Short-Latency Autogenic Idubition in Patients with Parkrnsonian Rqgdity P. J. Delwaide, J. L. Pepin, and A. Maertens de Noordhout
The spinal Ib interneuron efficacy has been compared in 11 control subjects and 19patients with parkinsonian rigidity. In normal subjects, gastrocnemius medialis nerve stimulation induces an inhibition of the soleus H reflex for 3 to 8 msec with a peak at 5 msec of 83.72 f 7.28% of the control value of H reflex. In parkinsonian patients, inhibition is reduced or even replaced by facilitation, which also peaks at 5 msec. The departures from normal values correlate with rigidity intensity assessed by the Webster scale. Increase in rigidity is associated, first, with a reduction of inhibition and, from a score of 2 or more, with facilitation replacing the normal inhibition. In addition to providing an electrophysiological index of rigidity, reduction in autogenic inhibition might be one of the neurophysiological mechanisms underlying rigidity. In association with the known hyperactivity of the Ia inhibitory interneuron in Parkinson's disease, reduction of activity of Ib interneuron could be explained by an increased activity in the reticularis gigantocellularis nucleus; its efferent tracts both inhibit Ib interneurons and activate la interneurons. Delwaide PJ, Pepin JL, Maertens de Noordhout A. Short-latency autogenic inhibition in patients with parkinsonian rigidity. Ann Neurol 1991;30:83-89
Neurophysiological mechanisms underlying parkinsonian rigidity are not well understood. Functional changes restricted to the spinal level and its reflex circuits were first advocated El}, but experimental evidence supporting that view has proved unreliable, especially after the demonstration by microneurography that Ia discharges are not increased 12). An alternative explanation has been proposed on the basis of increased M2 responses recorded in rigid subjects after abruptly stopping an ongoing movement {3]. This observation suggested hyperactivity in long loop reflex pathways originating in primary endings of neuromuscular spindles and relaying through the motor cortex. Although popular, this hypothesis is also questionable because it is not specific and does not explain many clinical observations, for instance that rigidity is not reinforced by the speed of the imposed movement. Further studies are thus needed to specify which neurophysiological mechanisms are responsible for muscle stiffness in parkinsonism. Many spinal mechanisms have been studied in rigid patients, but to date, only one has been shown to be abnormal; tendon jerks {4}, motoneuron pool excitability El, 51, presynaptic inhibition {6},and Renshaw's recurrent inhibition are within the range of agematched normal values, but reciprocal inhibition mediated through the Ia inhibitory interneuron is increased 17-31. Ib inhibition has not yet been studied, although
modifications of its function could theoretically explain rigidity. Ib fibers, originating in Golgi tendon organs, are large-diameter fibers that project onto interneurons, located at Rexed's lamina V and VI, which in turn inhibit homonymous motoneurons (autogenic inhibition). In fact, other afferents converge on the Ib interneuron, that is, Ia, cutaneous, and joint afferents. The Ib interneuron is facilitated by both the corticospinal and rubrospinal tracts but inhibited by the dorsal reticulospinal and the noradrenergic reticulospinal tracts. If inhibition brought about by Ib interneurons is reduced, the mechanisms maintaining a linear relationship between length and tension [lo} would be altered and muscle stiffness might result. Testing Ib interneuron function in humans is possible through a technique proposed by PierrotDeseilligny and colleagues Ell}, which is based on the assumption that the primary afferents originating in the gastrocnemius medialis and projecting onto the soleus motoneurons are chiefly, if not entirely, Group Ib. Although there is no direct proof that in humans the afferent projection from gastrocnemius to soleus is from Group Ib, not Ia fibers, indirect evidence indicates that the proposed test is accurate, that is, type of stimulated fibers, latency of the inhibition, modulation of activity by cutaneous afferents, differences in anatomical data between cats and primates { 123. The aim of the present study was to test the functioning of Ib
From the University Department of Neurology, HBpital de la Citadelle, Liege, Belgium.
Address correspondence to Dr Delwaide, University Department of Neurology, HBpital de la Citadelle, Bd du 12" de L g n e 1, 4000 'ege' Be'gium'
Received Oct 9, 1990, and in revised form Dec 31. Accepted for publication Jan 7, 1991.
Copyright 0 1991 by the American Neurological Association
83
Some Clinical Featares of the Patients
Webster Rating Scale Subjects
Sex
B.S. B.L. D.J. L.C. R.L. L.R. H.B.
F M M F F F F F M M M M M F M
Age (yr)
Duration Stage" (yr) I I11 Ill 111 IV I11 I1 I I 11 I11 V 111 I1
4 2
Bradykinesia
1
2 0
1 0 1
No Yes No Yes
F F
70
111
5
Y.L.
F
70
IV
20
2
2
R.E.
M
73
v
17
3
1
C.J.
D.A. H.M.
2
0 1 2
4 8 3
2 2 2
3
0 2
1
1
Yes
1
1
3
1
0
1
2 0
1
0 0 2 3 2 1
2
No NO Yes
3 2 10 8
5
1
1 0
2 1 0
0
2
I11
8 10
1
0
1
0
I11
No
Tremor
M.E. L.E.
H.Y. L.L.
of Motor Performance Treatment
Rigidity
58 60 60 60 61 62 63 64 64 65 65 65 66 67 68 70
S.S. B.M. S.L.
Fluctuations
Cogwheel Phenomenon
1 0 1 1 2
0 1 2
3 3 1
Yes
No No
No Yes No
L-Dopa None None r-Dopa L-Dopa L-Dopa
+ bromocriptine
+ anticholinergics + anticholinergics
No
Anticholinergics r-Dopa Anticholinergics Amantadine None L-Dopa + bromocriptine r-Dopa L-Dopa
No
No
Anticholinergics
Yes
No
Yes No
L-Dopa L-Dopa
Yes
Yes
d dopa + bromocriprine
Yes
Yes
L-Dopa
No NO No Yes Yes
No
No No No No Yes
No
+ bromocriptine + amantadine
aHoehn & Yahr Scale. interneurons in rigid parkinsonian patients using Pierrot-Deseilligny's technique and to look for possible involvement of that mechanism in the pathophysiology of rigidity.
Subjects and Methods Eleven control subjects and 19 parkinsonian patients were studied. All gave informed oral consent for the experimental procedure that was approved by the local ethics committee. The control subjects (6 men, 5 women; mean age, 64.6 2 7.6 years; range, 54-79 years) were hospitalized for nonneurological diseases and did not receive any psychotropic drugs. The 19 parkinsonian patients were admitted for adjustment of medication so that good control of medication was obtained. All experiments described herein were performed in the morning before the first daily dose of antiparkinsonian drugs. The Table lists the main characteristics (namely, items 1,2, and 6 of the Webster Rating Scale) of the patients at the time of the electrophysiological test. Sixteen patients were treated with L-dopa plus carbidopa; some also took bromocriptine or anticholinergics. Three were studied before any treatment was begun, and the test procedure was repeated after they reached efficacious drug levels. Subjects were seated comfortably in an armchair especially designed for reflex studies. Experiments were conducted in a silent, constant-temperature (20°C) room. The leg under study was rigidly fixed with the knee flexed at 120 degrees and the ankle at 100 degrees. The technique for testing Ib inhibition was described in detail by Pierrot-Deseilligny and colleagues [ I l l , and we followed those procedures. A brief summary follows. The full
84 Annals of Neurology Vol 30 No 1 July 1991
procedure took less than 1 hour. Electrical reflex responses of soleus were recorded through Ag-AgCl surface electrodes fixed over the Achilles tendon and 3 cm proximally after gentle skin abrasion. The ground electrode was placed over the calf. After amplification and filtering (bandpass filter between 80 Hz and 2.5 kHz), the signals were digitized (sampling frequency, 5 kHz). The test response, the soleus H reflex, was evoked by stimulation of the posterior tibial nerve through Disa bipolar cutaneous electrodes (Disa Electronic, Copenhagen, Denmark) fixed over the nerve at the upper part of the popliteal fossa. A single rectangular 1-msec shock was given at an intensity such that the amplitude of the reflex corresponded to 15% of the soleus maximum motor response (M response). Stimulation was repeated randomly at intervals longer than 7 seconds. The test response was conditioned by a I-msec electrical shock applied through Disa bipolar cutaneous electrodes fixed over the gastrocnemius medialis (GM) nerve at the lower part of the popliteal fossa, 6 to 8 cm distal to electrodes stimulating the posterior tibial nerve. Intensity was adjusted to 98% of the motor threshold for the GM nerve. It has been systematically verified that intense GM nerve stimulation (four times motor threshold) does not elicit a response in soleus and was thus selective for the GM nerve. Twenty unconditioned H reflexes were averaged as basal values and compared with 10 conditioned responses obtained every millisecond at increasing delays from 2 to 10 msec (GM nerve stimulation preceding posterior tibial nerve stimulation). At the end of the sequence of conditioned responses, 20 unconditioned responses were again averaged to verify stability compared with the initial values. In case of more than 10% variation, the whole sequence was rejected.
0
2
4
6
8
]Oms
Fig I . Autogenic inhibition in normal subjects and parkinsonian patients. (A)An example of amplitude changes of the soleus H rejex in a normal subject expressed in terms of interstimuli intervals. Inhibition is seen when conditioning Jtimulation precedes Sy 5 msec the stimulation evoking the conditioned response. At 8 m e c , inhibition has disappeared. (B) An example in 1 parkinsonian patient (D.J.) of amplitude changes of the sokus H reflex in twms of interstimuli intervals. Contvaqi t o A, facilitation is Jeen at an interstimuli interval of 5 mec. At 8 msec. facilitation has disappeared. (C) Mean curve obtained from 11 normal subjects (interrupted line) and 19 parkinsonian patients (continuous line) indicating the time course of soleus H reflex changes afev conditioning by gastrocnemius medialis n m e stimulation. In ordinate, changej in soleus H reflex amplitude expressed in percentage of control values (c. H rejex). In abscissa, interstimuli intervals in milliseconds. Vertical bars, standzrd error oftbe mean (SEMI. In n o m l subjects, clear inhibition is seen, which is replaced in pavkinsonian patients by moderate facilitation.
Changes in amplitude of the H reflexes were expressed as percentages of control values and plotted in terms of delays separating conditioning and test stimulations. The H reflex was conditioned in all subjects when soleus was completely relaxed, which was verified by continuous auditory feedback monitoring of electromyographic (EMG) activity. In 5 control subjects, however, a slight voluntary tonic contraction (10% of maximal muscle strength) was requested in the soleus and tibialis anterior alternately, and the conditioning of H reflexes was repeated under these conditions. Significance of differences between control subjects and parkinsonian patients was established by analysis of variance. In parkinsonian patients, differences between subgroups were established by the Wilcoxon test. Possible correlations between electrophysiological results and clinical status were sought by analysis of linear regression.
Results Control Stlbjects Figure 1A illustrates soleus H reflex amplitude reduction in a control subject when tibialis posterior nerve stimulation follows 5 msec after gastrocnemius medi-
alis nerve stimulation. In the illustrated subject, intensity of the conditioning stimulation was 10 mA, that is, 98% of the GM nerve motor activation threshold. Under basal conditions, soleus H reflex amplitude was 6.1 mV that, when conditioned, dropped to 4.5 mV or 74% of its control value. In Figure 1C is represented the mean curve obtained in control subjects of H reflex amplitude changes in terms of delays separating GM and tibialis posterior nerve stimulations. Amplitude was clearly reduced from 3 to 8 msec with maximum reduction at 5 msec. At that delay, H reflex amplitude was reduced to 83.72% (SD, +7.28%) of its control value. The time course and magnitude of the effects are similar to those previously described by PierrotDeseilligny and colleagues { 111 and attributed to Ib inhibition. Similar curves have been obtained from all control subjects. The mean intensity of the conditioning electrical shock given to the gastrocnemius medialis nerve was 11.8 mA (SD, k l . 2 mA). In all subjects, the background EMG activity was monitored for the duration of the experiments to ensure that no active contraction was influencing the results. Moreover, to rule out possible interference by undetected muscle activity, this influence was studied in 5 subjects. The experimental procedure was applied during slight tonic voluntary contractions of soleus and tibiahs anterior, respectively. During soleus contraction, the test H reflex is increased to 172% (SD, *34%) of its control value and the conditioning shock reduces its amplitude. The ratio between conditioned and test H reflex amplitudes is slightly but not significantly increased to 86.3% (SD, k7.896) compared with that obtained at rest in this series of 5 subjects (84.3%; SD, t7.8%). When tibialis anterior is contracted slightly, the test H reflex is reduced to 38% (SD, 2 1 4 % ) of its value at rest and the ratio between conditioned and test reflex amplitude (80.1%; SD, 27.6%) indicates that inhibition is discretely but not significantly enhanced, and not reduced.
Parkinsonian Patients Figure 1B exemplifies the results obtained with the same technique in a parkinsonian patient (D.J.). Soleus H reflex amplitude was 5.9 mV when unconditioned. When the test stimulation was preceded ( 5 msec) by stimulation of the GM nerve at an intensity of 11 mA (which corresponds to 98% of the threshold for GM nerve motor fibers), the amplitude of soleus H reflex rose to 6.4 mV, that is, 108% of control value. The curve plotted at different delays from 2 to 10 msec does not reveal any inhibition. Figure 1C shows the mean curve obtained at various delays for the 19 parhnsonian patients. At 4- to 8-msec delay, the mean H reflex amplitude is increased. The peak value is seen after 5 msec and reaches 107.5% (SD, 216.01%) of Delwaide et al: IB Inhibition in PD Rigidity 85
control values. The amplitude differences between parkinsonian patients and control subjects are statistically significant at 4, .5 ( p < 0.001) and 6 to 8 msec ( p < 0.02). The mean intensity of the GM nerve stimulation in parkinsonian patients was not statistically different from that used in control subjects (10.6 mA; SD, 21.4 mA). The results were analyzed in terms of rigidity intensity expressed by the Webster scale. Four subgroups were formed. Mean age was 63.8 years for Group 0 (with no rigidity), 66.0 years for Group 1 (mild rigidity), 63.4 years for Group 2 (moderate rigidity), and 69.0 years for Group 3 (marked rigidity). Mean curves obtained for the four groups differ. The amplitude of the conditioned H reflex was reduced in Group 0 without detectable rigidity. Reduction followed a time course similar to that observed in control subjects reaching a maximum after a delay of 5 msec. At that point, the mean value corresponded to 89.25% (SD, ?8.3%) of control values. There was no statistical difference with the mean value obtained in normal subjects. In Group 1 (mild rigidity), the mean amplitude of conditioned values was not reduced and peaked at 5 msec, corresponding to 105.6% (SD, 28.03%) of control values. In Group 2 (moderate rigidity), the mean amplitude of conditioned H reflexes at 5 msec reached 115.92% (SD, *12.86%) and, in Group 3 (severe rigidity), 129.68% (SD, k7.4%) of control values. At a delay of 5 msec, differences in H reflex amplitudes are statistically significant between Groups 0 and 1 ( p < 0.001) and Groups 1 and 2 ( p < 0.001). Between Groups 2 and 3, the differences are not significant. Correlation between rigidity intensity and conditioned H reflex amplitudes at 5-msec delay was sought (Fig 2A) and is linear with r = 0.849 ( p < 0.001). With tremor, the correlation coefficient is r = 0.195 ( p = 0.423) and with bradykinesia, r = 0.247 ( p = 0.308). In 3 de novo patients (H.Y., D.J., and B.L.), the experimental procedure was repeated three times, as follows: first, while patients were free of drug and were moderately rigid; second, when patients were treated either with an anticholinergic drug (5- or 10-mg of procyclidine hydrochloride) or by a single dose of 500 mg of L-dopa plus carbidopa and still exhibiting mild rigidity; third, when patients were no longer rigid due to daily treatment with 15 mg of procyclidine or 500 mg of L-dopa (plus carbidopa). Figure 2B represents changes in amplitude of the conditioned reflex in these patients. When untreated with a rigidity score of 2 , H.Y. had an H reflex facilitation of 120% of its control values at 5 msec. When treated with 3 x 5 mg of procyclidine, rigidity was not detectable and H reflex amplitude at 5 msec after conditioning reached 82% of its control value. Because treatment with that drug 86 Annals of Neurology Vol 30 No 1 July 1991
A
%
0
% 120r
140r
II
c. 1 3 0 -
n
* 1 c
10090
,-I
110-
il
J
I ,I
' ',"
100-
l
-
SOL
.-' ,,/,'
G.
, 0
J
1
2
Riiidily
3
SOL
The spinal Ib interneuron efficacy has been compared in 11 control subjects and 19patients with parkinsonian rigidity. In normal subjects, gastrocnemius medialis nerve stimulation induces an inhibition of the soleus H reflex for 3 to 8 msec with a peak at 5 msec of 83.72 f 7.28% of the control value of H reflex. In parkinsonian patients, inhibition is reduced or even replaced by facilitation, which also peaks at 5 msec. The departures from normal values correlate with rigidity intensity assessed by the Webster scale. Increase in rigidity is associated, first, with a reduction of inhibition and, from a score of 2 or more, with facilitation replacing the normal inhibition. In addition to providing an electrophysiological index of rigidity, reduction in autogenic inhibition might be one of the neurophysiological mechanisms underlying rigidity. In association with the known hyperactivity of the Ia inhibitory interneuron in Parkinson's disease, reduction of activity of Ib interneuron could be explained by an increased activity in the reticularis gigantocellularis nucleus; its efferent tracts both inhibit Ib interneurons and activate la interneurons. Delwaide PJ, Pepin JL, Maertens de Noordhout A. Short-latency autogenic inhibition in patients with parkinsonian rigidity. Ann Neurol 1991;30:83-89
Neurophysiological mechanisms underlying parkinsonian rigidity are not well understood. Functional changes restricted to the spinal level and its reflex circuits were first advocated El}, but experimental evidence supporting that view has proved unreliable, especially after the demonstration by microneurography that Ia discharges are not increased 12). An alternative explanation has been proposed on the basis of increased M2 responses recorded in rigid subjects after abruptly stopping an ongoing movement {3]. This observation suggested hyperactivity in long loop reflex pathways originating in primary endings of neuromuscular spindles and relaying through the motor cortex. Although popular, this hypothesis is also questionable because it is not specific and does not explain many clinical observations, for instance that rigidity is not reinforced by the speed of the imposed movement. Further studies are thus needed to specify which neurophysiological mechanisms are responsible for muscle stiffness in parkinsonism. Many spinal mechanisms have been studied in rigid patients, but to date, only one has been shown to be abnormal; tendon jerks {4}, motoneuron pool excitability El, 51, presynaptic inhibition {6},and Renshaw's recurrent inhibition are within the range of agematched normal values, but reciprocal inhibition mediated through the Ia inhibitory interneuron is increased 17-31. Ib inhibition has not yet been studied, although
modifications of its function could theoretically explain rigidity. Ib fibers, originating in Golgi tendon organs, are large-diameter fibers that project onto interneurons, located at Rexed's lamina V and VI, which in turn inhibit homonymous motoneurons (autogenic inhibition). In fact, other afferents converge on the Ib interneuron, that is, Ia, cutaneous, and joint afferents. The Ib interneuron is facilitated by both the corticospinal and rubrospinal tracts but inhibited by the dorsal reticulospinal and the noradrenergic reticulospinal tracts. If inhibition brought about by Ib interneurons is reduced, the mechanisms maintaining a linear relationship between length and tension [lo} would be altered and muscle stiffness might result. Testing Ib interneuron function in humans is possible through a technique proposed by PierrotDeseilligny and colleagues Ell}, which is based on the assumption that the primary afferents originating in the gastrocnemius medialis and projecting onto the soleus motoneurons are chiefly, if not entirely, Group Ib. Although there is no direct proof that in humans the afferent projection from gastrocnemius to soleus is from Group Ib, not Ia fibers, indirect evidence indicates that the proposed test is accurate, that is, type of stimulated fibers, latency of the inhibition, modulation of activity by cutaneous afferents, differences in anatomical data between cats and primates { 123. The aim of the present study was to test the functioning of Ib
From the University Department of Neurology, HBpital de la Citadelle, Liege, Belgium.
Address correspondence to Dr Delwaide, University Department of Neurology, HBpital de la Citadelle, Bd du 12" de L g n e 1, 4000 'ege' Be'gium'
Received Oct 9, 1990, and in revised form Dec 31. Accepted for publication Jan 7, 1991.
Copyright 0 1991 by the American Neurological Association
83
Some Clinical Featares of the Patients
Webster Rating Scale Subjects
Sex
B.S. B.L. D.J. L.C. R.L. L.R. H.B.
F M M F F F F F M M M M M F M
Age (yr)
Duration Stage" (yr) I I11 Ill 111 IV I11 I1 I I 11 I11 V 111 I1
4 2
Bradykinesia
1
2 0
1 0 1
No Yes No Yes
F F
70
111
5
Y.L.
F
70
IV
20
2
2
R.E.
M
73
v
17
3
1
C.J.
D.A. H.M.
2
0 1 2
4 8 3
2 2 2
3
0 2
1
1
Yes
1
1
3
1
0
1
2 0
1
0 0 2 3 2 1
2
No NO Yes
3 2 10 8
5
1
1 0
2 1 0
0
2
I11
8 10
1
0
1
0
I11
No
Tremor
M.E. L.E.
H.Y. L.L.
of Motor Performance Treatment
Rigidity
58 60 60 60 61 62 63 64 64 65 65 65 66 67 68 70
S.S. B.M. S.L.
Fluctuations
Cogwheel Phenomenon
1 0 1 1 2
0 1 2
3 3 1
Yes
No No
No Yes No
L-Dopa None None r-Dopa L-Dopa L-Dopa
+ bromocriptine
+ anticholinergics + anticholinergics
No
Anticholinergics r-Dopa Anticholinergics Amantadine None L-Dopa + bromocriptine r-Dopa L-Dopa
No
No
Anticholinergics
Yes
No
Yes No
L-Dopa L-Dopa
Yes
Yes
d dopa + bromocriprine
Yes
Yes
L-Dopa
No NO No Yes Yes
No
No No No No Yes
No
+ bromocriptine + amantadine
aHoehn & Yahr Scale. interneurons in rigid parkinsonian patients using Pierrot-Deseilligny's technique and to look for possible involvement of that mechanism in the pathophysiology of rigidity.
Subjects and Methods Eleven control subjects and 19 parkinsonian patients were studied. All gave informed oral consent for the experimental procedure that was approved by the local ethics committee. The control subjects (6 men, 5 women; mean age, 64.6 2 7.6 years; range, 54-79 years) were hospitalized for nonneurological diseases and did not receive any psychotropic drugs. The 19 parkinsonian patients were admitted for adjustment of medication so that good control of medication was obtained. All experiments described herein were performed in the morning before the first daily dose of antiparkinsonian drugs. The Table lists the main characteristics (namely, items 1,2, and 6 of the Webster Rating Scale) of the patients at the time of the electrophysiological test. Sixteen patients were treated with L-dopa plus carbidopa; some also took bromocriptine or anticholinergics. Three were studied before any treatment was begun, and the test procedure was repeated after they reached efficacious drug levels. Subjects were seated comfortably in an armchair especially designed for reflex studies. Experiments were conducted in a silent, constant-temperature (20°C) room. The leg under study was rigidly fixed with the knee flexed at 120 degrees and the ankle at 100 degrees. The technique for testing Ib inhibition was described in detail by Pierrot-Deseilligny and colleagues [ I l l , and we followed those procedures. A brief summary follows. The full
84 Annals of Neurology Vol 30 No 1 July 1991
procedure took less than 1 hour. Electrical reflex responses of soleus were recorded through Ag-AgCl surface electrodes fixed over the Achilles tendon and 3 cm proximally after gentle skin abrasion. The ground electrode was placed over the calf. After amplification and filtering (bandpass filter between 80 Hz and 2.5 kHz), the signals were digitized (sampling frequency, 5 kHz). The test response, the soleus H reflex, was evoked by stimulation of the posterior tibial nerve through Disa bipolar cutaneous electrodes (Disa Electronic, Copenhagen, Denmark) fixed over the nerve at the upper part of the popliteal fossa. A single rectangular 1-msec shock was given at an intensity such that the amplitude of the reflex corresponded to 15% of the soleus maximum motor response (M response). Stimulation was repeated randomly at intervals longer than 7 seconds. The test response was conditioned by a I-msec electrical shock applied through Disa bipolar cutaneous electrodes fixed over the gastrocnemius medialis (GM) nerve at the lower part of the popliteal fossa, 6 to 8 cm distal to electrodes stimulating the posterior tibial nerve. Intensity was adjusted to 98% of the motor threshold for the GM nerve. It has been systematically verified that intense GM nerve stimulation (four times motor threshold) does not elicit a response in soleus and was thus selective for the GM nerve. Twenty unconditioned H reflexes were averaged as basal values and compared with 10 conditioned responses obtained every millisecond at increasing delays from 2 to 10 msec (GM nerve stimulation preceding posterior tibial nerve stimulation). At the end of the sequence of conditioned responses, 20 unconditioned responses were again averaged to verify stability compared with the initial values. In case of more than 10% variation, the whole sequence was rejected.
0
2
4
6
8
]Oms
Fig I . Autogenic inhibition in normal subjects and parkinsonian patients. (A)An example of amplitude changes of the soleus H rejex in a normal subject expressed in terms of interstimuli intervals. Inhibition is seen when conditioning Jtimulation precedes Sy 5 msec the stimulation evoking the conditioned response. At 8 m e c , inhibition has disappeared. (B) An example in 1 parkinsonian patient (D.J.) of amplitude changes of the sokus H reflex in twms of interstimuli intervals. Contvaqi t o A, facilitation is Jeen at an interstimuli interval of 5 mec. At 8 msec. facilitation has disappeared. (C) Mean curve obtained from 11 normal subjects (interrupted line) and 19 parkinsonian patients (continuous line) indicating the time course of soleus H reflex changes afev conditioning by gastrocnemius medialis n m e stimulation. In ordinate, changej in soleus H reflex amplitude expressed in percentage of control values (c. H rejex). In abscissa, interstimuli intervals in milliseconds. Vertical bars, standzrd error oftbe mean (SEMI. In n o m l subjects, clear inhibition is seen, which is replaced in pavkinsonian patients by moderate facilitation.
Changes in amplitude of the H reflexes were expressed as percentages of control values and plotted in terms of delays separating conditioning and test stimulations. The H reflex was conditioned in all subjects when soleus was completely relaxed, which was verified by continuous auditory feedback monitoring of electromyographic (EMG) activity. In 5 control subjects, however, a slight voluntary tonic contraction (10% of maximal muscle strength) was requested in the soleus and tibialis anterior alternately, and the conditioning of H reflexes was repeated under these conditions. Significance of differences between control subjects and parkinsonian patients was established by analysis of variance. In parkinsonian patients, differences between subgroups were established by the Wilcoxon test. Possible correlations between electrophysiological results and clinical status were sought by analysis of linear regression.
Results Control Stlbjects Figure 1A illustrates soleus H reflex amplitude reduction in a control subject when tibialis posterior nerve stimulation follows 5 msec after gastrocnemius medi-
alis nerve stimulation. In the illustrated subject, intensity of the conditioning stimulation was 10 mA, that is, 98% of the GM nerve motor activation threshold. Under basal conditions, soleus H reflex amplitude was 6.1 mV that, when conditioned, dropped to 4.5 mV or 74% of its control value. In Figure 1C is represented the mean curve obtained in control subjects of H reflex amplitude changes in terms of delays separating GM and tibialis posterior nerve stimulations. Amplitude was clearly reduced from 3 to 8 msec with maximum reduction at 5 msec. At that delay, H reflex amplitude was reduced to 83.72% (SD, +7.28%) of its control value. The time course and magnitude of the effects are similar to those previously described by PierrotDeseilligny and colleagues { 111 and attributed to Ib inhibition. Similar curves have been obtained from all control subjects. The mean intensity of the conditioning electrical shock given to the gastrocnemius medialis nerve was 11.8 mA (SD, k l . 2 mA). In all subjects, the background EMG activity was monitored for the duration of the experiments to ensure that no active contraction was influencing the results. Moreover, to rule out possible interference by undetected muscle activity, this influence was studied in 5 subjects. The experimental procedure was applied during slight tonic voluntary contractions of soleus and tibiahs anterior, respectively. During soleus contraction, the test H reflex is increased to 172% (SD, *34%) of its control value and the conditioning shock reduces its amplitude. The ratio between conditioned and test H reflex amplitudes is slightly but not significantly increased to 86.3% (SD, k7.896) compared with that obtained at rest in this series of 5 subjects (84.3%; SD, t7.8%). When tibialis anterior is contracted slightly, the test H reflex is reduced to 38% (SD, 2 1 4 % ) of its value at rest and the ratio between conditioned and test reflex amplitude (80.1%; SD, 27.6%) indicates that inhibition is discretely but not significantly enhanced, and not reduced.
Parkinsonian Patients Figure 1B exemplifies the results obtained with the same technique in a parkinsonian patient (D.J.). Soleus H reflex amplitude was 5.9 mV when unconditioned. When the test stimulation was preceded ( 5 msec) by stimulation of the GM nerve at an intensity of 11 mA (which corresponds to 98% of the threshold for GM nerve motor fibers), the amplitude of soleus H reflex rose to 6.4 mV, that is, 108% of control value. The curve plotted at different delays from 2 to 10 msec does not reveal any inhibition. Figure 1C shows the mean curve obtained at various delays for the 19 parhnsonian patients. At 4- to 8-msec delay, the mean H reflex amplitude is increased. The peak value is seen after 5 msec and reaches 107.5% (SD, 216.01%) of Delwaide et al: IB Inhibition in PD Rigidity 85
control values. The amplitude differences between parkinsonian patients and control subjects are statistically significant at 4, .5 ( p < 0.001) and 6 to 8 msec ( p < 0.02). The mean intensity of the GM nerve stimulation in parkinsonian patients was not statistically different from that used in control subjects (10.6 mA; SD, 21.4 mA). The results were analyzed in terms of rigidity intensity expressed by the Webster scale. Four subgroups were formed. Mean age was 63.8 years for Group 0 (with no rigidity), 66.0 years for Group 1 (mild rigidity), 63.4 years for Group 2 (moderate rigidity), and 69.0 years for Group 3 (marked rigidity). Mean curves obtained for the four groups differ. The amplitude of the conditioned H reflex was reduced in Group 0 without detectable rigidity. Reduction followed a time course similar to that observed in control subjects reaching a maximum after a delay of 5 msec. At that point, the mean value corresponded to 89.25% (SD, ?8.3%) of control values. There was no statistical difference with the mean value obtained in normal subjects. In Group 1 (mild rigidity), the mean amplitude of conditioned values was not reduced and peaked at 5 msec, corresponding to 105.6% (SD, 28.03%) of control values. In Group 2 (moderate rigidity), the mean amplitude of conditioned H reflexes at 5 msec reached 115.92% (SD, *12.86%) and, in Group 3 (severe rigidity), 129.68% (SD, k7.4%) of control values. At a delay of 5 msec, differences in H reflex amplitudes are statistically significant between Groups 0 and 1 ( p < 0.001) and Groups 1 and 2 ( p < 0.001). Between Groups 2 and 3, the differences are not significant. Correlation between rigidity intensity and conditioned H reflex amplitudes at 5-msec delay was sought (Fig 2A) and is linear with r = 0.849 ( p < 0.001). With tremor, the correlation coefficient is r = 0.195 ( p = 0.423) and with bradykinesia, r = 0.247 ( p = 0.308). In 3 de novo patients (H.Y., D.J., and B.L.), the experimental procedure was repeated three times, as follows: first, while patients were free of drug and were moderately rigid; second, when patients were treated either with an anticholinergic drug (5- or 10-mg of procyclidine hydrochloride) or by a single dose of 500 mg of L-dopa plus carbidopa and still exhibiting mild rigidity; third, when patients were no longer rigid due to daily treatment with 15 mg of procyclidine or 500 mg of L-dopa (plus carbidopa). Figure 2B represents changes in amplitude of the conditioned reflex in these patients. When untreated with a rigidity score of 2 , H.Y. had an H reflex facilitation of 120% of its control values at 5 msec. When treated with 3 x 5 mg of procyclidine, rigidity was not detectable and H reflex amplitude at 5 msec after conditioning reached 82% of its control value. Because treatment with that drug 86 Annals of Neurology Vol 30 No 1 July 1991
A
%
0
% 120r
140r
II
c. 1 3 0 -
n
* 1 c
10090
,-I
110-
il
J
I ,I
' ',"
100-
l
-
SOL
.-' ,,/,'
G.
, 0
J
1
2
Riiidily
3
SOL
