Patients with Spastic Hemiplegia at Different of Early/Late Reflex Responses

38638ournal ofNeurology, Neurosurgery, and Psychiatry 1993;56:386-392

Patients with spastic hemiplegia at different J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.4.386 on 1 April 1993. Downloaded from recovery stages: evidence of reciprocal modulation of early/late reflex responses. I K Ibrahim, M A R El-Abd, V Dietz

Abstract of about 25 and 50 ms respectively. Although Reflex electromyographic (EMG) muscle it was suggested that VI was an unmasked H responses were recorded from abductor reflex, the origin of V2 could not be deter- pollicis brevis (APB) and tibialis anterior mined. Recently, interest has focused upon (TA) muscles of fifty patients with spas- the behaviour and origin of the late compo- tic hemiplegia. Responses in the muscles nents of mechanically and electrically induced were evoked during voluntary muscle reflexes.) 4 Milner-Brown et a13 interpreted contraction (about 20% of maximum vol- the findings of Upton et al in such a way that untary effort) by submaximal but they considered the possibility that V2 has a suprathreshold electrical stimulation of long-loop pathway via the motor cortex. the median (at the wrist) and common Conrad and Aschoff5 supported this conclu- peroneal (at the neck of the fibula) sion and also suggested that cutaneous signals nerves respectively. Three EMG peaks also contribute to the afferent cortical input (Rl, R2 and R3) could be recorded after of the late response. Stanley7 has shown that the direct muscle response (M). There both responses are evoked independently was only a slight difference in Ri-R2 through reflex pathways: the first, with a latency interval of about 5 ms between mean afferent conduction velocity of 64 m/s upper and lower limbs on the unaffected and an estimated central delay of about side of the patients making it unlikely 0-8 ms and the second with a mean afferent that this late response of the lower limb conduction velocity of 43 m/s and an esti- involves a long loop pathway, although mated central delay of about 17 ms. The ori- this possibility cannot be discounted for gin of the long-latency reflex response has the later, R3, response. Reflex behaviour been the subject of controversy.8 The was analysed for three clinical identifi- supraspinal origin of the long-latency reflex able recovery stages of voluntary move- responses has been considered by several ments in the spastic limbs (synergistic, workers,9 12 but questioned by others.'3 ' isolated and useful movements). The Alternatively, there may be differences in the major finding was that an increase in the behaviour and pathways of this reflex in amplitude of the early response "Rl" upper and lower limb muscles," '9 as well as

was associated with a decreased ampli- in proximal and distal muscles.23 http://jnnp.bmj.com/ tude and delayed latency of the late In patients with spastic paresis, the behav- response "IR2" on the spastic side. The iour of early and late reflex activity is altered amplitude of Rl in the three different in that reduced inhibition of the former is recovery stages decreased significantly, contrasted by a reduction or loss of the whereas the amplitude of R2 increased latter.2-23 The enhanced activity of the early significantly with improvement of the stretch reflex is thought to arise from a reduc- functional stage of the limb. A significant tion in reflex threshold without significant negative linear correlation was found enhancement of the reflex gain.- - It was on September 25, 2021 by guest. Protected copyright. between Rl and R2 amplitude changes in argued that the absence of long latency upper as weli as lower limbs. A refrac- stretch reflexes is due to interruption of the Department of toriness of the motor neuron pool as a long-loop (transcortical) pathway. Others Physical Medicine, possible explanation for the decreased R2 attribute this (at least for leg muscles) to reci- University of mono- and Alexandria, amplitude could be discounted. These procal modulation of polysynaptic Alexandria, Egypt findings together with recent work on spinal reflex responses by brain centres.'27 I K Ibrahim reflex development in children support In the light of these views, the aim of this M A R El-Abd the hypothesis of reciprocal modulation work was to quantify the behaviour of early Department of of early and late reflex signals by and late reflex activities in spastic patients Clinical Neurology and Neurophysiology, supraspinal motor centres. with different degrees of clinical recovery. University of Freiburg, Freiburg, (J Neurol Neurosurg Psychiatry 1993;56:386-392) Germany V Dietz Methods Correspondence to: Subjects Dr V Dietz, Paraplegic Upton et al' have shown that early (VI) and Experiments were conducted on both sides of Centre Belgrist, Forchstr 340, CH8008 Zurich, late (V2) responses (in addition to direct 5 apparently healthy volunteers aged between Switzerland muscle response) may be recorded in the 37 and 50 years, and 50 stroke patients (32 Received 21 February 1992 intrinsic muscles of the hand, after stimula- males and 18 females) with spastic hemiple- and in revised form varied from 25 to 71 26 May 1992. tion of the ulnar or median nerves during a gia. The patients' ages Accepted 19 June 1992 voluntary contraction, with latencies to onset years [mean (SD) 57 4 (9 8) years]. The Reflexes in spasticity 387

Table Patients with spastic hemiplegia at different recover) stages: evidetnce of reciprocal

seated with the forearm and J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.4.386 on 1 April 1993. Downloaded from nmodulation ofearly/late comfortably reflex responses hand lying palm upwards on a pillow. Two Upper limbfinctional stage 9 mm silver/silver chloride disc surface SYN ISOL USEF Total (UL) recording electrodes were positioned so that Lower limlb the active recording electrode was over the fiunctional ni = 24 i = 16 nI = 10 II = 50 stage 48% 32% 20% 100% belly of the muscle and the reference recording electrode was at the metacarpophalangeal SYN n= 11 n=6 0 n = 17 34%X 22'% 12'%, upper joint of the thumb. The ground electrode was and placed at the wrist between the stimulating lower ISOL n = 13 n = 10 0 limbs and the recording electrodes. n = 23 46% 26%0, 20'% combined For TA muscle recording, the subject lay groups USEF 0 0 n=10 comfortably in a supine position with the n = 10 20°% 201% lower limb to be tested lying straight and Total (LL) n = 50 Total n = 50 100% 100% (patients) slightly elevated by a sandbag placed under the calf muscles. Bipolar surface recording Recovery stage of voluntary movements in upper and lower limbs and the distribution of the combined groups among the studied cases. n = number, UL = upper limb, LL = lower limb, electrodes were placed 3 cm apart over the SYN = synergistic movement, ISOL = isolated movement and USEF = useful movement. belly of the muscle at the junction of the upper and middle third of the leg. The active duration of the symptoms varied from 6 to 60 recording electrode was proximal to the Lefer- months [mean (SD) 16-1 (10-2) months). ence one. The ground electrode was placed The right to left ratio of the spastic side was between the stimulating and recording elec- 36/14. The stroke was caused in most trodes. instances by a thrombosis (n = 49) and only one case was attributed to an embolism. Stinmulations Each of the patients satisfied the following Two silver electrodes were used for nerve criteria: 1) exclusive unilateral motor disor- stimulation. They were covered with saline- der; 2) clinical signs of upper motor neuron soaked felt and spaced 2-5 cm apart in a syndrome at the affected side, that is, exten- holder. The electrodes were placed on the sor plantar response, increased muscle tone skin overlying the nerve at the point at which and exaggerated tendon tap reflexes; 3) no the lowest motor activation threshold could clinical evidence suggesting the presence of be achieved. To study TA muscle the stimu- any other neuromuscular disease (such as, lating electrodes were positioned over the neuropathy); 4) The patient had suffered the common peroneal nerve behind the neck of stroke at least six months previously so that the fibula whereas for the investigation of the adaptation to the existing neurological deficits APB muscle the electrodes were situated over could be guaranteed. In all patients the sensa- the median nerve at the wrist. The stimuli tions were not clinically affected. The unaf- were rectangular voltage pulses, 0 5 ms in fected side of the patients was clinically duration and delivered at 0 5 Hz. The inten- normal (the so-called unaffected side is sity was submaximal but suprathreshold described as frequently abnormal.28 The stage (about 55 V) and just evoked a weak visible of recovery of the spastic upper and lower muscle contraction. Toennies Electromyo-

limbs of the patients was classified into one of graphic apparatus was used for stimulation http://jnnp.bmj.com/ the following groups: stage (I) of synergistic and recording. movement characterised by flexion synergy in the upper limb and extension synergy in the Testing procedure lower limb. In these instances the patient During the testing, the subject was instructed moves the whole limb synergistically when he to mildly contract the muscle under investiga- or she tries to move one joint. Stage (II) of tion so that the electrical activity recorded isolated movement in which the patient is remained at approximately 20% of maximum capable of performing independent joint voluntary effort measured on the unaffected on September 25, 2021 by guest. Protected copyright. movements. Stage (III) in which the patient side, and to make the contractions as isomet- can perform functionally useful movements. ric as possible by maintaining the force. For upper limb these movements include Stimuli were delivered only while the subject manual and finger dexterity (the ability to was contracting his muscle consistently. work with the hand in grasping, placing and Muscle electrical activity was amplified (band turning motions and the ability to manipulate width of 15 Hz to 5 kHz) and averaged auto- small objects with the fingers), whereas in the matically (time locked to each stimulus) over lower limb complete clearance of the foot 64 or 128 or 256 sweeps (depending upon from the ground during walking can be the size of the evoked responses). In figures achieved. The table shows the distribution of where EMG recordings are shown, two such these three recovery stages of voluntary averaged responses have been superimposed. movement in the upper and lower limbs and The latencies (in ms) of reflex responses were the distribution of the combined groups measured as the time from the stimulus arti- among the cases studied. fact to the point at which the recording deflects from the base line. The amplitudes Recordings (in V) were measured peak to peak for the The reflex potentials were recorded from the largest response. To analyse the amplitude abductor pollicis brevis (APB) and tibialis changes of reflexes between the three groups, anterior (TA) muscles. response amplitude was normalised by divid- For APB muscle recording, the subject was ing the spastic side value by the unaffected 388 Ibrahim, El-Abd, Dietz

side one in each patient. Side Spastic J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.4.386 on 1 April 1993. Downloaded from Unaffected Further recordings were made from APB muscles of two healthy subjects and three A: Abd Poll Brev m patients (unaffected side) during isotonic contractions. The subject was instructed to abduct and adduct the thumb smoothly and 1 1.1 - I alternatively perpendicularly to the palm. The frequency of movement (0 3 Hz) was con- trolled by auditory signals from a mechanical metronome. Stimuli were delivered only during the abduction movement. -rr,LL 3-2<~~i-IL Statistical methods zr Statistical analysis of data was performed MOis 10 ms 200 ~v I using an SPSS/PCtm package run on a com- puter system. The different tests used were: paired t test, one way analysis of variance B: Tib Ant m (ANOVA), multiple range test (Scheffe's test), correlation coefficient (r) (Pearson test), and linear regression.

Results REFLEX EMG RESPONSES a) Unaffected side Distinct, early (Rl) and late (R2) waves (in addition to the direct muscle "M" response) could be recorded from APB and TA in all patients and healthy subjects. A later (R3) wave was infrequently observed and could be recorded from APB in only 9 patients and from TA in only 3 patients. Figure I Two superimposed, averaged EMG responses recorded froni abductor pollicis Figure 1 shows two superimposed averaged brevis (A) and tibialis anterior (B) muscles on the unaffected (left traces) and the spastic EMG reflex patterns from the linaffected and (right traces) sides in a stage I patient (capable ofsynergistic movements). RI of the . s in a stage I spastic side has a very high amplitude while R2 is absent. The shape ofM wave is the spastic upper and low. distorted in the unaffected TA because ofhigh averaged amplitude.

Abd poll brev m Unaffected Side Spastic Rl R2

A: Abd Poll Brev m http://jnnp.bmj.com/ I WI I >d Iim |RRlII I___ a) R2 IR1 Al11 -P -mI IR2 'a . I A E

11 ?r d on September 25, 2021 by guest. Protected copyright. .1.11 I_I. 10 ms 9 Ant tibial m 50 .Lv 10 Ms. 1 Rl FL'f 50 C__-0In R2 B B: Tib Ant m 5

1 O

E I <: ** I IkI I I 50AVI |.|wA1| | Spastic Unaffected

Figure 3 Mean (SD) values of RI and R2 amplitude on the unaffected and the spastic side of the 50 patients. In the Figure 2 EMG responses in a stage II patient (capable of isolated movenlents). R2 has a spastic side RI amplitude is increased while R2 amplitude delayed latency and a decreased anmplitude on the spastic side. is decreased. **: p < 0 01, ***: p < 0 001. Reflexes in spasticity 389

Figure 4 Mean (SD) These latency and amplitude values for Rl A Abd poll brev m B Ant tibial m J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.4.386 on 1 April 1993. Downloaded from values ofRI (top) and R2 (bottonm) anmplitude and R2 are about the same as those obtained (normalised data: spastic Rl from healthy subjects. It was noticed that Rl side value was divided by latency in the unaffected lower limb was the unaffected side value in slightly longer than that of the unaffected each patient) in abductor pollicis brevis (left traces) upper limb. This difference (about 2-5 ms) and tibialis anterior (right can be explained by the limb length differ- traces) muscles for each of ences. R2 and even later EMG peak, R3, the three functional recovery stages. R2 could latencies in the unaffected lower limbs are not be recordedfrom the considerably longer than that of the unaf- tibialis anterior muscle of fected upper limbs (about 10 ms for R2 and stage I patients. As linmb about 20 ms for R3) a latency that cannot be function recovers, RI 5) amplitude decreases while explained solely by difference in limb length. R2 increases significantly (p < 0O01;ANOVA). b) Spastic side The difference between 4) R2 each two stages is The major finding on the spastic side of significant (p < 0 01; 0. patients was an increased amplitude of Rl Scheffe's test) in all E and a decreased amplitude with increased responses. latency of R2 compared with the unaffected side. RI was observed in all spastic upper and lower limbs while R2 could be seen in the APB and TA muscles of 26 and 25 patients respectively. R3 was not usually observed from the spastic side. It is seen from fig 1 that the spastic side RI wave had a very high amplitude in comparison to that of Rl | Useful 5 Isolated recorded from the unaffected side, whereas = Synergist R2 was not present. These recordings were taken from a stage I patient. Figure 2 displays another example of the EMG reflex pattern of subject (capable of synergistic movements both the unaffected and spastic upper and only). The reflex potentials started with a lower limbs but in this case from a stage II negative wave and were always biphasic. The patient (capable of performing isolated move- main deflection was negative for APB and ments). Although the R2 wave is present on positive for TA. The mean (SD) latencies of the spastic side it shows an increased latency APB and TA were; for RI 28-9 (2-3) ms and and diminished amplitude in comparison 31-4 (3 0) ms respectively, for R2, 48-9 (5-7) with the unaffected side at this stage of ms and 57-5 (5-1) ms respectively and, for R3 functional recovery. was 68-1 (14-5) ms and 90 7 (15-8) ms The mean (SD) latencies of the RI wave in respectively. The mean (SD) amplitude of RI APB and TA muscles on the spastic side

in APB and TA was 161-3 (90 7) puV and [29-1 (2-6) and 31-6 (2-7) ms respectively)] http://jnnp.bmj.com/ 132-9 (64-0) uV respectively, of R2, 94-6 were about the same as those obtained from (45 2) ,iV and 65- 1 (35 4) 1iV respectively the unaffected side. However, the mean and, of R3, 45X5 (9-5) uV and 60-25 latencies of the R2 wave in APB and TA (49-9) pV respectively. muscles on the spastic side [55.9 (8-9) and 601 (7.7) ms respectively] was increased compared with the unaffected side (p < 0-001 < A: Abd Poll Brev m B: TibAntm for APB, and p 0-05 for TA). Figure 3 shows that the mean (SD) ampli- on September 25, 2021 by guest. Protected copyright. 6-0 6-0 - tude of the RI wave of APB and TA on the "a spastic side [521-1 (426-3) and 423-4 (348-1) c ,uV respectively)] was considerably increased I* 4-5 (p < 0-001) compared with the unaffected

* * side. The mean (SD) amplitude of the R2 ** wave of APB and TA on the spastic side -3-0- 3-0 + 0) [51-8 (26-1) and 38-7 (22-0) pV respectively] 'a) ** * was significantly decreased compared with * + * the unaffected side < E 155- (p 0-01). * R 1*5 .i: R QUANTIFICATION OF RI AND R2 RESPONSES IN THE THREE FUNCTIONAL STAGES 0o 0 , , Figure 4 shows that the mean amplitude 0-3 0-6 0-9 0-6 0-8 1 (normalised data) of RI in each of the three R2 Amplitude (relative units) functional groups decreases significantly (p < 0 01), while the mean amplitude of Figure 5 Scatter plot ofRJ-R2 amplitude values (norn,nalised data) and the R2 increases significantly (p < 0 01) with corresponding regression lines of abductor pollicis brevis ("A: n = 26) and tibialis anterior improvement of the functional stage of the (B: n = 25) muscles on the spastic side. A significant ne?gative linear correlation is present limb. and R2 in APB = -0 < 0 as well as in TA = p < between Rl (r 74; p 001) (r -0-59; vti 0-01). *: one case, + : two cases. This reciprocal activity was further quanti- 390 Ibrahinm, El-Abd, Dietz

Figure 6 The prevalence tion upon the amplitude of the reflex of the R2 response recorded ---- Lower limb J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.4.386 on 1 April 1993. Downloaded from 120 response in APB muscles of two healthy sub- from APB (solid line) and Upper limb TA (broken line) muscles jects and the unaffected side of three patients in each of the three was investigated. It was evident that during functional stages. ;t 80- isotonic thumb movements the amplitude of a) both reflexes, particularly that of R2, was increased compared with isometric contraction. The amplitude of RI increased by 30 to C 40 co0. 300% and that of R2 increased by 120 to <: 500%. Consequently, under these conditions, a) several EMG patterns were characterised by very high RI amplitudes (in the range of Syn Isol Usef those seen in the spastic side in the absence Functional recovery stage of R2) which were, however, associated with large R2 (see fig 7). fied using scatter plots and regression line correlating the amplitude change (normalised values) of RI and R2 (fig 5). A significant Discussion The main findings of our study were: 1) evi- negative linear correlation was found between dence for a different response both responses in APB (r = -0 74; p < origin of late "R2" in upper and lower limb muscles; 0 001) as well as TA (r = -0-56; p < 0 01) 2) muscles. The normalisation of the data to the increased amplitude of RI associated with decreased unaffected side was responsible for uncover- amplitude and delayed latency of R2 in and correlation of reci- ing this significant correlation as using the spastic limbs, 3) absolute values of amplitude of the spastic procal behaviour of RI and R2 response side revealed insignificant negative correlation amplitudes in the different recovery stages of hemiplegia. in APB (r = -0 1 p > 0 05) as well as in TA (r = - 0 23 p > 0 05). This is due to the wide Possible origin oflate response range of normal Rl and R2 amplitudes (see amplitude values). The R1-R2 amplitude For the early response (RI) an oligosynaptic can (absolute values) correlation on the unaf- spinal pathway be assumed for both upper and lower limb." Also for fected side was found to be insignificant in the late response (R2) in hand muscles a transcortical APB (r = 0-29 p > 0 05) as well as in TA (r = most to 0 18 p> 0-05). pathway is probable according earlier works.8 1920 This may be different The prevalence of R2 responses was differ- for the lower limb muscles.'6 In this an Ri ent in each of the recovery stages in the upper study -R2 interval of about 20 ms in hand and of (n = 26) and lower (n = 25) limbs. Figure 6 about 25 ms in leg muscles was found. The shows that R2 seldom appears in poorly func- differ- ence (about 5 ms) between the upper tioning limbs (stage I) but is much more evi- limb and the lower limb argues dent as recovery progresses through stages II against the possibility of a transcortical for R2 and III in which the limb can at least perform pathway in

lower limb as (using the fastest http://jnnp.bmj.com/ isolated movements. It was also noticed that conducting fibres) an extra of 10 ms small indi- in spastic limbs where R2 was not seen there delay (in viduals) to 20 ms is was a large RI peak. The R2 latency changes (in large individuals) needed for an impulse to travel the distance among the three groups were insignificant between the lumbar and cervical motor neu- (not shown in a Fig). ron pools twice during the reflex transmission see et RI AND R2 RESPONSES DURING ISOTONIC (for calculation Darton al'8). Further- CONTRACTIONS OF HEALTHY APB MUSCLE more, following muscle stretches, the M1-M2 was to same on September 25, 2021 by guest. Protected copyright. The influence of the type of muscle contrac- interval shown be about the in hand and foot muscles.'8 The calculated, R1-R3 interval in the hand (about 40 ms) Isometric Isotonic and in the leg (about 60 ms) makes a Contraction transcortical pathway for R3 possible also for lower limb muscles. r-I M Rl IR2 -;9iQ m | A i I Reflex behaviour in spasticity IIR2 The short-latency reflex "RI" could be l R obtained in all patients. Its latency was simi- f I l irC L1I lar to that of the unaffected side, however, its I amplitude was usually larger. This RI l r I IAT LI IT enhancement corresponds to the well-known Ii exaggerated tendon tap reflexes in spasticity. XIXll I i1I 1X For some time it has been suggested that ;200AVI I I1 T Im 200 tLVl 110Ms increased gamma-motor neuron drive is responsible for the exaggerated reflexes, how- Figure 7 An examiiple of EMG reflex response pattenis recordedfromt abductor pollicis ever, microneurographic recordings of single brevis muscle of a healthy subject durinzg isonmetric and isotonic contraction. During isotonic muscle spindle afferents in humans have pro- contractioni Rl anid R2 amplitudes increased. Notice that under isotonic condition the Rl anmplitude reaches a high value similar to sonme of those observed in spastic limbs. R2 vided no evidence of overactivity of afferent amplitude, in contrast to RIIR2 behaviour in spastic linmbs, increased. spindle discharges due to exaggerated fusimo- Reflexes in spasticity 391

tor drive in-spastic patients.29 30 The enhanced movements of limbs which has been spastic J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.4.386 on 1 April 1993. Downloaded from amplitude of monosynaptic stretch reflex in described recently.37 Correlation of the reflex spastic paresis may arise from a reduction in activity to the clinical stages of recovery has reflex threshold without significant enhance- shown a reciprocal relationship between RI ment of the reflex gain.2425 The most proba- and R2 amplitudes. From stage (I) of syner- ble cause for that may be a decreased gistic movement to stage (III) of useful move- presynaptic inhibition of Ia fibres as sug- ment, RI amplitude decreased while that of gested in other studies.3'-33 R2 increased significantly in a proportional In this work, the late reflex response R2 manner (fig 4). This is also underlined by a was either absent or reduced in amplitude significant negative linear correlation between and its onset was delayed on the spastic side. both responses in upper as well as in lower The absence of a long-latency response in limbs (fig 5). This negative correlation was spastic patients has been previously described statistically insignificant in previous work23 in APB following electrical stimulation of the probably because the correlated data were not median nerve,6 and in the triceps surae2' and normalised to the response recorded on the the wrist flexors23 following stretch. The pos- unaffected side of the patient. A similar sibility that the cortical lesion had disrupted insignificant correlation was also found in this the pathway for R2 as suggested elsewhere9 34 35 study when absolute amplitude values were is unlikely for lower limb muscles according used. The normalisation was performed in to our observations. In our study response our work to account for the wide range of patterns with an absent R2 component were reflex amplitudes observed. A reciprocal associated with large RI peaks (fig 1). A simi- modulation of RI and R2 reflexes seen here lar change to that observed in spastic patients in spastic patients during different stages of with large short-latency and small long- recovery was also described for small children latency reflex responses has also been during development where a progressive described in young children with immature decrease of short-latency reflex response (El) CNS.2636 As an explanation for the absence of was associated with the progressive increase the late reflex response, it was argued that an of longer-latency reflex response (E2) increased motor neuronal refractoriness after (evoked by cutaneous stimulation) with a large early response may have limited sub- increasing age in children.38 sequent later overactivity.2' 23 This hypothesis Our findings support the proposal of recip- is based on the observation that in normal rocal modulation of these reflex mechanisms, subjects the long-latency response decreased the function of which depends on supraspinal in amplitude as the short-latency response control. The close relationship between the became larger by increasing the acceleration stage of recovery and the behaviour of early of the stretch.'7 However, our work shows and late reflexes we describe in spastic that among healthy subjects and the unaf- patients may serve for the documentation of fected limbs there were examples of large RI the course of rehabilitation by simple electro- waves associated with large R2 responses dur- physiological means. ing isotonic contraction. Under this condition several RI amplitudes in healthy subjects and We thank Dr I Gibson for scrutinising the English text. This work was supported by the Channel System of the Egyptian the unaffected limbs were in the range of Mission Department and the Deutsche Forschungsgemein- http://jnnp.bmj.com/ those seen in spastic limbs where R2 was schaft (SFB 325). absent (figs 1 and 7). In healthy subjects and unaffected limbs, however, the amplitude of R2 was also increased. Furthermore, follow- 1 Upton ARM, McComas AJ, Sica REP. Potentiation of late response evoked in muscles during effort. J Neiinl ing muscle stretches in spastic patients it Vekirosiurg Pssvchliat' 1971 ;34:699-711. could be shown that exaggerated early 2 Cacia MR, McComas AJ, Upton ARM, Blogg T. reflexes cannot account for the depression of Cutaneous reflexes in small muscles of the hand. Jf

Neurol Neiurosusrg Psschaiatrs 1973;36:960-77. on September 25, 2021 by guest. Protected copyright. the late responses.'3 Thus motor neuronal 3 Marsden CD, Merton PA, Morton HB. Is the human stretch reflex cortical rather than spinal? Lau,cet refractoriness is unlikely to be the sole cause 1973;1:759-61. of the R2 depression in spasticity. It is 4 Merton PA. The properties of the human muscle servo. Bra,i Res 1974;71:475-8. assumed that both responses are modulated 5 Milner-Brown HS, Stein RB, Lee RG. Synchronization of reciprocally by a supraspinal mechanism. The human motor units: possible role of exercise and supraspinal reflexes. J Neiurtl N)Seitroig Psvslhiarnr 1975; dynamic RI-R2 relationship described in nor- 38:245-54. mal subjects seems to be static in patients 6 Conrad B, Aschoff JC. Effects of voluntary isometric and isotonic activity on late transcortical reflex components with spastic paresis. in normal subjects and hemiparetic patients. Elctrt)eicephlal)gr Cli,, Neirtr)ph-vsio11 977;42: 107-16. 7 Stanley EF. Reflex evoked in human thenar muscles dur- Relation between reflex responses and the stage of ing voluntary activity and their conduction pathwavs. J recovery NAei,r,ol Neitrosslrg Pss.vciatr, 1 978,41:1016-23. 8 Matthews PBC. The human stretch reflex and the motor In our study it was found that the R2 cortex. TINS 1991;14(3):87-91. response is generally observed in stage II and 9 Evarts EV. Motor cortex reflex associated with learned movement. Science 1973;179:501-3. III, that is, when the limb is at least capable 10 Lee RG, Tatton WG. Motor responses to sudden limb of performing isolated movements (fig 6). displacements in primates with specific CNS lesions and in human patients with motor svstem disorders. Ca,, _J When it appeared in poor functioning limbs it Vei?ol S i 1975;2:285-93. was usually of very low amplitude. The defec- 11 Marsden CD, Merton PA, Morton HB. Stretch reflex and R2 servo action in a variety of human muscles. J PI,ssial tive reflex may therefore contribute to dis- (Lo-,d) 1976;259:531-60. turbed motor control in spastic patients. This 12 Chan CWY, Melvill JG, Kearnev RE, Watt DGD. The late electromyographic response to limb displacement in suggestion agrees with the loss of long-latency man. I: Evidence for supraspinal contribution. Electra- reflex modulation during functional arm e?iceph,alogr CliM, Neitraphsll 1979;46: 173-8 1. 392 Ibrahim, El-Abd, Dietz

13 Houk JC. Regulation of stiffness by skeletomotor reflexes. 27 Baldissera F, Hultborn H, Illert M. Integration in spinal Annu Rev Physiol 1979;41:99-114. neuronal systems. In: Handbook of Physiology. The J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.56.4.386 on 1 April 1993. Downloaded from 14 Bawa P, Tatton WG. Motor unit responses in muscles Nervous System. Motor Control. Washington, DC: Am stretched by imposed displacements of the monkey Physiol Soc 1981:509-95. wrist. Exp Brain Res 1979;37:417-37. 28 Thilmann AF, Fellows SJ, Garms E. Pathological stretch 15 Hagbarth KE, Hagglund JV, Wallin EV, Young RR. reflexes on the "good" side of hemiparetic patients. J Grouped spindle and electromyographic responses to Neurol Neurosurg Psychiatry 1990;53:208-14. abrupt wrist extension movements in man. J7 Physiol 29 Vallbo AB, Hagbarth KE, Torebjork HE, Wallim BC. (Lond) 1981;312:81-96. Somatosensory, proprioceptive and sympathetic tract 16 Berger W, Dietz V, Quintern J. Corrective reactions to activity in human peripheral nerves. Physiol Rev 1979; stumbling in man: Neuronal coordination of bilateral 59:919-57. leg muscle activity during gait. J Physiol (Lond) 1984; 30 Burke DA. Critical examination of the case for and against 357:109-25. fusimotor involvement in disorders in muscle tone. In: 17 Berardelli A, Hallett M, Kaufman C. Stretch reflexes of Desmedt JE ed: Motor control mechanisms in health triceps surae in man. J Neurol Neurosurg Psychiatry and disease. Advances in Neurology New York: Raven 1982;45:513-25. Press 1983;39:133-50. 18 Darton K, Lippold OCJ, Shahani M, Shahani U. Long- 31 Iles JF, Robert, RC. Presynaptic inhibition of monosynap- latency spinal reflexes in humans. J Neurophysiol 1985; tic reflexes in the lower limbs of subjects with upper 53:1604-18. motoneurone diseases. Y Neurol Neurosurg Psychiatry 19 Jenner JR, Stephens JA. Cutaneous reflex responses and 1986;49:937-44. their central nervous pathways studied in man. Jf Physiol 32 Lance JW, DeGail P, Nielson PD. Tonic and Phasic (Lond) 1982;333:405-19. spinal cord mechanisms in man. J Neurol Neurosurg 20 Thilmann AF, Schwarz M, Topper R, Fellows SJ, Noth J. Psychiatry 1966;29:535-44. Different mechanisms underlie the long-latency stretch 33 Somerville J, Ashby P. Hemiplegic spasticity: Neuro- reflex responses of active human muscle at different physiological studies. Arch Phys Med Rehab 1978;59: joints. J Neurol Neurosurg Psychiatry 1991;444:631-43. 592-6. 21 Berardelli A, Hallett M, Simon SR. Stretch reflexes of tri- 34 Hammond PH, Merton PA, Sutton GC. Nervous ceps surae in patients with upper motor neurone syn- gradation of muscular control. Brit Med Bull 1956; dromes. 7 Neurol Neurosurg Psychiatry 1983;46:54-60. 12:214-18. 22 Ashby P, Burke D. Stretch reflexes in the upper limb of 35 Marsden CD, Merton PA, Morton HB. Servoaction in spastic man. J Neurol Neurosurg Psychiatry 1971 ;34: human voluntary movement. Nature (Lond) 1972;238: 765-71. 140-3. 23 Cody FWJ, Richardson HC, MacDermott N, Ferguson 36 Forssberg H. Ontogeny of human locomotor control. I. IT. Stretch and vibration reflexes of wrist flexor muscle Infant stepping, supported locomotion and transition in spasticity. Brain 1987;110:433-50. to independent locomotion. Exp Brain Res 1985;57: 24 Powers RK, Meyer JM, Rymer WZ. Quantitative relations 480-93. between hypertonia and stretch reflex threshold in 37 Dietz V, Trippel M, Berger W. Reflex activity and muscle spastic hemiparesis. Ann Neurol 1988;23: 115-24. tone during elbow movements of patients with spastic 25 Powers RK, Campbell DL, Rymer WZ. Stretch reflex paresis. Ann Neurol 1991;30:767-79. dynamics in spastic elbow flexion muscles. Ann Neurol 38 Evans AL, Harrison LM, Stephens JA. Maturation of the 1989;25:32-42. cutaneomuscular reflex recorded from the first dorsal 26 Berger W, Altenmuller E, and Dietz V. Normal and interosseous muscle in man. J Physiol (Lond) 1990; impaired development of children's gait. Hum Neurobiol 428:425-40. 1984;3:163-170. http://jnnp.bmj.com/ on September 25, 2021 by guest. Protected copyright.