H reflexes and F waves are commonly recorded electrophysiologic responses ( Figs. 18-1 , 18-2 , and 18-3 ). H reflexes are reflexes that are produced by afferent conduction in large afferent fibers and by efferent conduction in alpha motor neurons. By contrast, F waves are produced by antidromic activation (“backfiring”) of the alpha motor neurons. Despite their differences ( Table 18-1 ), they commonly are discussed together. H reflexes and F waves are studied for similar clinical problems, are found at comparable latencies, reflect conduction to and from the spinal cord, and involve motor neuron activation.
|Feature||H Reflexes||F Waves|
|Response||Reflex||Antidromic firing of motor neurons|
|Afferent fibers||Ia afferents||Alpha motor fibers|
|Efferent fibers||Alpha motor fibers||Alpha motor fibers|
|Increasing stimulus intensity||Inhibits response||Facilitates response|
|Motor units||Different from M wave||Same as in M wave|
In 1918, Hoffman described a reflex response in calf muscles that followed stimulation of the posterior tibial nerve and was comparable in latency to the Achilles reflex. In recognition of Hoffman’s original contribution, the response was named the H reflex by Magladery and McDougal.
Because H reflexes involve conduction from the periphery to and from the spinal cord, they occur at latencies considerably longer than the latency of a direct motor response. A necessary condition for establishing an H reflex is that this “late” response must be larger than the preceding direct motor response. This condition can occur only with central amplification of the motor response caused by reflex activation of motor neurons. Careful observation of H reflexes as they increase in amplitude often reveals a decrease in latency. This decrease is consistent with the orderly activation of smaller and then larger motor neurons, with an associated increase in axonal conduction velocities, as would be expected in a reflex response.
The arc of the H reflex includes conduction in large, fast-conducting Ia fibers. In that sense, the H reflex is similar to the phasic myotatic (“deep tendon”) reflex produced by muscle stretch. Although the phasic myotatic reflex generally is considered monosynaptic and there is a monosynaptic component to the H reflex, evidence exists that the H reflex also has oligosynaptic components. These include contributions from spindle afferents, as well as inhibitory Ib effects from Golgi tendon organs. Achilles and calf H reflexes are generally present together, and calf H-reflex amplitudes have been correlated positively with prominence of the Achilles reflex. Unlike the phasic myotatic reflex, however, the H reflex does not involve muscle spindle activation. This difference at times may explain the presence of calf H reflexes in the absence of Achilles reflexes.
H reflexes are inhibited as the stimulus intensity is increased from submaximal to that required for eliciting a maximal direct (M) response. This relationship has been explained by “collision” of orthodromic impulses with impulses conducted antidromically in motor axons. In fact, this mechanism has little or no role in the inhibition of H reflexes that occurs with increasing stimulus intensity. To be complete, such collision must occur distal to the motor neurons. Allowing for rise times for motor neuron excitatory postsynaptic potentials (EPSPs) of at least 3 msec, as well as for at least one synapse, the differences in afferent and efferent conduction in H reflexes necessary for collision inhibition have not been established. Large H reflexes are obtained from calf muscles, even with supramaximal stimulation, if the stimuli are timed appropriately with phasic contractions of the muscles. This indicates that H-reflex inhibition does not depend on peripheral collision of afferent and efferent impulses. Experimental studies in normal and spastic subjects are consistent with central inhibition of H reflexes occurring as stimulus intensities are increased.
Inhibitory mechanisms present in the spinal cord can explain H-reflex inhibition. Inhibitory interneurons (Renshaw cells) are activated by antidromic stimulation, are distributed widely throughout the motor neuron pool, and discharge more strongly and with shorter latency as stimulus intensity increases. H reflexes may be monosynaptic, whereas H-reflex inhibition by Renshaw cells would involve two synapses. Any resulting difference in the onset of inhibitory and excitatory effects may be as brief as 0.3 msec and therefore well within a reasonable physiologic range given motor neuron EPSP rise times. Direct connections between motor neurons are also present and could contribute to H-reflex inhibition. Single-fiber electromyographic (EMG) studies of H reflexes support a process of active inhibition involving inhibitory synapses with stimuli of increasing intensity.
In infants younger than 2 years, H reflexes are distributed widely. Beyond infancy, H reflexes are found regularly in calf muscles (primarily the soleus) and homologous forearm flexors. They are also commonly present in the quadriceps and occasionally in plantar foot muscles. This restricted distribution of H reflexes with age is caused by changes associated with physiologic maturation of the central nervous system (CNS).
The fraction of the soleus motor neuron pool activated in an H reflex is usually about 50 percent but can be as high as 100 percent. The ratio of the peak-to-peak maximum H-reflex to maximum M-wave amplitude (H/M ratio) provides a measure of motor neuron pool activation and therefore excitability. The H/M ratio for calf H reflexes is normally less than 0.7 but there is considerable variation. A recent study has recommended evaluating M responses and H reflexes over the entire stimulus range in order to standardize H-reflex responses better, not only between studies and subjects but also from the same subject. This method allows not only for normalization of the H reflex itself but also for the relative stimulus intensity needed to elicit a response.
H reflexes involve the activation of a portion of the segmental motor neuron pool and therefore are enhanced by maneuvers that increase excitability of the motor neuron pool. H reflexes can be produced by facilitation maneuvers (e.g., contraction or post-tetanic potentiation) in muscles where they are not otherwise present, such as the small hand muscles. H reflexes, particularly those of low amplitude, are depressed by prior stimulation. This is related to homosynaptic depression rather than presynaptic inhibition from flexor afferents.
H reflexes have been used to explore the physiology of the CNS. A recent review has emphasized the problems with equating H-reflex responsiveness with motor neuron excitability. Such equating of these two phenomena is limited, for example, by the effects of presynaptic inhibition and postactivation depression on H reflexes. In general, H reflexes are subject to a complex of segmental and suprasegmental influences related to both the physical and mental state of the individual. With these limitations, H reflexes remain an important tool for investigating the central control of movement.
H reflexes have been used to study patterns of central reflex organization, as reviewed elsewhere. Studies have shown changes in H reflexes consistent with an influence on the reflex of both group Ia and group II afferents. H reflexes have been used to analyze patterns of reciprocal inhibition. The methodology for this in the forearm has been well described. Myotatic arcs between proximal and distal muscles also affect H reflexes. Activation of the quadriceps, for example, inhibits the soleus H reflex. Presynaptic inhibition of Ia afferent fibers is decreased on fibers projecting on motor neurons of contracting muscles, but it is increased on afferents to motor neurons supplying muscles that are not involved in the contraction. Passive cyclic movements of the legs produce inhibition of H reflexes; this is probably because of presynaptic inhibition that can be integrated at the spinal cord level.
Characteristic changes in H-reflex amplitude can be defined if test stimuli are given at varying intervals after a conditioning stimulus. These excitability or recovery curves vary with the level of stimulation ( Fig. 18-4 ) and are not established clearly until 1 year of age.
H reflexes are inhibited by vibration as a result of vibration-induced activation of large afferent fibers with consequent peripheral “busy line” interference, presynaptic inhibition of afferent input, and activation of spindles in antagonist muscles. Vibration of the tibialis anterior muscle has been used to assess the degree of presynaptic inhibiton of Ia projections from the quadriceps to soleus motor neurons.
Technique of Recording H Reflexes
H reflexes are obtained readily with percutaneous stimulation and surface recording techniques. Responses are stable, but only under similar conditions of stimulation and recording. The stimulating cathode should be proximal to avoid anodal block. Stimulus pulses of long duration (1 msec) are used to activate the large sensory fibers preferentially. Stimulus frequency should be 0.2 Hz or less to allow recovery of postactivation depression of the H reflex from a prior stimulus.
A series of responses should be obtained. Starting with submaximal stimuli and increasing to supramaximal stimulation, one should verify whether the “late” response can be larger than the preceding direct motor response, determine which H reflex has the largest amplitude, and demonstrate that inhibition of the H reflex occurs with increasing stimulus intensity. Latencies should be measured to the onset of the responses (either negative or positive), and amplitudes should be measured from peak to peak.
For calf H reflexes, the posterior tibial nerve is stimulated in the popliteal fossa. Bipolar stimulation is usually adequate. Use of an anode with a large surface area at the patella can decrease stimulus artifact and may provide more discrete cathodal excitation of the nerve in the popliteal fossa. Recordings are made from the soleus muscle. Although techniques vary, a standard and convenient location for the active electrode is medial to the tibia at a point that is half the distance between the stimulation site and the medial malleolus, with the indifferent electrode placed on the Achilles tendon. H reflexes in the forearm are recorded readily from the flexor carpi radialis muscle. The recording electrode is placed at the junction of the upper one-third and lower two-thirds of the distance between the medial epicondyle and the radial styloid. The median nerve is stimulated percutaneously in the cubital fossa. At times, H reflexes in normal subjects can be recorded from the quadriceps with stimulation of the femoral nerve in the femoral triangle and from the abductor hallucis with stimulation of the tibial nerve at the ankle.
H reflexes are recorded routinely with the muscle at rest. Contraction of the recording muscle will enhance H reflexes by facilitating the motor neuron pool. Such contraction can help to identify H reflexes in muscles in which H reflexes are normally present, as well as elicit H reflexes in muscles where they are not normally found. This facilitation of H reflexes by contraction is sometimes clinically useful and demonstrates that the normal distribution of H reflexes is based on physiologic, not structural, factors.
Uses of H-Reflex Studies
The upper limit of normal latency for the H reflex of the soleus is 35 msec, and that of the flexor carpi radialis is 21 msec. H-reflex latencies are related directly to leg or arm length, height, and, to a lesser degree, age (see Fig. 18-4 ). For clinical purposes, it is best to use regression equations including height and age as variables. Normal values for infants and children are available. With careful technique, the upper limits of normal for side-to-side latency differences have been reported to be as low as 1.5 msec for calf muscles (mean, 0.09 ± 0.70 standard deviation [SD]) and 1.0 msec for forearm flexor muscles (mean, 0.002 ± 0.42 SD). For routine clinical work and consistent with a criterion of 3 SD from the mean, 2 msec should be allowed for side-to-side differences with recordings from the calf, and 1.5 msec allowed for recordings from the forearm. Side-to-side latency differences may be larger in the elderly. The upper limit of normal for the interside ratio of H-reflex amplitudes for calf muscles has been reported as 2, but a preferable figure for general clinical work is probably 3. This is similar to the interside ratio for H-reflex amplitudes in the forearm.
Calf and forearm H reflexes are usually present in normal subjects. As with phasic myotatic reflexes, however, this is not always true, and symmetrically absent H reflexes are not necessarily abnormal. The percentage of absent responses increases in the elderly.
Disorders of the Peripheral Nervous System
H reflexes are a sensitive test for polyneuropathies and may be abnormal even in mild neuropathies. H reflexes involve conduction in proximal as well as distal fibers. These studies, therefore, can define proximal nerve injury and may be abnormal even when studies of distal function are unremarkable. Absent H reflexes are characteristic of acute inflammatory demyelinating polyneuropathy (Guillain–Barré syndrome). This loss of H reflexes occurs early and may be an isolated finding in patients studied within several days after onset of illness. H reflexes may be abnormal in asymptomatic patients with possible neuropathic dysfunction, and in plexopathies and radiculopathies. H reflexes in the forearm flexor muscles may be abnormal with C6 or C7 root injury, and calf H reflexes may be abnormal with S1 radiculopathies. H reflexes are affected by injury to either the posterior or anterior roots. Examination of these reflexes therefore can be important in the electrodiagnostic evaluation of radiculopathies by documenting nerve injury, even when needle EMG is unrevealing owing to sparing of the anterior roots. Eliciting the H reflex by direct stimulation of the S1 root may enhance their utility for detecting an S1 radiculopathy. Increase in the stimulus threshold has been noted in patients with neurogenic claudication when studied after reproduction of their symptoms with walking.
Disorders of the Central Nervous System
H reflexes are a recognized probe for noninvasive study of the reflex organization of the CNS and of motor control in general, as reviewed elsewhere. H-reflex studies have also been used to monitor spinal cord plasticity.
The H reflexes may be abnormally widespread in patients with CNS lesions and upper motor neuron signs. The presence in adults of H reflexes in muscles where they are not normally present (e.g., in the tibialis anterior or small hand muscles) may be clinically useful for documenting dysfunction of the central motor system.
Patients with even mild hemiparesis have decreased potentiation of H reflexes with muscle contraction, and this finding is consistent with decreased background facilitation of motor neurons. At the same time, H/M ratios tend to be increased in patients with CNS lesions and upper motor neuron signs, and recruitment curves are altered in a manner consistent with increased excitability of the central motor neuron pool. Conversely, H reflexes may be depressed when excitability of the central motor neuron pool is decreased. As mentioned previously, however, a simplistic relation between motor neuron excitablity and H reflexes cannot be assumed, given the complexity of influences affecting the H reflex, including pre- and postsynaptic events.
Changes in H reflexes after CNS lesions are time dependent. H reflexes are depressed acutely after spinal cord injury. Increased H/M ratios develop during the weeks to months following a cerebrovascular lesion associated with the appearance of features of the upper motor neuron syndrome (e.g., increased tone, brisk reflexes, and extensor plantar responses). In patients with chronic upper motor neuron lesions, vibratory inhibition of H reflexes is less than expected, possibly because of decreased presynaptic inhibition. In contrast, vibratory inhibition of H reflexes may be enhanced in patients with acute cerebral lesions. H reflexes are preserved relatively well acutely after spinal shock, at a time when both Achilles reflexes and H-reflex recovery curves are depressed. Alterations in soleus H reflexes have been related to specific features of the upper motor neuron syndrome: in particular, decreased vibratory inhibition of H reflexes with hypertonia, increased H/M ratios with increased reflexes, and late facilitation of the recruitment curve with clonus.
Within several months after a central injury, H-reflex excitability curves can show an abnormally rapid pattern of recovery ( Fig. 18-5 ) associated with increased H/M ratios. These patterns differ from those in patients with parkinsonian rigidity or cerebellar hypotonia.