Quantitative measures of spasticity in post-stroke patients

Clinical Neurophysiology 111 (2000) 1015±1022

www.elsevier.com/locate/clinph

Quantitative measures of spasticity in post-stroke patients Fabrizio Pisano a,*, Giacinta Miscio a, Carmen Del Conte a, Danilo Pianca a, Elisa Candeloro a, Roberto Colombo b a

Department of Neurology, Unit of Neurophysiology, `Salvatore Maugeri' Foundation, IRCCS, Rehabilitation Institute 28010 Veruno (NO), Italy b Department of Bioengineering, `Salvatore Maugeri' Foundation, IRCCS, Rehabilitation Institute 28010 Veruno (NO), Italy Accepted 17 February 2000

Abstract Objective: Quantitative evaluation of muscle tone in post-stroke patients; correlation of biomechanical indices with conventional clinical scales and neurophysiological measures; characterization of passive and neural components of muscle tone. Methods: Mechanical stretches of the wrist ¯exor muscles of 53 post-stroke patients were imposed by means of a torque motor at constant speed. Patients were clinically studied using the Ashworth scale for spasticity and the Medical Research Council score for residual muscle strength. The neurophysiological measures were Hoffmann re¯ex latency, Hmax/Mmax ratio, stretch re¯ex threshold speed (SRTS), stretch re¯ex (SR) latency and area, passive (ISI) and total (TSI) stiffness indices. Results: Hmax/Mmax ratio, SR area, ISI and TSI values were signi®cantly higher in patients, while SRTS was signi®cantly lower. TSI, SRTS and SR area were highly correlated to the Ashworth score. Conclusions: This EMG-biomechanical technique allows an objective evaluation of changes in muscle tone in post-stroke patients, providing easily measurable, quantitative indices of muscle stiffness. The linear distribution of these measures is particularly indicated for monitoring changes induced by treatment. The apparatus seems suitable to characterize neural stiffness, while dif®culties were found in isolating the passive components, because of the occurrence of tonic EMG activity in most spastic patients. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Spasticity; Muscle stiffness; Electromyography; Stretch re¯ex; H re¯ex

1. Introduction Spasticity is attributed to increased muscle tone associated with hyperre¯exia according to Lance (1980) who de®ned spasticity as a `velocity-dependent increase of tonic stretch re¯exes with exaggerated tendon jerks'. Previous efforts to quantify spastic hypertonia have concentrated on clinical scales (Ashworth, 1964; Pedersen, 1969; Bohannon and Smith, 1987), electromyographic and biomechanical analysis of limb resistance to passive or voluntary movements (Gottlieb et al., 1978; Bohannon, 1987; Cody et al., 1987; Ibrahim et al., 1993; Toft et al., 1993; Fellows et al., 1994), gait analysis (Dietz et al., 1981; Berger et al., 1982; Ada et al., 1998) and a host of electrophysiological re¯ex studies (Burry, 1972; Norton et al., 1972; Lance, 1980; Eisen, 1987; Katz and Rymer, 1989; Hilgevoord et al., 1994). In spite of this broad range of techniques, no

* Corresponding author. Tel.: 139-322-884-711/723; fax: 139-322-830294. E-mail address: neu®[email protected] (F. Pisano).

uniformly useful objective measurements have emerged in the clinical practice. An objective, quantitative measure would achieve widespread clinical acceptance only if its variations broadly paralleled an accepted clinical scale. An important criterion that objective parameters have to ful®l to gain everyday clinical acceptance is consistency and sensitivity (Katz et al., 1992). Clinical scales, such as those proposed by Ashworth (1964), offer qualitative information, but lack temporal and interexaminer reproducibility and suffer from a clustering effect in that most of the patients are grouped within the middle grades (Katz et al., 1992). Nevertheless they have been and continue to be widely used in the study of spasticity (Penn et al., 1989) and are the present yardstick against which newer, more exact methods must be compared. The resistance clinically appreciated even in normals during passive movement of a limb is called stiffness: it expresses the increment in force developed by the muscles in response to a change in length. On this premise we previously evaluated muscle tone in normals at the upper

1388-2457/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 1388-245 7(00)00289-3

CLINPH 99685

1016

F. Pisano et al. / Clinical Neurophysiology 111 (2000) 1015±1022

limb by employing a torque motor which imposed wrist extension displacements at different speeds. We also quanti®ed the passive and neural components of muscle stiffness (Pisano et al., 1996). In the present paper we applied the same methodology to a group of post-stroke patients, presenting different degrees of muscle hypertonia and paresis, with the aim to: (a) quantitatively evaluate the increase in muscle tone, (b) correlate the computed biomechanical indices with the conventional clinical scales and neurophysiological measures, (c) characterize the intrinsic and neural components of muscle tone in the hemiparetic patients.

2. Materials and methods Fifty-three (39 males and 14 females) post-stroke patients were examined; their mean ^ SD age was 45:3 ^ 13:7 years (range 21±70). All patients gave their informed consent to the investigation, which was approved by the local Ethics Committee. Patients with severe Wernicke's aphasia, hyperalgesia, ®xed contractures in the limb examined or neuropathic signs were excluded from the study. At the time of the examination, patients with a mild degree of hypertonia were not receiving antispastic or myorelaxant drugs; the others underwent a wash-out period after their informed consent. Clinical features of the post-stroke patients are reported in Table 1. For control values we referred to the data in 48 normal subjects described in the previous paper quoted above (Pisano et al., 1996). Each patient underwent the following evaluations: 1. clinical examination of muscle tone of the forearm ¯exors, by means of the modi®ed Ashworth scale, which is articulated on six levels (Bohannon and Smith, 1987; Katz et al., 1992). The Medical Research Council (MRC) score, on 6 levels, was adopted as a measure of residual strength in the wrist ¯exor muscles (Medical Research Council, 1943); 2. The Hoffmann (H) re¯ex was recorded by placing disposable surface electrodes (Dantec 13L20) in correspondence to the muscle belly of the ¯exor carpi radialis (FCR) muscle with an interelectrode distance of 3 cm, and stimulating the median nerve at the elbow. Stimulus duration was 1 ms (Panizza et al., 1989). We measured

both the H re¯ex latency and Hmax/Mmax ratio. A twochannel electromyograph (Medelec Mystro Plus), antialising ®ltered in the 20 Hz to 2 kHz band, was employed; 3. For torque analysis, subjects were comfortably seated beside a servo-controlled DC torque motor. The forearm was ®xed in an adjustable support and the hand was strapped semipronated to the handle of a manipulandum coupled to a computer-controlled torque motor (CemParvex T4C2D). The hand and wrist were initially positioned at a relative angle of 1808. This position was chosen in order to compare the ®ndings of the post-stroke patients with those of the normal group. The torque motor induced wrist extension at a constant speed (ramp movements); the angle displacement was 508. The motor shaft was connected to torque, tachometer and position transducers. EMG activity of the FCR and extensor carpi radialis muscle was recorded by means of disposable surface electrodes. Fig. 1 shows the system apparatus. Data from the wrist extensor muscle have been reserved for a future study, and will not be commented on in the present paper. The position, torque and recti®ed EMG signals were all acquired at a sampling rate of 1000 Hz and averaged. The display of acquired signals is reported in Fig. 2. The onset and end points of the SR from the ¯exor muscles were automatically measured at the point where the signal exceeded and returned below a threshold level corresponding to the mean ^ 3 SD of baseline activity (Bedingham and Tatton, 1984). From this the measures of SR latency and area were obtained. Linear regression analysis of the torque/position curves was calculated. The slope of the regression line was considered as a measure of wrist stiffness. The ®rst and last 58 of the displacements were excluded from regression calculation because of acceleration and deceleration artifacts. All subjects performed three sessions a day at the same hour. The following parameters were measured: 1. SRTS, the minimum velocity able to evoke the SR in the wrist ¯exor muscles, in at least 5 of 10 extension displacements. The sequence of extension displacements was delivered at pseudo-random intervals (between 10 and 20 s) by increments of 108/s, starting from 108/s until SRTS was reached; 2. Latency and area of the SR, evoked at 5008/s (the maximal speed allowed by the torque motor to enable compar-

Table 1 Clinical features of post-stroke patients (n ˆ 53; 39 m, 14 f) Age (years)

Distance from stroke (months)

Ashworth scale

MRC

Type of lesion

45.3 ^ 13.7 Range 21±70

30.7 ^ 40.7 Range: 1±228

2.4 ^ 1.2 Range 0±5

1.3 ^ 13 Range 0±4

12 hemorrhagic 41 ischemic

F. Pisano et al. / Clinical Neurophysiology 111 (2000) 1015±1022

1017

Fig. 1. Block diagram of the system for the quantitative evaluation of muscle tone.

ison with the controls). In order to avoid fatigue and habituation, there was at least a 10 s resting period between stretches (Rothwell et al., 1986); 3. The stiffness indices, corresponding to the slope of the regression line calculated on the torque/position curve, at the speeds of 10 and 2008/s (Fig. 3). The value obtained at 108/s identi®ed exclusively the contribution of the passive muscle properties and the tendon-articular

structures and was named ISI. At 2008/s the resistance due to both the contribution of the non-re¯ex and neural components was expressed by a total stiffness index named TSI. Each parameter was calculated on 10 displacements and the average value was considered for analysis. In all the sessions subjects were instructed to completely

Fig. 2. Signals acquired from a post-stroke patient. From the top: EMG from the ¯exor muscles; EMG from the extensor muscles; the position curve from 0 to 508; the torque curve during the 508 displacement.

Fig. 3. Stiffness indices ISI and TSI in a post-stroke patient. The indices correspond to the slope of the regression lines calculated on the torque/ position curves at 108/s and 2008/s, respectively.

1018

F. Pisano et al. / Clinical Neurophysiology 111 (2000) 1015±1022

relax before the onset of the displacements and `not to intervene voluntarily' during the passive movement. We also measured the basal EMG area (mV ms) over a 100 ms interval preceding the displacement in order to check the relaxation state. 2.1. Statistics All variables were not normally distributed (ShapiroWilk test). Thus, all data except for categorical variables (Ashworth scale, MRC) were normalized by means of a logarithmic transformation, and analysed by parametric statistics. The mean values ^ SD of each parameter for both normals and patients are reported in the tables. Student's unpaired t test was used to compare the results between normals and patients and to compare males and females. To perform the correlations between clinical parameters, neurophysiological and biomechanical measures, we used linear regression analysis. Spearman's rank correlation coef®cient was used with categorical data. The statistical program used was StatView 5.0 by the SAS Institute Inc. 3. Results 3.1. H re¯ex study The H re¯ex, recordable in the FRC muscle in 28 out of the 48 controls, was detected in all 53 patients. The mean H re¯ex latency was 16:6 ^ 1:9 ms (range 14.5±19.2) in normals and 17:1 ^ 1:9 ms (range 15.2±19.6) in patients; the difference between the two populations was not signi®cant (P , 0:43). The Hmax/Mmax ratio was 0:2 ^ 0:07 (range 0.1±0.3) and 0:5 ^ 0:2 (range 0.2±0.8), respectively, in normals and patients: the difference was statistically signi®cant (P , 0:001). 3.2. Torque analysis As previously reported (Pisano et al., 1996), SR was elicitable in only 26 out of the 48 normal subjects, in spite of the very high speed used (5008/s). Therefore, parameters regarding SRTS, SR latency and area necessarily refer to these 26 controls. In the present study, TSI was calculated at 2008/s to reduce the artifacts in the signals recorded. At such velocity, only in 14 out of the 26 normal subjects did we obtain a TSI value suitable for comparison with the patients. ISI and basal EMG area concern all the 48 normal subjects. The above-mentioned parameters were instead measured in the whole post-stroke patient population, except for ISI which was measured in a subgroup of 19 patients. SRTS was signi®cantly lower in patients than in normals (P ˆ 0:0001). Only two of the 26 normal subjects presented a SRTS lower than 1008/s, while 50 out of 53 patients showed a threshold speed lower than 1008/s.

Fig. 4. Distribution of the stretch re¯ex threshold speed (SRTS) in the hemiparetic population (n ˆ 53) and controls (n ˆ 26). On the x axis, angular speed; on the y axis, number of subjects.

The diagram in Fig. 4 represents the distribution of the SRTS in controls and post-stroke patients. The comparison of the SR latency values, obtained in response to a mechanical stretch at 5008/s, did not reveal signi®cant differences between patients and controls (P ˆ 0:67). SR area at 5008/s was signi®cantly larger in patients (P ˆ 0:0001). In addition in most patients a late EMG activity occurred after the stretch, but did not enter the measure of the SR area; such activity was never recorded in normal subjects. Fig. 5 shows the EMG re¯ex activity from the wrist ¯exor muscles in a representative normal subject (A) and patient (B), recorded after 5008/s stretches. For each subject, the waves represent the average of 10 traces. In all the 48 normal subjects, at 108/s, no EMG activity at rest was present and the recorded signal was only due to ampli®er and background noise. The ISI value, that we derived from the slope of the torque/position curve at such velocity, thus expresses a resistance to stretch solely due to the passive properties of the muscle. Of the patients under study, in only 19 did the torque-

Fig. 5. The averaged-recti®ed EMG activity in a normal subject (upper trace-A) and in a post-stroke patient (lower trace-B) recorded after mechanical stretches at 5008/s (10 sweeps).

F. Pisano et al. / Clinical Neurophysiology 111 (2000) 1015±1022

1019

3.3. Correlation between the measures

Fig. 6. Flexor EMG recordings after extension displacement at 108/s in 3 representative post-stroke patients. Group I: EMG activity is absent during the test. Group II: EMG activity is present after the start of displacement. Group III: EMG activity is continuously present.

imposed movement at 108/s not evoke any EMG activity from the ¯exor muscles (Group I). Consequently, only in these cases was it possible to calculate the ISI index, and its value was signi®cantly higher in patients than in controls (P ˆ 0:0019). Sex signi®cantly discriminated ISI between males and females in both normals and patients: i.e. males presented higher values than females probably because of the larger mass of the forearm muscle in the former. In the remaining 34 patients a tonic EMG stretch re¯ex activity of variable intensity had been recorded from ¯exors after the ®rst degrees of displacement at 108/s (n ˆ 9, Group II), or even at rest, i.e. before the displacement occurred (n ˆ 25, Group III) (Fig. 6). This tonic re¯ex EMG activity recorded at 108/s in patients was never recorded in controls. TSI calculated at 2008/s was measured in only 14 normal subjects, because this velocity failed to evoke SR activity in several normal subjects. Also TSI was signi®cantly higher in patients than in controls (P ˆ 0:007). The mean ^ SD values of the parameters derived by means of the torque analysis in normals and patients are reported in Table 2.

H re¯ex latency and Hmax/Mmax did not show signi®cant correlation to the Ashworth scale (r ˆ 0:01 and r ˆ 0:03, respectively). On the other hand, the Ashworth scale showed a signi®cantly high correlation with the measures obtained by the mechanical stretch, namely SRTS and TSI (respectively: r ˆ 20:46; P ˆ 0:0001, and r ˆ 0:55; P ˆ 0:0001). In addition these measures were only mildly correlated to Hmax/Mmax ratio, and not correlated to H re¯ex latency. ISI was not correlated to the Ashworth scale; this is not due to a clustering of the 19 patients in the low levels; in fact their scores at the Ashworth scale were distributed as follows: score 4 ˆ 1 patient, score 3 ˆ 4 patients, score 2 ˆ 9 patients, score 1 ˆ 3 patients and score 0 ˆ 2 patients. TSI was also fairly well correlated to SRTS (r ˆ 0:61; P ˆ 0:0001), and to SR latency and area (respectively: r ˆ 20:33; P ˆ 0:02, and r ˆ 0:42; P ˆ 0:003). Table 3 reports the correlations between the re¯exologic measures and stiffness indices obtained by the biomechanical set-up and the clinical measure of spasticity expressed by the Ashworth scale; also reported is the correlation between the clinical scale and the conventional neurophysiological parameters expressed by the H re¯ex. The correlation coef®cients (r) and P values are listed. MRC did not show signi®cant correlation with any of the measures; thus the corresponding correlation coef®cients were not reported in Table 3.

4. Conclusions In the present study we applied an EMG-biomechanical technique to assess muscle tone in passive conditions in 53 post-stroke patients. Patients were also clinically evaluated by means of the

Table 2 Torque analysis ®ndings in normals and patients

Basal EMG (mVms) SR latency (5008/s) (ms) SR area (5008/s) (mVms) SR threshold speed (8/s) ISI (a.u.)

TSI (a.u.)

Normals (n ˆ 48)

Patients (n ˆ 53)

P value

104.7 ^ 29.1 (n ˆ 48) (33.5±160) 39.9 ^ 8.1 (n ˆ 26) (22±59) 413 ^ 291 (n ˆ 26) (74±1332) 243 ^ 92.1 (n ˆ 26) (70±500) 5.3 ^ 2.7 (n ˆ 48) (0.9±10.7) Males: 6 ^ 2.1 P ˆ 0.001 Females: 3.5 ^ 1.3 8 ^ 3.2 (n ˆ 20) (5.8±11.8)

187.8 ^ 150.7 (n ˆ 53) (33.5±953) 41 ^ 9 (n ˆ 53) (22±69) 4481 ^ 3984 (451±23025) 75 ^ 34.3 (50±200) 8.0 ^ 3.4 (n ˆ 19) (3±13.9) Males: 8.8 ^ 3.3 P ˆ 0.0075 Females: 6.1 ^ 2.5 47.2 ^ 30.1 (n ˆ 53) (6.1±170)

0.0002 0.67 0.0001 0.0001 0.0019

0.007

1020

F. Pisano et al. / Clinical Neurophysiology 111 (2000) 1015±1022

Table 3 Correlations between clinical, physiological and biomechanical measures a

SR latency (ms) SR area (mVms) SRTS (8/s) TSI (a.u.) ISI (a.u.) a

Ashworth scale

H latency (ms)

H/M

Basal EMG (mVms)

20.37** 0.36** 20.46**** 0.55**** 0.37

0.15 0.07 0.31 0.07 0.29

20.52* 0.54** 20.49** 0.51* 0.25

20.07 0.14 20.36* 0.6*** 0.12

*P , 0:05; **P , 0:01; ***P , 0:001; ****P , 0:0001.

modi®ed Ashworth scale, which has been shown to have a high interrater reliability (Bohannon and Smith, 1987; Katz et al., 1992) and is considered the gold standard for the clinical measure of muscle tone (Lee et al., 1989). All subjects were assessed by the conventional neurophysiological technique of H re¯ex study. It is well known that the H re¯ex can identify an enhancement of excitability in the monosynaptic re¯ex arc (Schieppati, 1987), but, when used as an objective measure of spasticity, the Hmax/ Mmax ratio shows a large intersubject variability (Levin and Chan, 1993). In fact patients with a similar clinical degree of spasticity can have different H re¯ex amplitudes and, conversely, patients with comparable sizes of re¯ex response can have different grades of tone (Cody et al., 1987). Indeed a broad overlapping of H re¯ex values has been described in spastics and controls (Delwaide, 1984; Katz et al., 1992); the poor correlation between these values and spasticity scales limits the use of this test in clinical practice. Also in the present study the application of the H re¯ex study yielded rather incoherent ®ndings. In fact we could not demonstrate a correlation between Hmax/Mmax ratio and the severity of spasticity, expressed by the Ashworth scale. For these reasons we agree with other authors claiming that studies utilizing the H re¯ex must be interpreted with caution (Matthews, 1966; Delwaide, 1984; Young and Wierzbicka, 1985; Katz et al., 1992). On the contrary the available biomechanical techniques offer a more reliable measure of spastic hypertonia: similarly to the ®ndings of other studies (Lee et al., 1987; Powers et al., 1988; Powers et al., 1989; Katz and Rymer, 1989), SRTS was found markedly lower in poststroke patients than in normals. The 1008/s velocity represents a sort of cut-off value that discriminates between patients and controls: in fact at a lower speed, only two normal subjects can be found, while 50 out of the 53 patients are gathered here. The lack of difference in the SR latency between normals and patients is in line with data from the literature (Verrier et al., 1984). The SR latency values are much longer if compared to the H re¯ex latency, both in normals and patients. This could be simply due to the passive situation in which we examined our subjects. This experimental approach, even at the maximum velocity provided by the torque motor (5008/s), probably does not allow a synchro-

nous activation of the Ia ®bers, in the way that electrical stimulation of the H re¯ex does. In this study, by using wrist extension movements at different velocities, we were also able to identify two easily measurable parameters quantifying wrist stiffness: ISI and TSI. At 108/s ISI expresses the passive stiffness due to connective tissue and tendon articular structures. Because of the heterogeneity of our patient population and the high degree of re¯ex hyperexcitability in most of them, we were actually able to measure ISI only in a subgroup of 19 patients, who were the only ones who did not show EMG activity in the ¯exor muscles, either at rest or during the 108/s stretch. In this subgroup ISI values were signi®cantly higher than in controls. This ®nding would con®rm the occurrence of changes in the passive properties of the spastic muscles. (Edstrom, 1970; Herman et al., 1974; Hufschmidt and Mauritz, 1985; Dietz et al., 1986; Sinkjaer and Magnussen, 1994). From a practical point of view ISI helps identify the visco-elastic contribution to passive movement rigidity as a whole, hence enabling a more correct rehabilitative approach. In the 34 patients where ISI was not measurable, we observed at 108/s a torque-evoked tonic re¯ex EMG activity, which was never observed in the controls, thus con®rming a previous study performed on the passive biceps brachii, in a similar experimental condition (Thilmann et al., 1991). In this latter, the authors interpreted the EMG activity as a `late stretch re¯ex activity'. TSI, i.e. index of total stiffness, was obtained at 2008/s; TSI includes the contribution of the short latency SR besides the aforementioned late activity. Its values were signi®cantly higher in patients than controls. As demonstrated in the study performed in normals (Pisano et al., 1996), SR is powerful enough to signi®cantly modify the mechanical response to external perturbations. This is particularly evident in post-stroke patients where TSI was correlated to SR latency and area, con®rming the close relationship of re¯ex stiffness not only with the mere occurrence of SR, but also with SR magnitude. The occurrence of SR is strongly correlated to the level of background activity, that facilitates the SR elicitation. It follows that the EMG basal level linearly increases the total neural stiffness (TSI) (Herman and Schaumberg, 1968).

F. Pisano et al. / Clinical Neurophysiology 111 (2000) 1015±1022

The increase in the gain of EMG activity described in spastic subjects in passive conditions (Ibrahim et al., 1993; Thilmann et al., 1991), has been con®rmed also in our study with the ®nding of an extra-EMG activity related to the pathological SR. Moreover passive conditions offer a direct comparison with clinical experience and allow to obtain quantitative information about muscle tone even in those severely paretic patients unable to perform active motor tasks. The main ®nding of our experimental protocol is that most of the measures drawn out by means of the biomechanical technique are well correlated to the Ashworth scale. This is particularly relevant for SRTS and TSI. As claimed at the beginning, this correlation represents an indispensable condition for the clinical acceptance of an objective measure. We would also underline the possibility that this device provides an objective quanti®cation of spasticity in contrast to the qualitative and subjective description of muscle hypertonia given by the clinical scale, necessarily dependent on the examiner's perception. Furthermore this EMG-biomechanical evaluation of muscle tone offers widely linearly distributed values, thus avoiding the clustering effect due to the grouping in 6 levels of the clinical scale, and enhancing even mild but signi®cant changes within a single level. In conclusion, we believe that the application of biomechanical techniques combined with EMG recordings represents a useful tool to quantify spasticity in the wrist ¯exor muscles. The strong correlation of the biomechanical measures with the clinical scores allows their application in clinical practice, in particular to objectively evaluate changes secondary to pharmacological or physical therapies. Acknowledgements This research was partially ®nanced by the Istituto Superiore di SanitaÁ with 1% National Health Service Funds, art. 502. We wish to thank Prof Marco Schieppati for his critical review of the manuscript, Rosemary Allpress for help with the English, Fabio Comazzi for statistical support and Stefania Bicelli for technical assistance. References 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. Ashworth B. Preliminary trials of carisoprodol in multiple sclerosis. Practitioner 1964;192:540±542. Bedingham W, Tatton WG. Dependence of EMG responses evoked by imposed wrist displacements on pre-existing activity in the stretched muscles. Can J Neurol Sci 1984;11:272±280. Berger W, Quintern J, Dietz V. Pathophysiology of gait in children with cerebral palsy. Electroencephalogr Clin Neurophysiol 1982;53(5):538± 548. Bohannon RW. Variability and reliability of the Pendulum test for spasticity using Cybex II isokinetic dynamometer. Phys Therapy 1987;67:659±661.

1021

Bohannon RW, Smith MB. Interrater reliability on a modi®ed Ashworth scale of muscle spasticity. Phys Therapy 1987;67:206±207. Burry HC. Objective measurement of spasticity. Dev Med Child Neurol 1972;14:508±510. Cody FWJ, Richardson HC, MacDermott N, Ferguson T. Stretch and vibration re¯exes of wrist ¯exor muscles in spasticity. Brain 1987;110:433± 450. Delwaide PJ. Contribution of human re¯ex studies to the understanding and management of the pyramidal syndrome. In: Shahani BT, editor. Electromyography in CNS Disorders: central EMG, Boston, MA: Butterworths, 1984. pp. 77±109. Dietz V, Quintern J, Berger W. Electrophysiological studies of gait in spasticity and rigidity: evidence that altered mechanical properties of muscle contribute to hypertonia. Brain 1981;104:431±449. Dietz V, Ketelsen UP, Berger W, Quintern J. Motor unit involvement in spastic paresis: relationship between leg muscle activation and histochemistry. J Neurol Sci 1986;75:89±103. Edstrom L. Selective changes in the sizes of red and white muscle ®bres in upper motor lesions and Parkinson. J Neurol Sci 1970;11:550±573. Eisen A. Electromyography in disorders of muscle tone. Can J Neurol Sci 1987;14:501±505. Fellows SJ, Kaus C, Ross HF, Thilmann AF. Agonist and antagonist EMG activation during isometric torque development at the elbow in spastic hemiparesis. Electroencephal Clin Neurophysiol 1994;93(2):106±112. Gottlieb GR, Agarwal GC, Penn R. Sinusoidal oscillation of ankle as a means of evaluating spastic patients. J Neurol Neurosurg Psychiatry 1978;41:32±39. Herman R, Schaumberg H. Alterations in dynamic and static properties of the stretch re¯ex in patients with spastic hemiplegia. Arch Phys Med Rehabil 1968;49(4):199±204. Herman R, Freedman W, Mayer N. Neurophysiologic mechanisms of hemiplegic and paraplegic spasticity: implications for therapy. Arch Phys Med Rehabil 1974;55:338±343. Hilgevoord AA, Koelman JH, Bour LJ, Ongerboer de Visser BW. Normalization of soleus H-re¯ex recruitment curves in controls and a population of spastic patients. Electroenceph clin Neurophysiol 1994;93(3):202±208. Hufschmidt A, Mauritz KH. Chronic transformation in spasticity: peripheral contribution to increased tone. J Neurol Neurosurg Psychiatry 1985;48:676±685. Ibrahim IK, Berger W, Trippel M, Dietz V. Stretch-induced electromyographic activity and torque in spastic elbow muscles. Brain 1993;116:971±989. Katz R, Rymer WZ. Spastic hypertonia: mechanisms and measurement. Arch Phys Med Rehabil 1989;70:144±155. Katz R, Rovai GP, Brait C, Rymer WZ. Objective quanti®cation of spastic hypertonia: correlation with clinical ®ndings. Arch Phys Med Rehabil 1992;73:339±347. Lance JW. Control of muscle tone, re¯exes and movement: Robert Wartenberg Lecture. Neurology 1980;30:1303±1313. Lee WA, Boughton A, Rymer WZ. Absence of stretch re¯ex gain enhancement in voluntarily activated spastic muscle. Exp, Neurol 1987;98:317± 335. Lee KC, Caron L, Kinnin E, Patterson V. The Ashworth scale: a reliable and reproducible method of measuring spasticity. J Neurol Rehab 1989;3:205±209. Levin MF, Chan HC. Are H and stretch re¯exes in hemiparesis reproducible and correlated with spasticity? J Neurol 1993;240:63±71. Matthews WB. Ratio of maximum H re¯ex to maximum M response as a measure of spasticity. J Neurol Neurosurg Psychiatry 1966;29:210±214. Medical Research Council. Aids to the investigation of peripheral nerve injuries, War Memorandum, London: HMSO, 1943. pp. 11±46. Norton BJ, Bomze HA, Chaplin H. An approach to the objective measurement of spasticity. Phys Therapy 1972;52:15±23. Panizza M, Nilsson J, Hallett M. Optimal stimulus duration for the H re¯ex. Muscle Nerve 1989;12:576±579.

1022

F. Pisano et al. / Clinical Neurophysiology 111 (2000) 1015±1022

Pedersen E. In: Thomas CC, editor. Spasticity: mechanism, measurement, and management, Spring®eld, 1969. pp. 36±54. Penn RD, Savoy SM, Corcos D. Intrathecal baclofen for severe spinal spasticity. New Engl J Med 1989;320:1517±1521. Pisano F, Miscio G, Colombo R, Pinelli P. Quantitative evaluation of normal muscle tone. J Neurol Sci 1996;135(2):168±172. Powers RK, Marder-Meyer J, Rymer WZ. Quantitative relations between hypertonia and stretch re¯ex threshold in spastic hemiparesis. Ann Neurol 1988;98:115±124. Powers RK, Campbell DL, Rymer WZ. Stretch re¯ex dynamics in spastic elbow ¯exor muscles. Ann Neurol 1989;25:32±42. Rothwell JC, et al. Habituation and conditioning of the human long latency re¯ex. Exp Brain Res 1986;63:196±214. Schieppati M. The Hoffman re¯ex: a means of assessing spinal re¯ex excitability and its descending control in man. Prog Neurobiol 1987;28:345±376.

Sinkjaer T, Magnussen I. Passive, intrinsic and re¯ex-mediated stiffness in the ankle extensors of hemiparetic patients. Brain 1994;117:355±363. Thilmann AF, Fellows SJ, Garms E. The mechanism of spastic muscle hypertonus. Brain 1991;114:233±244. Toft E, Sinkjaer T, Andreassen S, Hansen HJ. Stretch responses to ankle rotation in multiple sclerosis patients with spasticity. Electroenceph clin Neurophysiol 1993;89(5):311±318. Verrier MC, Tatton WG, Blair RDG. Characteristics of EMG responses to imposed limb displacement in patients with vascular hemiplegia. Can J Neurol Sci 1984;11:288±296. Young RR, Wierzbicka MD. Behaviour of single motor units in normal subjects and in patients with spastic paresis. In: Delwaide PJ, Young RR, editors. Clinical neurophysiology in spasticity: contribution to assessment and pathophysiology, Amsterdam: Elsevier, 1985. pp. 27± 40.