Motor Nerve Conduction

The CMAP as a population summation

The compound muscle action potential (CMAP) is the spatially and temporally summed extracellular field of every muscle fibre depolarised by a supramaximal nerve stimulus, recorded by a surface electrode placed over the motor point in a belly–tendon montage. Its negative-peak amplitude, measured from baseline to the trough of the initial negativity, is the canonical index of the number of conducting motor axons and the muscle fibres they innervate. Amplitude is the quantity most sensitive to axon loss, but it is a non-linear readout: because surviving motor units enlarge by collateral reinnervation, amplitude is partially defended until axonal dropout is substantial, which is precisely why a "normal" CMAP never excludes chronic neurogenic disease and must be read alongside the needle examination.

The amplitude measurement is only interpretable when stimulation is supramaximal — typically the current is advanced 20–30% beyond the intensity that first recruits the last axon, so that the response plateaus and becomes reproducible. Sub-maximal stimulation spuriously lowers amplitude and is the single most common reason a normal nerve is mislabelled axonal. Area under the negative phase carries equivalent information to amplitude for axon counting and is the preferred currency when waveforms are dispersed, because area is conserved across phase shifts that erode peak amplitude.

Distal latency is a composite, not a velocity

The distal (onset) latency is the interval from stimulus artifact to the first reproducible deflection of the CMAP from baseline. It is irreducibly a composite of three serial delays: conduction time along the terminal motor axon, the neuromuscular transmission delay at the endplate (roughly 0.5–1.0 ms of synaptic and junctional time), and the muscle fibre activation and propagation time. Because the latter two components are conduction-independent fixed costs, distal latency cannot be divided by distance to yield a velocity — doing so contaminates the estimate with the constant NMJ and activation overhead and systematically underestimates true axonal speed. Distal latency is therefore reported as a latency, compared against a nerve-specific normative limit, and interpreted as a sentinel of distal demyelination (as in the prolonged terminal latency of carpal tunnel syndrome) rather than converted into a conduction metric.

Conduction velocity by two-site subtraction

A true motor conduction velocity is recoverable only by stimulating at two points along the nerve and exploiting the fact that the fixed distal overhead is shared by both responses. The proximal and distal onset latencies each contain the same NMJ and activation delay; subtracting them cancels that constant, leaving a pure inter-site conduction time:

• CV (m/s) = inter-site distance (mm) ÷ (proximal latency − distal latency, ms). The surface distance is measured along the anatomical course of the nerve between cathodes.
• Velocity reflects only the fastest, largest-diameter fibres — the ones that arrive first and define the onset — so it is relatively insensitive to loss of small slow fibres and can remain near-normal in pure axonal disease until the fast population is decimated.

Because the measurement is a difference, it amplifies any error in either latency or in the distance estimate; a 1 cm error in surface measurement over a 25 cm forearm segment shifts velocity by several m/s. This is the structural reason short-segment velocities are noisier than long-segment ones and why "inching" demands meticulous technique.

Temporal dispersion versus conduction block

When the proximal CMAP is smaller than the distal CMAP, two mechanistically distinct phenomena compete to explain it, and distinguishing them is the diagnostic crux of demyelinating disease. Temporal dispersion is the consequence of differential slowing: along a longer proximal segment, the spread of conduction velocities across individual fibres widens the arrival-time distribution, so single-fibre potentials that once summed in phase now overlap their positive and negative phases. The result is phase cancellation — a broader, more polyphasic, lower-amplitude waveform whose negative-peak amplitude falls while its area is comparatively preserved, and whose total duration lengthens. Dispersion alone does not signify that any axon has stopped conducting; it is a reordering of arrival times, not a loss of impulses.

Conduction block, by contrast, is the genuine failure of impulse propagation across a focal demyelinated segment whose safety factor has fallen below unity. Here, fibres that conduct distal to the lesion simply do not conduct through it, so the proximal response loses both amplitude and area. The conventional electrodiagnostic threshold is a >50% drop in negative-peak amplitude or area on proximal versus distal stimulation, with a duration increase modest enough (commonly defined as <30%) to exclude dispersion as the sole cause. Partial conduction block is the electrophysiological signature of an acquired, multifocal demyelinating process and, when present outside common compression sites, is a near-defining feature of CIDP and multifocal motor neuropathy.

F-waves and reference values

The antidromic motor volley that ascends to the anterior horn also generates a small recurrent discharge — the F-wave — which uniquely samples the proximal segment and root that surface stimulation cannot reach; it is treated in depth in the late-responses module. For routine interpretation, nerve-specific normative limits anchor every judgement: median and ulnar motor distal latencies are generally abnormal beyond ~4.2 ms and ~3.3 ms respectively, forearm and leg velocities below roughly 50 m/s (upper limb) and 40 m/s (lower limb) are slow, and amplitudes below the laboratory lower limit signal axon loss or block. These thresholds are temperature-corrected; an unwarmed limb is the commonest source of spurious slowing.