H reflex behavior in Parkinson's disease patients and patients with extrapyramidal and pyramidal signs combined.

The H reflex was elicited in 21 normal subjects, 48 patients with Parkinson's disease (PD) and 22 patients with pyramidal and extrapyramidal signs combined (PESC). In most normal subjects (90.5%), in 29.2% of PD patients and 54.5% of patient with PESC the threshold for sensory fibers was lower than for motor fibers, and the H reflex was obtained before the M response for all duration stimuli in both legs. In 9.5% of normal subjects, 39.6% of PD patients with mild and moderate rigidity (according to the motor part of UPDRS) and 31.8% of patients with PESC, the threshold for the H reflex and M response was the same or the M response threshold was lower in at least one of the legs for short stimulus duration (0.1-0.2 ms). In 31.2% of PD patients (most of them with severe rigidity) and 13.7% of patients with PESC, the threshold for M response was lower for all stimulus duration in at least one of the legs, and it was obtained before H reflex.These very significant differences in behavior of the H reflex in PD patients (Fisher exact test, p<0.0001) that almost disappear in patients with PESC, could be possibly explained by changes in agonist-antagonist inhibition.     

H reflex threshold in Parkinson's disease patients for different stimulus duration

Electrical stimulus, with duration starting at 0.1 ms and gradually increased to 1.0 ms, was used for eliciting the H reflex in 14 normal subjects and 19 patients with Parkinson's disease (PD). In 71.1% of normal subjects and in 13.2% of PD patients the H reflex to M response threshold ratio (H/M TR) was <1 and the H reflex was obtained before the M response for all duration stimuli. For all stimulus durations a significant difference between the H/M TR in normal subjects and PD patients was found (t test 0.002-0.007). The duration effect was found to be highly significant-H/M TR for short stimulus duration was greater than for long stimulus durations (p<0.001). The optimal stimulus duration for evaluating H reflex behavior in PD patient was 0.2 ms.These very significant differences in behavior of the H reflex in PD patients could be used as another parameter in the assessment of extrapyramidal rigidity in PD patients.     

Stretch reflex of quadriceps femoris and its relation to rigidity in Parkinson''s disease

Much evidence suggests that parkinsonian rigidity is due to hyperactivity of a reflex arc. While tendon jerk and H reflex are not modified in Parkinson's disease (PD), the long-latency component of stretch reflex (LLR) shows an increased size in PD. It has been proposed that this modification could account for rigidity. We studied in 14 PD patients and 8 normal subjects the stretch reflex of the quadriceps femoris. The muscle was stretched by a torque motor in two experimental sets: at rest and with voluntary background activity. Latency, duration and size of the reflex were compared in two groups; correlation between size of the reflex and rigidity was investigated. A lower threshold for the reflex was found in PD patients in trials at rest, and LLR showed increased size and duration in trials with background activity. No clear relationships between these data and rigidity were demonstrated.     

Paradoxical modulation of tendon tap reflex during voluntary contraction in Parkinson's disease.

OBJECTIVE: Inadequate supraspinal modulation of spinal motor control mechanisms such as alpha-gamma coactivation is supposed to cause difficulty in maintaining proper voluntary contraction in Parkinson's disease (PD). METHODS: Subjects were 42 patients with PD and 20 normal volunteers. Soleus H-reflex and tendon tap reflex (T-reflex) were recorded. The maximal reflexes (H(max) and T(max)) at rest were recorded first. Next, the stimulus intensities were fixed to obtain a reflex size of around 25% of M(max) at rest for both H- and T-reflexes, and the reflexes were recorded at rest, during tonic plantarflexion (TPF), and at the onset of plantarflexion. RESULTS: H(max) at rest was 55% and T(max) 30% in normal subjects, while they were 36 and 31%, respectively, in PD. The size ratio of T(max) and H(max) at rest in PD was larger than normal. In PD, the size of H-reflex increased with TPF as in normal subjects, but T-reflex decreased. These changes in T-reflex were correlated with the grade of rigidity, bradykinesia, and time for 10 m gait. H-reflex had no such correlations. CONCLUSIONS: T-reflex was abnormally modulated in PD especially during tonic contraction. SIGNIFICANCE: Inappropriate supraspinal modulation of the spinal reflex pathways disturbs motor performance in PD.     

Lateralized alpha-motoneuron excitabilities during lying and standing of healthy individuals in relation to Parkinsonian rigidity

Abstract

Objectives: To elucidate mechanisms of Parkinsonian rigidity by assessing excitability of alpha-motoneurons innervating right and left soleus muscles in healthy controls and Parkinson’s disease (PD) patients with rigidities in the right, left and both legs.

Methods: One group of 45 controls was recruited and 60 PD patients in three groups: rigidities, predominantly in the right, left and both legs. H-reflex (H) and muscle response (M) were recorded from right and left soleus muscles during stimulations of the posterior tibial nerve at the popliteal fossa while lying and standing. The H/M ratio was taken as an index for motoneuron excitability.

Results: Mean H/M ratios were significantly different on the right and left sides, modified by postural changes in controls and PD patients. Analysis of variance showed that in healthy subjects the H/M ratio was: standing>lying (right), lying>standing (left). In right leg rigidity patients, the H/M ratio was greatest during standing, and smallest during lying. In left leg rigidity patients, the H/M ratios on the right and left sides were equally independent of posture. In controls, left H/M>right while lying, <right while standing. In right leg rigidity patients, right H/M>left, but <right in left leg rigidity patients, independent of posture. There was no side difference in patients with rigidity in both legs.

Conclusions: (i) motoneuron excitability may show side and postural differences in healthy individuals and PD patients; (ii) posture may be associated with lateralized motoneuron excitability in these subjects; and (iii) Parkinsonian rigidity may have spinal motor origins.     

Neurophysiological changes following traumatic spinal lesions in man.

Neurophysiological observations were made on normal subjects and on 57 patients who had had injuries to the spinal cord. The amplitude of the muscle compound action potential (M response) recorded from triceps surae in response to supramaximal stimulation of the tibial nerve was reduced in the patients indicating that there are changes in motor units below the level of a spinal lesion in man. In the patients who were clinically spastic it was found that: (1) The proportion of the triceps surae motoneuron pool reflexly activated either by tapping the Achilles tendon or by stimulating the tibial nerve just below the threshold of the alpha motoneuron axons (H reflex) was greater than in normal subjects. This can be explained by an increase in the excitability of central reflex pathways. (2) Vibration of the tendo Achilles depressed the H reflex less effectively than in normal subjects. This may indicate altered transmission in the premotoneuronal portion of the H reflex pathway. (3) The H reflex elicited 50 and 100 ms after a standardised conditioning stimulus to the tibial nerve and expressed as percentage of the unconditioned reflex was greater than in normal subjects. This could reflect a change in the excitability of motoneurons or of interneurons.     

Subthalamic nucleus stimulation restores corticospinal facilitation in Parkinson's disease

We have previously shown that in patients with Parkinson's disease (PD), high‐frequency stimulation (HFS) of the subthalamic nucleus (STN) modifies spinal excitability via subcortical reticulospinal routes. To investigate whether STN‐HFS also modifies spinal excitability via transcortical routes in PD, 10 patients with PD (9 men, 1 woman; 58.3 ± 8.3 years) were investigated in the medical OFF‐state with or without STN‐HFS. The H‐reflex of the right soleus muscle was recorded during slight plantar flexion at 20% of maximum force. A conditioning transcranial stimulus was applied at 95% of active motor threshold to the contralateral primary motor leg area (M1) 0–5 ms after eliciting the H‐reflex. The same paradigm was applied to 8 healthy individuals (5 men, 3 women; 50.8 ± 3.0 years). Transcranial magnetic stimulation (TMS) facilitated the H‐reflex amplitude in healthy controls. A facilitatory effect of the corticospinal input on the H‐reflex was also found in patients with PD, but only with STN‐HFS switched on. When STN‐HFS was discontinued, the H‐reflex was no longer facilitated by the TMS pulse. Accordingly, analysis of variance showed a main effect of stimulation (F = 11.15; P = 0.005), ISI (F = 6.1; P = 0.003), and an interaction between stimulation and group (PD vs. control) (F = 8.9; P = 0.01). STN‐HFS restores the normal facilitatory drive of a transcranially evoked motor cortical response to the spinal motoneuron pool. In addition to subcortical routes, STN‐DBS also alters spinal excitability via transcortical pathways. © 2008 Movement Disorder Society.     

Reflex studies and MRI in the restless legs syndrome

Brainstem and spinal pathways of untreated patients with idiopathic restless legs syndrome (RLS) were examined using magnetic resonance imaging (MRI), blink reflex, first and second exteroceptive suppression (ES1, ES2) of temporalis muscle, and H reflex. MRI of 25 patients elicited no structural lesions beyond age-related atrophy or white matter lesions on proton density- and T2-weighted coronal and axial images. All patients showed a normal latency of the soleus H reflex (mean·SD latency=31.22·2.81 ms) and the H/M ratio was 48·17%. The duration and onset latency of the direct and indirect blink reflex responses were normal in all patients compared with those of controls (p>0.5). There was no significant difference in ES1 and ES2 latencies or duration between patients and controls (p>0.5). These results suggest that the etiology of RLS symptoms does not involve structural lesions.     

Standardisation of the electrical elicitation of the human flexor reflex.

The threshold and latency of the human flexor reflex were recorded by different kinds of electrical stimuli in order to find the optimal stimulus, defined as the lowest amount of currency and the shortest possible duration. Stimulation was given over the posterior tibial nerve of the foot. The reflex response was recorded from the tibialis anterior muscle and the number and duration of the pulses and the inter-phase interval were varied. A train of five square wave pulses with a duration of 0.5 ms separated by 1 ms were found most suitable in quantitative studies of the reflex. Twenty-four normal persons were investigated in this way as a basis for normal values. The day-to-day variations of the reflex were small in normal subjects.     

Homonymous and heteronymous monosynaptic reflexes in biceps brachii

Using poststimulus time histograms, it has been reported that stimulation of the median nerve at the elbow produces a monosynaptic EPSP in voluntarily active single motoneurons of the human biceps brachii. The present study was undertaken to: (i) determine whether such stimulation could evoke a reproducible reflex response in biceps brachii; and (ii) establish the optimal conditions for eliciting the reflex under clinical conditions. Twelve normal subjects were studied. No reflex response was recordable when biceps brachii was relaxed. A reflex response with a mean latency of 14.0 ms (±0.96 ms) could be recorded during a background voluntary contraction. The response was small (0.5–4.5% of the maximal M wave) but symmetrical, and could be obtained in all subjects. The responsible afferents appear to be rapidly conducting fibers from forearm flexor muscles and the latencies of the response were consistent with a monosynaptic reflex. Reflex amplitude increased with stimulus intensity and contraction strength. Stimulus rate did not affect amplitude significantly. It is concluded that a reproducible heteronymous monosynaptic reflex can be recorded from the contracting biceps brachii on stimulation of the median nerve at the elbow. Although smaller than the homonymous H reflex evoked by stimulation at Erb's point, it was technically easier to demonstrate that the EMG potential was of reflex origin (rather than part of an M wave). These reflexes should be of value in the assessment of the C-5/C-6 segments and the upper trunk of the brachial plexus. © 1995 John Wiley & Sons, Inc.     

Optimal stimulus duration for the H reflex

Authorities advocate different stimulus durations to produce an H reflex. In order to find the optimal stimulus duration for recording H reflexes, the recruitment curves for H reflexes and M responses were studied in 10 healthy subjects. The H reflex was recorded in the upper and lower extremities, and the durations of the electrical stimulus used ranged from 0.1 to 3 msec. The amplitude of the H reflex and the relation between the H reflex and M response changed with stimulus duration. H reflexes are brought out to advantage using a stimulus duration between 0.5 and 1 msec.     

Extinction of the soleus H reflex induced by conditioning stimulus given after test stimulus

OBJECTIVE: To quantify the extinction of the soleus H reflex induced by a conditioning stimulus above the motor threshold to the post-tibial nerve applied 10-12 ms after a test stimulus (S2 method). METHODS: Ten healthy subjects participated. The sizes of extinction induced by a test stimulus above the motor threshold (conventional method) and by the S2 method were measured. RESULTS: The size of the conditioned H reflex decreased as the intensity of the S2 conditioning stimulus increased. The decrease was less than that induced by the conventional method. The difference between the two methods correlated highly with the amount of orthodromically activated recurrent inhibition. When the S2 conditioning stimulus evoked an M wave that was roughly half of the maximum M wave, the decrease in the size of the conditioned H reflex depended on the size of the unconditioned H reflex. CONCLUSION: The S2 method allows us to observe extinction without changing the intensity of the test stimulus. The amount of the extinction depends partially on the size of the unconditioned H reflex. The difference in the sizes of extinction between the S2 and conventional methods should relate to recurrent inhibition.     

EXTINCTION OF THE SOLEUS H REFLEX INDUCED BY CONDITIONING STIMULUS GIVEN AFTER TEST STIMULUS

Objective. To quantify the extinction of the soleus H reflex induced by a conditioning stimulus above the motor threshold to the post-tibial nerve applied 10 12 ms after a test stimulus (S2 method). Methods. Ten healthy subjects participated. The sizes of extinction induced by a test stimulus above the motor threshold (conventional method) and by the S2 method were measured. Results. The size of the conditioned H reflex decreased as the intensity of the S2 conditioning stimulus increased. The decrease was less than that induced by the conventional method. The difference between the two methods correlated highly with the amount of orthodromically activated recurrent inhibition. When the S2 conditioning stimulus evoked an M wave that was roughly half of the maximum M wave, the decrease in the size of the conditioned H reflex depended on the size of the unconditioned H reflex. Conclusion. The S2 method allows us to observe extinction without changing the intensity of the test stimulus. The amount of the extinction depends partially on the size of the unconditioned H reflex. The difference in the sizes of extinction between the S2 and conventional methods should relate to recurrent inhibition.     

Cardiac sympathetic dysfunction in Parkinson's disease--relationship between results of 123I-MIBG scintigraphy and autonomic nervous function evaluated by the Valsalva maneuver]

We examined whether the results of 123I-MIBG scintigraphy reflect cardiac sympathetic nerve function in patients with Parkinson's disease (PD). The subjects were 62 patients with Parkinson's disease (age, 65.4 +/- 6.3 years) and 53 controls (65.2 +/- 7.1 years). All subjects underwent 123I-MIBG scintigraphy and QTc interval measurement on ECG. Hemodynamic autonomic function was estimated by the Valsalva maneuver in 37 subjects (63.9 +/- 5.2 years) randomly selected from the patients with PD. As control, the Valsalva maneuver was also done in 20 randomly selected controls (64.1 +/- 5.0 years), and 123I-MIBG scintigraphy was performed in 21 controls (67.7 +/- 5.3 years old). The subjects rested in a supine position for 20 min and were given an intravenous injection of 111 MBq 123I-MIBG. Relative organ uptake was determined by the region of interest (ROI) in the anterior view and the ratio of average pixel count in the heart (H) to that in the mediastinum (M) was calculated (H/M ratio) for early (after 15 min) and delayed (after 3 hrs) periods. The Valsalva maneuver was done by having the subjects exhale into a mouthpiece at an expiratory pressure of 40 mmHg for 15 seconds. Blood pressure and RR intervals were measured during the Valsalva maneuver by tonometry, using a noninvasive blood pressure monitoring system (ANS 508, Nihon Colin Co., Ltd.). Baroreceptor reflex sensitivities (BRS) of the second phase (BRS II) and fourth phase (BRS IV) of the Valsalva maneuver were calculated, and blood pressure elevations during the late second phase (IIp) and fourth phase (IVp) were measured. QTc was greater in the patients with PD (417 ms) than in the control subjects (409 ms). The H/M ratios of the early and delayed images in the patients with PD (1.76, 1.61) were significantly lower than those in the control subjects (2.56, 2.45). The early and delayed H/M ratios significantly correlated with the severity of disease according to Hoehn-Yahr stage. QTc interval and IVp significantly correlated with early and delayed H/M ratios. No other significant correlations were detected. The early H/M ratio in the patients with PD who had IVp within the normal range was lower than the early H/M ratio in control subjects. Our results show that early and delayed H/M ratio correlates with cardiac autonomic function, evaluated on the basis of QTc interval and the Valsalva maneuver, but not with baroreceptor reflex sensitivity or vasomotor autonomic function. Our findings suggest that silent cardiac autonomic dysfunction may be evaluated by 123I-MIBG scintigraphy, because early and delayed H/M ratios were lower in the patients with PD who had normal IVp than in the control subjects.     

Reflex effects of induced muscle contraction in normal and spinal cord injured subjects

The modulation of the soleus H reflex in response to functional electrical stimulation (FES) of the rectus femoris (RF) muscle and its overlying skin was examined in 11 normal adults and 6 patients with a clinically defined complete spinal cord injury (SCI). Stimulation of RF at twice motor threshold (MT) resulted in a long-lasting (>1,000 ms) and significant reduction (50-70% of control) in the size of the soleus H reflex in all normal subjects tested. For five of the SCI subjects, 2MT stimulation of RF induced a 55-60% reduction in the soleus H reflex that was also long-lasting (>160 ms). In the remaining SCI subject, 2MT stimulation resulted in an initial period of significant H-reflex facilitation (0-14 ms) that was followed by a longer-lasting inhibition commencing 60 ms after the cessation of the conditioning stimulation. Decreasing the strength of stimulation to below that required to generate a clear contraction in RF resulted in mixed facilitatory and inhibitory actions that were subject dependent. The changes in H-reflex excitability resulting from FES highlight the potential use of FES in the management of hypertonicity in SCI but also suggest that the central actions of FES need to be considered when FES gait restoration programs are designed.     

The Hoffmann reflex of human plantar foot muscles

Electrical stimulation of the tibial nerve in the popliteal fossa evoked an M wave (10.9 ms) and a late reflex response (38.1 ms) in the plantar foot muscles of all 10 volunteers. The late response had a somewhat lower electrical threshold than the corresponding M wave (8.5 versus 9 mA), and reached a maximum of amplitude when the stimulus intensity was increased, but was strongly suppressed by further increased intensity. A more distal stimulation of the tibial nerve at the ankle shortened the onset latency of the M wave and lengthened that of the late response. The reflex was facilitated by activation of synergists and inhibited by activation of antagonists. We showed that the late response was contaminated neither by volume conducted activity from the soleus muscle, as shown by intramuscular recordings from the abductor hallucis muscle, nor by a F wave, as shown by double stimulation. In summary, we conclude that this late response in human plantar foot muscles corresponded to an H reflex, which may be used to assess alterations of distal motoneuronal excitability. © 1998 John Wiley & Sons, Inc. Muscle Nerve 21:732–738, 1998.     

Abnormalities of prepulse inhibition do not depend on blink reflex excitability: a study in Parkinson's disease and Huntington's disease.

OBJECTIVE: Prepulse inhibition of the blink reflex is a robust phenomenon with an interesting physiology and a large potential for clinical applicability. In the study presented here we investigated whether the blink reflex inhibition by a prepulse (BRIP) is influenced by the blink reflex excitability recovery (BRER). METHODS: The study was undertaken in 20 patients with Parkinson's disease (PD), 20 patients with Huntington's disease (HD) and 20 healthy volunteers. BRER was determined by measuring the size of the response to a test supraorbital nerve stimulus as a percentage of the response to a conditioning stimulus at inter-stimuli intervals of 100-1000 ms. BRIP was determined as the percentage reduction induced in the response to a supraorbital nerve stimulus by either a low intensity auditory click or a weak third finger somatosensory stimulus, applied with a leading interval of 50-110 ms. RESULTS: There was a negative correlation between the percentage BRER and the percentage BRIP (Pearson's correlation coefficient of -0.37). BRER was enhanced in 14 PD patients (70%) and 6 HD patients (30%), while it was depressed in 10 HD patients (50%). BRIP was significantly reduced in 15 PD patients (75%) and 16 HD patients (80%). No significant correlation was found between abnormally enhanced BRER and abnormally reduced BRIP in all patients as a group (chi(2)=2.4;P=0.11). A weak correlation was found in PD patients (P=0.019) and no correlation was observed in HD patients (P=0.8). CONCLUSIONS: Our results indicate that an abnormally reduced BRIP was not always accompanied by an abnormally enhanced BRER in patients with HD. The two tests likely assess specific and distinct brainstem functions, and provide different types of information. While BRIP may be the result of a widespread integrative processing of sensory stimuli, BRER likely reflects the excitability of a chain of brainstem inter-neurons. SIGNIFICANCE: BRER and BRIP provide independent information on the state of functionally separate circuits that converge on trigemino-facial brainstem inter-neurons.     

Parametric Studies on Electroacupuncture-Like Stimulation in a Rat Model: Effects of Intensity, Frequency, and Duration of Stimulation on Evoked Antinociception

We have found that electroacupuncture-like stimulation of defined sites in the hindlimb of the rat inhibits a nociceptive withdrawal reflex. The lightly anaesthetized rat was used and tail withdrawal from a noxious radiant heat stimulus was the nociceptive reflex. Standard stimulation of hindlimb meridian points femur-futu (ST-32), fengshi (GB-31), and zusanli (ST-36) consisted of a 2-ms square voltage pulse at 4 Hz for a duration of 20 min, applied at 20 times the threshold to evoke muscle twitch. This produced two types of inhibition of the reflex; one was an increase in the latency of up to 80% during the stimulation, termed the brief antinociception, and the other was a post stimulation increase of up to 60% lasting greater than 1 h, termed the persistent antinociception. When the stimulus intensity was reduced to 10 times threshold, the latency during stimulation increased up to 50%, but the persistent response did not occur. Stimulation at threshold produced neither effect. When the train duration was altered, 10 min of stimulation produced only the brief effect, whereas 40 min of stimulation produced both effects, although the persistent effect lasted only 20 min. Stimulation at 6 Hz produced responses similar to those at 4 Hz, whereas stimulation at 2 Hz produced smaller effects. At 8 Hz, only the brief antinociception was elicited. With a pulse duration of 0.2 ms, the brief response was observed but the persistent response was markedly attenuated, whereas 5 ms produced responses similar to those with 2 ms. These data suggest that high-intensity, low-frequency electrical stimulation of meridian points in the rat hindlimb produces both brief and persistent antinociceptive effects on the tail withdrawal reflex, and both effects are dependent upon the parameters of stimulation. The persistence of the latter effect beyond the period of stimulation suggests events occurring after direct synaptic activity, possibly mediated via plastic changes at spinal and/or supraspinal levels.     

Sympathetic sudomotor and vasoconstrictive neural function in patients with Parkinson's disease

To evaluate sympathetic sudomotor and vasoconstrictive neural function in Parkinson's disease (PD), we simultaneously recorded sympathetic skin response (SSR) and skin blood flow (SVR; skin vasomotor reflex), as well as skin sympathetic nerve activity (SSNA) measured in peroneal nerves by microneurography, comparing 12 patients with idiopathic PD with 16 healthy controls. Resting SSNA frequency (8.8 ± 4.3 bursts/min) was significantly lower in PD patients than in controls (p < 0.01). Frequency increases in response to performing mental arithmetic were slightly smaller in PD patients than in controls. PD patients exhibited normal SSNA reflex latencies compared with controls. Although no significant relationship was found between resting SSNA frequency and disease duration or degree of disability, a significantly negative correlation between increases in SSNA with mental arithmetic and PD duration was observed. Occurrence of SSR and SVR following SSNA bursts induced by electrical stimuli was reduced in PD (p < 0.05). In patients with PD, sympathetic sudomotor and vasoconstrictive neural function was decreased at rest, but SSNA reflex latencies in the legs were nearly normal. Since responses of peripheral target organs may be impaired, both central and peripheral factors may contribute to autonomic symptoms in PD.     

Soleus H reflex extinction in controls and spastic patients: ordered occlusion or diffuse inhibition?

Extinction of the soleus H reflex at higher stimulus intensities is commonly attributed to retrograde conduction of action potentials in motor axons. This study was designed to gain further insight into the mechanisms underlying the extinction. The decrease of the H reflex was quantified in a group of controls and spastic patients, with and without depression of the H response by continuous tendon vibration. Response amplitudes were normalized as a percentage of the maximal M wave amplitude. Stimuli were normalized as a multiple of the M wave threshold. After normalization, the mean M recruitment curves, and similarly the fractions of motor axons activated, were equal in each group. In contrast, the mean H reflex amplitudes at the M threshold were different. The mean H reflex decrease, between 1.0 and 1.5 times the M threshold, was found to be the same fraction of the maximal H reflex amplitude in each group. The largest motor fibres, belonging to the largest motoneurones, are traditionally thought to have the lowest threshold for electrical excitation. Collision or retrograde inactivation should therefore preferentially affect the largest motoneurones, employed in only the largest H reflexes, at the lowest stimulus intensities. Our results are contrary to this hypothesis. Renshaw and/or Ib inhibition is likely to play a role in the initial decrease of the H reflex at higher stimulus intensities.