Quantitative EMG, MUNE & MUNIX

Quantitative MUAP analysis and decomposition

In the classical quantitative method, the examiner samples 20 distinct motor-unit action potentials at slight voluntary effort and measures, for each, its amplitude, duration, number of phases and turns. The mean duration is the most robust discriminator: it lengthens with reinnervation (neurogenic remodelling enlarges the unit's territory and desynchronises its fibres) and shortens with myopathic fibre loss. Comparison against age- and muscle-matched reference values — historically the Buchthal tables — converts the morphology into a statistical statement.

Modern systems automate this through decomposition. The interference pattern recorded at moderate effort is mathematically separated into its constituent motor-unit trains by template matching: the algorithm identifies recurring waveform shapes, assigns each firing to a template, and resolves the overlapping discharges of several simultaneously active units. Surface and intramuscular high-density arrays extend this to decompose dozens of concurrent trains, recovering not only each potential's shape but its individual firing rate and recruitment behaviour. Decomposition thus delivers quantitative MUAP parameters and firing statistics from a natural contraction rather than from painstakingly isolated single units.

Parameters, outlier analysis, and interference-pattern methods

Quantitative interpretation rests on two complementary strategies. Mean-value analysis compares the sample mean of a parameter against the normative mean. Outlier analysis asks instead how many individual potentials fall outside the normal range — a single grossly abnormal unit among twenty normal ones may flag early pathology that the mean obscures. A pattern of a few markedly enlarged, long-duration outliers is the quantitative signature of early reinnervation. Where isolating individual potentials is impractical, interference-pattern analysis (turns/amplitude, the Willison method, or activity/cloud methods) characterises the fully recruited signal statistically, trading single-unit resolution for speed and operator-independence.

Motor unit number estimation (MUNE)

The maximal compound muscle action potential (CMAP) is the summed response of every functioning motor unit in a muscle. If one could measure the size of a single, representative motor-unit potential (SMUP) recorded at the same surface electrode, then the count of units follows from a deceptively simple ratio:

MUNE = maximal CMAP size ÷ mean SMUP size

The entire methodological challenge of MUNE is obtaining an unbiased estimate of the mean SMUP. The historical approaches differ chiefly in how they sample single units:

• Incremental stimulation — slowly increasing stimulus intensity recruits axons one at a time; quantal steps in the rising CMAP are taken as individual SMUPs. Simple, but corrupted by alternation, in which axons of similar threshold fire in shifting combinations and inflate the apparent number of discrete increments.
• Multiple-point stimulation — single SMUPs are collected from many sites along the nerve at just-threshold intensity, sidestepping alternation by sampling one all-or-none unit per site.
• Statistical (Poisson) MUNE — exploits the variance of submaximal responses; under Poisson assumptions the variance of CMAP fluctuation at fixed stimulus intensity yields the mean SMUP size without isolating individual units.
• Spike-triggered averaging — uses a voluntarily activated single unit (captured on a needle) as the trigger to extract its surface-recorded SMUP, sampling the larger, voluntarily recruited population rather than the low-threshold axons favoured by stimulation.

Assumptions, error sources, and the role in trials

Every MUNE method inherits the same vulnerable premises: that the sampled units are representative of the whole pool, that SMUP size cancellation from phase overlap is negligible, and that the surface electrode captures all contributing units. Stimulation-based methods oversample low-threshold axons; phase cancellation between asynchronous units causes the summed CMAP to underestimate the arithmetic sum of SMUPs, biasing counts upward; and reinnervation enlarges SMUPs, so an unchanged MUNE can mask ongoing denervation that collateral sprouting is compensating for. MUNIX shares the cancellation problem and adds a dependence on consistent, well-graded voluntary effort.

Despite these limitations, MUNE and MUNIX occupy a unique niche: they are the only electrodiagnostic measures that estimate the denominator of the disease — the number of surviving motor units — rather than the morphology of the survivors. In ALS clinical trials this makes them valuable as objective, quantitative endpoints that decline steadily and detect therapeutic slowing of motor-unit loss earlier than strength or function scales.