Late Responses, Repetitive Nerve Stimulation, Single-Fiber EMG, and Autonomic Testing

Section 1: F-Waves and the H-Reflex — Two Late Responses That Are Not the Same

~0:00 – F-Waves and the H-Reflex — Two Late Responses…

Bottom line: the F-wave is NOT a reflex but an antidromic motor backfire requiring supramaximal stimulation, with variable latency and morphology because a different ~1-5% of anterior horn cells (AHCs) backfire each time; the H-reflex IS a true monosynaptic Group Ia reflex requiring submaximal stimulation, with constant latency and morphology, and disappears at supramaximal intensity through antidromic collision; the tibial H-reflex is the most useful electrodiagnostic test for S1 radiculopathy.

The single most important board fact is that the F-wave is not a reflex. It has no afferent limb, no synapse, and no sensory component. Mechanism: a supramaximal stimulus to a motor nerve sends an action potential in two directions. Orthodromically, distally toward the muscle, producing the M-wave (CMAP). Antidromically, proximally back along the motor axon to the AHC body. When the antidromic impulse reaches the AHC, approximately 1-5% of motor neurons “backfire”: they re-excite and generate a new orthodromic action potential traveling back down the same motor axon to the muscle, producing the late F-wave. The letter F stands for “foot” because Magladery and McDougal first recorded it from foot muscles in 1950.

Because a different random 1-5% subset of AHCs backfires with each successive stimulus, the F-wave has variable latency and variable morphology. F-waves require supramaximal stimulation for three reasons: all motor axons must be activated to maximize backfiring; supramaximal stimulation eliminates the H-reflex through antidromic collision; and reproducibility requires consistent activation of the motor neuron pool. Ten to twenty consecutive F-waves are recorded per nerve (minimum 10 for reliable minimal latency). Key parameters: minimal F-wave latency (the most commonly used clinical parameter, reflecting the fastest motor fibers’ conduction from stimulation site to spinal cord and back, plus ~1 ms AHC turnaround); chronodispersion (max minus min latency); F-wave persistence (% of stimuli producing an F-wave, normal >50-80%); the F ratio = (F latency − M latency − 1) / (2 × M latency). Normal minimal F-wave latencies are approximately 25-32 ms for median and ulnar at the wrist and approximately 45-56 ms for tibial and fibular at the ankle. A side-to-side difference greater than 2 ms is abnormal.

F-wave clinical applications center on assessing proximal motor segments inaccessible to routine NCS. In Guillain-Barre syndrome (GBS), F-wave abnormalities (prolonged latency, decreased persistence, or absent F-waves) may be the earliest electrodiagnostic finding because GBS frequently affects roots first; distal NCS can be normal in the first few days while F-waves are already abnormal. In chronic inflammatory demyelinating polyneuropathy (CIDP), prolonged latencies, increased chronodispersion, and decreased persistence suggest preferential proximal demyelination. In radiculopathy, F-wave sensitivity is only ~10-20% and F-waves cannot localize to a specific root because multiple roots contribute motor fibers to each peripheral nerve. In amyotrophic lateral sclerosis (ALS), the depleted motor neuron pool produces repeater F-waves (identical morphology on consecutive stimuli) because the same few neurons backfire repeatedly; repeaters are a hallmark of motor neuron disease.

The H-reflex is fundamentally different. It is a true monosynaptic reflex, the electrical analog of the muscle stretch reflex (named for Paul Hoffmann, 1910). The arc has three parts: the afferent limb is Group Ia sensory fibers from muscle spindle primary endings (the largest, lowest-threshold sensory fibers); the central relay is a single synapse between Ia afferents and alpha motor neurons; the efferent limb is alpha motor neurons traveling via the ventral root to the muscle. Because the same motor neuron pool fires through the same arc each time, the H-reflex has constant latency and constant morphology. The standard clinical study is the tibial (soleus) H-reflex: stimulate the tibial nerve in the popliteal fossa with a long-duration stimulus (0.5-1.0 ms) to preferentially activate the large Ia afferents, record from the soleus midway between the popliteal crease and calcaneus. Normal latency is approximately 28-35 ms (height-dependent); side-to-side latency difference greater than 1.0-1.5 ms is abnormal; H/M ratio is approximately 0.5-0.7.

The H-reflex behavior with increasing stimulus intensity is a board favorite. At very low intensity, Ia afferents (lowest threshold) activate first; the H-reflex appears with no M-wave. As intensity rises to submaximal, more Ia fibers recruit, the H-reflex grows, and a small M-wave appears. At moderate intensity, the H-reflex reaches maximum amplitude. As intensity approaches supramaximal, the H-reflex amplitude decreases because the antidromic motor volley collides with and cancels the orthodromic H-reflex impulses returning from the cord. At supramaximal intensity, the H-reflex disappears completely, the M-wave is maximal, and the F-wave appears. The H-reflex requires submaximal stimulation; the F-wave requires supramaximal stimulation.

Clinical applications: the tibial H-reflex is the most useful electrodiagnostic test for S1 radiculopathy, the electrical equivalent of the ankle jerk. An absent or prolonged unilateral tibial H-reflex strongly suggests S1 root dysfunction; side-to-side comparison is more useful than absolute values; sensitivity is approximately 50% overall (up to 80% when acute). H-reflexes are prolonged or absent bilaterally in polyneuropathies (especially demyelinating); bilateral absence can be the earliest finding in GBS or diabetic polyneuropathy. Important caveat: H-reflexes may be absent bilaterally in normal individuals over age 60; unilateral absence remains significant at any age. The flexor carpi radialis (FCR) H-reflex tests the C6-C7 arc with reported sensitivity ~72% and specificity ~85% for cervical radiculopathy, but is not widely used due to technical difficulty.