13 reported that the MAS may not provide a valid measure of spasticity but a measure of resistance to passive movement in an acute stroke population. Although stretch reflex is highly dependent on the stretch velocity and different velocities give different results on the reflex contributions, the velocity used in obtaining the Ashworth Scale may vary substantially from 1 case to another.49 In this study, tendon reflexes conducted at isometric conditions essentially eliminated nonreflex contributions associated with limb movement and manifested the reflex contributions, which made the reflex evaluations more consistent. As shown in table 3, the parameters of the instrumented tendon reflex test had significant correlations with the MAS, but they gave more comprehensive and quantitative measures, including both reflex gain and reflex threshold, of hyperactive reflexes in spastic hypertonia. The results suggested that the MAS measured not just 1 component of spastic hypertonia, which is a multifaceted phenomenon, but would be influenced by multiple factors in addition to simple reflex measurements.9, 50 This finding agrees with a recent report,51 which suggested the Tardieu Scale, but not MAS, as a clinical measure to differentiate contracture from spasticity.

As shown in table 3, there were no significant correlations between the reflex and nonreflex components. This finding implicates that the quantitative changes of the reflex and nonreflex components of chronic spastic conditions may not always be commensurable to each other. It should be taken into consideration that the patient population in this study was made up of patients with a chronic condition who had relatively longer periods after the onset of stroke (6.7±2.9y), which might have affected the properties of spasticity in many ways, such as different severity of initial spasticity, various treatment strategies of spasticity, and/or different levels of ambulation. In chronic stroke patients, it would be necessary to quantify the 2 components of spasticity separately for a comprehensive measurement of spastic hypertonia. Isolated quantification of spasticity, distinguishing between reflex and nonreflex components, would accelerate to target treatment interventions appropriately.52

Study Limitations 

The reflex properties measured by this method are not completely out of the influence of the nonreflex properties because the force or electromyographic activity in response to a tendon tap can be affected by the changes in muscles. For example, the mechanical changes, such as shortening, in the musculotendinous complex could alter the reflex responses by changing the resting length or initial stiffness of the musculotendinous complex slightly.

Technically, the current method of separate quantification is still demanding and complicated. Further simplification is needed for convenient uses in a clinical setting.

Back to Article Outline

Conclusions 

By using the instrumented Achilles’ tendon reflex test and motorized passive movement of the ankle, we were able to isolate and quantify the reflex and nonreflex components of the spastic ankles, showing that the spastic plantarflexors had higher reflex responses and stiffer passive mechanical properties. It was found that the properties of the 2 components showed significant relationships with their corresponding clinical measures, but the quantified parameters of the reflex and nonreflex components were not exactly correlated with each other in our subject population. These findings suggest that the 2 components of spasticity, the reflex and nonreflex, should be quantified separately to evaluate spastic hypertonia comprehensively, especially in chronic spastic patients. The measurements used in the study could be further simplified and applied to clinical settings to quantify spasticity, distinguishing between reflex and nonreflex components.

Suppliers

Back to Article Outline

Acknowledgment 

We thank the Medical Research Collaborating Center of Seoul National University Hospital for statistical support.

Back to Article Outline

References 

  1. Lance JW. Symposium synopsis. In:  Feldman RG,  Young RR,  Koella WP editor. Spasticity: disordered motor control. Chicago: Year Book Medical; 1980;p. 485–494
  2. Sanger TD, Delgado MR, Gaebler-Spira D, Hallett M, Mink JW. Task Force on Childhood Motor Disorders (Classification and definition of disorders causing hypertonia in childhood). Pediatrics. 2003;111:e89–e97
  3. Sommerfeld DK, Eek EU, Svensson AK, Holmqvist LW, von Arbin MH. Spasticity after stroke: its occurrence and association with motor impairments and activity limitations. Stroke. 2004;35:134–139
  4. Gelber DA, Good DC, Dromerick A, Sergay S, Richardson M. Open-label dose-titration safety and efficacy study of tizanidine hydrochloride in the treatment of spasticity associated with chronic stroke. Stroke. 2001;32:1841–1846
  5. Harlaar J, Ten Kate JJ, Prevo AJ, Vogelaar TW, Lankhorst GJ. The effect of cooling on muscle co-ordination in spasticity: assessment with the repetitive movement test. Disabil Rehabil. 2001;23:453–461
  6. Hurvitz EA, Conti GE, Brown SH. Changes in movement characteristics of the spastic upper extremity after botulinum toxin injection. Arch Phys Med Rehabil. 2003;84:444–454
  7. Meythaler JM, Guin-Renfroe S, Brunner RC, Hadley MN. Intrathecal baclofen for spastic hypertonia from stroke. Stroke. 2001;32:2099–2109
  8. Keenan MA, Fuller DA, Whyte J, Mayer N, Esquenazi A, Fidler-Sheppard R. The influence of dynamic polyelectromyography in formulating a surgical plan in treatment of spastic elbow flexion deformity. Arch Phys Med Rehabil. 2003;84:291–296
  9. Pandyan AD, Johnson GR, Price CI, Curless RH, Barnes MP, Rodgers H. A review of the properties and limitations of the Ashworth and modified Ashworth Scales as measures of spasticity. Clin Rehabil. 1999;13:373–383
  10. Patrick SK, Denington AA, Gauthier MJ, Gillard DM, Prochazka A. Quantification of the UPDRS Rigidity Scale. IEEE Trans Neural Syst Rehabil Eng. 2001;9:31–41
  11. Zhang LQ, Wang G, Nishida T, Xu D, Sliwa JA, Rymer WZ. Hyperactive tendon reflexes in spastic multiple sclerosis: measures and mechanisms of action. Arch Phys Med Rehabil. 2000;81:901–909
  12. Peng Q, Shah P, Selles RW, Gaebler-Spira DJ, Zhang LQ. Measurement of ankle spasticity in children with cerebral palsy using a manual spasticity evaluator. Conf Proc IEEE Eng Med Biol Soc. 2004;7:4896–4899
  13. Pandyan AD, Price CI, Rodgers H, Barnes MP, Johnson GR. Biomechanical examination of a commonly used measure of spasticity. Clin Biomech. 2001;16:859–865(Bristol, Avon)
  14. Milanov I. Clinical and neurophysiological correlations of spasticity. Funct Neurol. 1999;14:193–201
  15. Ivanhoe CB, Reistetter TA. Spasticity: the misunderstood part of the upper motor neuron syndrome. Am J Phys Med Rehabil. 2004;83(10 Suppl):S3–S9
  16. Burridge JH, Wood DE, Hermens HJ, et al. Theoretical and methodological considerations in the measurement of spasticity. Disabil Rehabil. 2005;27:69–80
  17. Sinkjaer T, Toft E, Larsen K, Andreassen S, Hansen HJ. Non-reflex and reflex mediated ankle joint stiffness in multiple sclerosis patients with spasticity. Muscle Nerve. 1993;16:69–76
  18. Nuyens GE, De Weerdt WJ, Spaepen AJ, Kiekens C, Feys HM. Reduction of spastic hypertonia during repeated passive knee movements in stroke patients. Arch Phys Med Rehabil. 2002;83:930–935
  19. Nichols TR, Houk JC. Improvement in linearity and regulation of stiffness that results from actions of stretch reflex. J Neurophysiol. 1976;39:119–142
  20. Hoffer JA, Andreassen S. Regulation of soleus muscle stiffness in premammillary cats: intrinsic and reflex components. J Neurophysiol. 1981;45:267–285
  21. Houk JC, Singer JJ, Goldman MR. An evaluation of length and force feedback to soleus muscles of decerebrate cats. J Neurophysiol. 1970;33:784–811
  22. Lee HM, Huang YZ, Chen JJ, Hwang IS. Quantitative analysis of the velocity related pathophysiology of spasticity and rigidity in the elbow flexors. J Neurol Neurosurg Psychiatry. 2002;72:621–629
  23. 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
  24. Mirbagheri MM, Barbeau H, Ladouceur M, Kearney RE. Intrinsic and reflex stiffness in normal and spastic, spinal cord injured subjects. Exp Brain Res. 2001;141:446–459
  25. Galiana L, Fung J, Kearney R. Identification of intrinsic and reflex ankle stiffness components in stroke patients. Exp Brain Res. 2005;165:422–434
  26. Zhang LQ, Chung SG, van Rey EM, et al. Evaluation and treatment of spastic hypertonia and contracture. In: 24th Annual Conference and the Annual Fall Meeting of the Biomedical Engineering Society: Proceedings of the Second Joint EMBS/BMES Conference. 2002;p. 2451–2452Oct 23-26; Houston (TX)
  27. Rydahl SJ, Brouwer BJ. Ankle stiffness and tissue compliance in stroke survivors: a validation of myotonometer measurements. Arch Phys Med Rehabil. 2004;85:1631–1637
  28. Zhang LQ, Huang H, Sliwa JA, Rymer WZ. System identification of tendon reflex dynamics. IEEE Trans Rehabil Eng. 1999;7:193–203
  29. Chung SG, Van Rey EM, Bai Z, Rogers MW, Roth EJ, Zhang LQ. Aging-related neuromuscular changes characterized by tendon reflex system properties. Arch Phys Med Rehabil. 2005;86:318–327
  30. Chung SG, Van Rey E, Bai Z, Roth EJ, Zhang LQ. Biomechanic changes in passive properties of hemiplegic ankles with spastic hypertonia. Arch Phys Med Rehabil. 2004;85:1638–1646
  31. Bohannon RW, Smith MB. Interrater reliability of a modified Ashworth scale of muscle spasticity. Phys Ther. 1987;67:206–207
  32. Norkin CC, White DJ. Measurement of joint motion: a guide to goniometry. Philadelphia: FA Davis; 1985;
  33. Zhang LQ, Chung SG, Bai Z, et al. Intelligent stretching of ankle joints with contracture/spasticity. IEEE Trans Neural Syst Rehabil Eng. 2002;10:149–157
  34. Perotto A. Anatomical guide for the electromyographer: the limbs and trunk. Springfield: CC Thomas; 1994;
  35. Vaughan CL, Davis BL, O’Connor JC. Dynamics of human gait. 2nd ed.. Western Cape: Kiboho; 1999;
  36. Burstein AH, Wright TM. Fundamentals of orthopaedic biomechanics. Baltimore: Williams & Wilkins; 1994;
  37. Field AP. Discovering statistics using SPSS for Windows: advanced techniques for the beginner. London: Sage; 2000;
  38. Glass GV, Hopkins KD. Statistical methods in education and psychology. 3rd ed.. Boston: Allyn & Bacon; 1996;
  39. Meythaler JM, DeVivo MJ, Hadley M. Prospective study on the use of bolus intrathecal baclofen for spastic hypotonia due to acquired brain injury. Arch Phys Med Rehabil. 1996;77:461–466
  40. Struys MA, Jonkman EJ, Strijers RL. Measurement of patellar and ankle tendon reflexes in normal subjects. Electromyogr Clin Neurophysiol. 1997;37:13–18
  41. Berger W, Horstmann G, Dietz V. Tension development and muscle activation in the leg during gait in spastic hemiparesis: independence of muscle hypertonia and exaggerated stretch reflexes. J Neurol Neurosurg Psychiatry. 1984;47:1029–1033
  42. Ada L, Vattanasilp W, O’Dwyer NJ, Crosbie J. Does spasticity contribute to walking dysfunction after stroke?. J Neurol Neurosurg Psychiatry. 1998;64:628–635
  43. Lamontagne A, Malouin F, Richards C. Contribution of passive stiffness to ankle plantarflexor moment during gait after stroke. Arch Phys Med Rehabil. 2000;81:351–358
  44. Siegler S, Moskowitz G, Freedman W. Passive and active components of the internal moment developed about the ankle joint during human ambulation. J Biomech. 1984;17:647–652
  45. Tardieu C, Lespargot A, Tabary C, Bret M. Toe-walking in children with cerebral palsy: contributions of contracture and excessive contraction of triceps surae muscle. Phys Ther. 1989;69:656–662
  46. Harlaar J, Becher J, Snijders C, Lankhorst G. Passive stiffness characteristics of ankle plantar flexors in hemiplegia. Clin Biomech. 2000;15:261–270(Bristol, Avon)
  47. Singer B, Dunne J, Singer K, Allison G. Evaluation of triceps surae muscle length and resistance to passive lengthening in patients with acquired brain injury. Clin Biomech. 2002;17:151–161(Bristol, Avon)
  48. Lin FM, Sabbahi M. Correlation of spasticity with hyperactive stretch reflexes and motor dysfunction in hemiplegia. Arch Phys Med Rehabil. 1999;80:526–530
  49. Pandyan AD, Price CI, Barnes MP, Johnson GR. A biomechanical investigation into the validity of the modified Ashworth Scale as a measure of elbow spasticity. Clin Rehabil. 2003;17:290–293
  50. Levin MF, Hui-Chan C. Are H and stretch reflexes in hemiparesis reproducible and correlated with spasticity?. J Neurol. 1993;240:63–71
  51. Patrick E, Ada L. The Tardieu Scale differentiates contracture from spasticity whereas the Ashworth Scale is confounded by it. Clin Rehabil. 2006;20:173–182
  52. Dunne JW, Singer BJ, Allison GT. Velocity dependent passive muscle stiffness. [letter] J Neurol Neurosurg Psychiatry. 2003;74:283
  • a JR3 Inc, 22 Harter Ave, Woodland, CA 95776.
  • b Delsys Inc, PO Box 15734, Boston, MA 02215.
  • c Kistler Instrument Corp, 75 John Glenn Dr, Amherst, NY 14228.
  • d SPSS Inc, 233 S Wacker Dr, 11th Fl, Chicago, IL 60606.

 Supported by the National Institute on Disability and Rehabilitation Research (grant no. H133G010066) and the National Institutes of Health (grant nos. HD044295, HD043664).

 No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit upon the authors or upon any organization with which the authors are associated.

PII: S0003-9993(08)00030-0

doi:10.1016/j.apmr.2007.09.051

Archives of Physical Medicine and Rehabilitation
Volume 89, Issue 4 , Pages 700-710, April 2008

Article Tools

* Image Usage & Resolution