Nerve Conduction Studies: Interpretation¶
A clinician teaching reference: how to read NCS/EMG, representative normal values, and condition patterns.
Overview — what NCS/EMG can and cannot answer; the localisation mindset¶
Electrodiagnostic studies — nerve conduction studies (NCS) plus needle electromyography (EMG) — are an extension of the clinical examination, not a stand-alone diagnosis. Read alongside the history and examination they answer a small number of focused questions: is there a lesion of the peripheral nervous system, where is it, what is its pathophysiology (axon loss versus demyelination), how severe is it, and how old is it? [1][2] They cannot, on their own, tell you the aetiology (compression, trauma, inflammation, ischaemia all look broadly similar), and a normal study never fully excludes a small-fibre, very proximal, or very recent lesion.
The discipline that makes the report useful to a surgeon is the localisation mindset. Every abnormality is interpreted as a question of "where along the pathway?" — anterior horn cell, root, plexus, trunk, or named nerve; and within a named nerve, which segment. The single most powerful localising principle in the upper limb is the relationship of the lesion to the dorsal root ganglion (DRG): because the sensory cell body sits in the DRG (in the intervertebral foramen, distal to the root), a lesion proximal to the DRG — a root avulsion or radiculopathy — leaves the peripheral sensory axon in continuity and the sensory nerve action potential (SNAP) is preserved, whereas a lesion distal to the DRG (plexus or nerve) abolishes it. This is the lever that distinguishes radiculopathy from plexopathy, and it is developed in the condition-pattern section below. [1][3][8]
Two practical caveats frame everything that follows. First, EDX is operator- and technique-dependent; the numbers in a report are only as good as electrode placement, distances, and temperature control. Second, the study is most reliable when read comparatively — symptomatic versus asymptomatic side, one nerve against its neighbour, distal against proximal — because each patient is, in effect, their own control.
The measurements and what they mean — onset/peak latency, amplitude, conduction velocity, F-waves, H-reflex¶
A motor study stimulates a nerve and records the compound muscle action potential (CMAP) over the target muscle; a sensory study records the SNAP over the nerve itself (orthodromic or antidromic). Four primary measurements, plus the late responses, carry the information. [1][2]
- Onset latency — time (ms) from stimulus to the first deflection of the response. It reflects conduction in the fastest fibres and the distal-most segment, and is the value used for distal motor latency and median–ulnar comparisons. [1][2]
- Peak latency — time to the peak of the SNAP; more reproducible than onset for sensory studies and conventionally used for sensory distal latencies and median–ulnar/median–radial comparisons. [4][7]
- Amplitude — CMAP in millivolts (mV), SNAP in microvolts (µV). Amplitude is a semi-quantitative measure of the number of functioning axons (and, for the CMAP, muscle fibres) contributing to the response. Low amplitude is the signature of axon loss. [1][2]
- Conduction velocity (CV) — m/s, computed from the latency difference between two stimulation sites divided by the inter-site distance (motor) or from latency and distance (sensory). CV reflects myelination and the integrity of the largest, fastest fibres. Marked slowing implies demyelination. [1][2]
Late responses interrogate the proximal segments that routine distal stimulation cannot reach:
- F-waves — a small late motor response produced by antidromic activation of the anterior horn cell that "backfires" orthodromically to the muscle. F-wave latency samples the entire motor axon including the root, and is prolonged or absent in proximal demyelination (e.g. early Guillain–Barré) and in radiculopathy/plexopathy, though it is a relatively non-specific and insensitive localiser. [1][2]
- H-reflex — the electrophysiological analogue of the monosynaptic stretch reflex (Ia afferent → spinal cord → motor efferent), routinely elicited from soleus (S1) and used mainly in the lower limb; upper-limb (flexor carpi radialis, C6/C7) H-reflexes are technically harder and less routine. Normal soleus H-reflex latency is on the order of ~28–35 ms (height-dependent). [1][2]
The mnemonic that survives contact with a real report: amplitude ≈ axon number; latency/CV ≈ myelination. Keep the two channels separate when you read.
A systematic way to read a study — sensory then motor; distal vs proximal; side-to-side; amplitudes; CV/latency; late responses; then needle EMG¶
A report is best read in a fixed order so nothing is missed: [1][2][4]
- Sensory studies first. SNAPs are the most sensitive early marker of an axonal lesion distal to the DRG and the key discriminator for the radiculopathy-versus-plexopathy question. Check each SNAP is present, then its amplitude, then its peak latency/CV.
- Motor studies. For each nerve check distal motor latency, CMAP amplitude (distal), and forearm/across-elbow CV. CMAP amplitude loss lags sensory loss in many compressive lesions.
- Distal versus proximal. Compare the response on stimulating distally and proximally along the same nerve. A drop in CMAP amplitude/area from proximal to distal stimulation (with preserved distal latency) defines conduction block — the hallmark of a focal demyelinating lesion between the two stimulation sites (e.g. ulnar block across the elbow). Disproportionate temporal dispersion (a longer, lower, more polyphasic proximal response) points to differential slowing across the segment.
- Side-to-side. For a unilateral problem, the contralateral nerve is the most useful normal reference; inter-side differences in latency or amplitude are often more informative than absolute values against a population range.
- CV/latency are then interpreted as a block: focal slowing or block localises; uniform slowing across many nerves suggests a generalised demyelinating process; near-normal CV with low amplitudes suggests axon loss.
- Late responses (F-waves, H-reflex) add proximal information when distal studies are normal or borderline.
- Needle EMG is the final and often most localising step. Assess, in each muscle: spontaneous activity at rest (fibrillation potentials and positive sharp waves — markers of active denervation; complex repetitive discharges, fasciculations, myokymia, myotonia), motor-unit action potential (MUAP) morphology (large, long, polyphasic units indicate chronic reinnervation; small, short units suggest myopathy), and recruitment (reduced/discrete recruitment with a fast firing rate indicates neurogenic axon/motor-unit loss). The distribution of abnormal muscles — sampled across nerves and across roots/myotomes, including paraspinals — is what converts EMG findings into an anatomical localisation. [1][2][3]
Axonal vs demyelinating — the core dichotomy¶
Almost every electrodiagnostic interpretation reduces to placing the lesion on the axonal–demyelinating axis, because the two have different mechanisms, prognoses, and (for the surgeon) implications. [1][2][5]
Axon loss (the typical picture in chronic compression that has progressed, traction injury, axonotmesis, and most toxic/metabolic neuropathies): - Low amplitude is the cardinal finding — SNAP first, then CMAP. - Latency and CV are relatively preserved (mild slowing only, because the fastest surviving fibres still conduct near-normally; CV typically stays above ~70–80% of the lower limit of normal). - Needle EMG shows fibrillations/positive sharp waves (after the denervation delay) and, later, reinnervation changes (large polyphasic MUAPs, reduced recruitment). - Prognosis depends on axon regrowth (~1 mm/day) and is generally slower and less complete.
Demyelination (focal entrapment with block, inflammatory demyelinating neuropathies, some hereditary neuropathies): - Slowing out of proportion to amplitude loss — prolonged distal latencies, slow CV, prolonged/absent F-waves. - Conduction block (amplitude/area drop on proximal vs distal stimulation across the affected segment) and temporal dispersion are specific for acquired segmental demyelination and are highly localising. - Needle EMG is often relatively normal early (no axon loss = no fibrillations), or shows only reduced recruitment due to block. - Prognosis is generally good and fast if remyelination occurs (no axon loss to regrow).
A further axis is uniform versus segmental. Uniform slowing across all nerves and segments suggests a hereditary demyelinating process (e.g. CMT1); patchy, segmental slowing with block and dispersion suggests an acquired (often inflammatory) demyelinating process. Most surgically relevant entrapments produce focal demyelination at the entrapment site, sometimes with superimposed axon loss once the lesion is chronic. [1][2][5]
Typical normal values (representative)¶
The values below are representative adult upper-limb figures drawn from widely used reference compilations and the AANEM Normative Data Task Force review. They are not absolute. Every reference range is laboratory- and technique-specific and varies with limb temperature (a cool limb slows conduction and prolongs latency, roughly 1.5–2.5 m/s and ~0.2 ms per °C), inter-electrode distance, the patient's age and height, and the recording montage. Always interpret a result against the issuing laboratory's own reference range and, ideally, the patient's asymptomatic side — never against the table alone. [1][4][5][6][7]
| Study (adult, surface, ~32–34 °C) | Typical normal limit | Parameter |
|---|---|---|
| Median motor (wrist→APB) | distal latency ≤ ~4.2–4.4 ms | onset latency |
| Median motor CMAP | ≥ ~4 mV | amplitude |
| Median motor forearm CV | ≥ ~49–50 m/s | conduction velocity |
| Median F-wave | ≤ ~31 ms (height-dependent) | minimum F latency |
| Ulnar motor (wrist→ADM) | distal latency ≤ ~3.3–3.6 ms | onset latency |
| Ulnar motor CMAP | ≥ ~6 mV | amplitude |
| Ulnar motor forearm CV | ≥ ~49–51 m/s | conduction velocity |
| Ulnar motor across-elbow CV | ≥ ~50 m/s (and < 10 m/s slower than forearm) | conduction velocity |
| Ulnar F-wave | ≤ ~32 ms (height-dependent) | minimum F latency |
| Radial motor (forearm→EIP) | distal latency ≤ ~3.0 ms | onset latency |
| Radial motor CMAP | ≥ ~2–3 mV | amplitude |
| Median sensory (digit II/III, 14 cm) | peak latency ≤ ~3.5 ms | peak latency |
| Median SNAP | ≥ ~20 µV (antidromic) | amplitude |
| Median sensory CV | ≥ ~50 m/s | conduction velocity |
| Ulnar sensory (digit V, 14 cm) | peak latency ≤ ~3.1 ms | peak latency |
| Ulnar SNAP | ≥ ~10–17 µV (antidromic) | amplitude |
| Ulnar sensory CV | ≥ ~50 m/s | conduction velocity |
| Radial sensory (snuffbox→thumb) | peak latency ≤ ~2.7 ms | peak latency |
| Radial SNAP | ≥ ~15–20 µV | amplitude |
Sources for the table: the AANEM Normative Data Task Force literature review (Chen et al. 2016; Dillingham et al. 2016), Buschbacher's reference-value studies, and the value ranges tabulated in Preston & Shapiro and Kimura. [4][5][6][7] Amplitude lower limits in particular vary substantially between montages and references — treat them as orientation, not thresholds.
Condition patterns — what they look like¶
Carpal tunnel syndrome (median neuropathy at the wrist)¶
The earliest and most sensitive abnormality is prolonged median sensory (and palmar/mixed) distal latency — focal slowing across the carpal tunnel — because sensory fibres are affected before motor fibres. As severity increases, the median motor distal latency prolongs (distal latency > ~4.2 ms is supportive), then SNAP and CMAP amplitudes fall as axon loss supervenes. [1][7][9]
The most useful early test is a comparison study that controls for temperature and technique by pitting the median nerve against an adjacent normal nerve over the same distance — typically the median–ulnar (ring-finger sensory or mixed palmar) latency difference or the median–radial (thumb) difference. A difference > ~0.4–0.5 ms is abnormal and raises sensitivity from roughly 75% to ~95%. [7][9] Conventional severity grading runs: mild (abnormal sensory/comparison latency only, normal motor); moderate (abnormal sensory and motor distal latency, preserved amplitudes); severe (the above plus reduced SNAP and/or CMAP amplitude, with needle-EMG denervation/reinnervation in APB). A genuinely normal, well-performed median sensory study and comparison study make clinically significant CTS unlikely. [7][9]
Ulnar neuropathy at the elbow (UNE)¶
Localising the ulnar nerve to the elbow rests on focal slowing or conduction block across the elbow segment, studied with the elbow flexed (~90°) and an adequate measured across-elbow distance (~10 cm). The classic AANEM/AAEM criteria are: across-elbow motor CV reduced below ~50 m/s; an across-elbow segment > 10 m/s slower than the below-elbow (forearm) segment; > ~20% CMAP amplitude/area drop from below-elbow to above-elbow stimulation (conduction block); and a significant change in CMAP configuration (temporal dispersion) above versus below the elbow. [3][10] When routine studies are inconclusive, short-segment incremental stimulation ("inching") at ~1–2 cm intervals across the elbow localises an abrupt latency jump or amplitude drop to the entrapment point and is the most sensitive method. [3][10] Needle EMG of FDP (ulnar head) and FCU helps separate elbow from wrist lesions and grades axon loss.
Cervical radiculopathy¶
The signature is a near-normal NCS with abnormal needle EMG in a myotomal pattern. Because the lesion is proximal to the DRG, the peripheral sensory axon stays in continuity and SNAPs are preserved even when the dermatome is numb — the single most important teaching point, and the discriminator from plexopathy. [1][3][8] Motor studies are usually normal unless there is substantial anterior-root axon loss (then the relevant CMAP may be low). The diagnosis is made on needle EMG: denervation/reinnervation in two or more muscles supplied by the same root but different peripheral nerves, plus, importantly, the cervical paraspinal muscles (innervated by the posterior primary ramus, which leaves the root very proximally — paraspinal abnormality places the lesion at/near the root and is often the earliest finding). [3][8]
Polyneuropathy¶
The typical acquired pattern is length-dependent and symmetric — abnormalities appear first and worst in the longest nerves (lower limb/sural before upper limb), with a distal-to-proximal gradient. Classify on the axonal–demyelinating axis: axonal polyneuropathy shows low-amplitude SNAPs/CMAPs with relatively preserved CV (the common diabetic/toxic/metabolic picture); demyelinating polyneuropathy shows marked slowing, prolonged distal latencies and F-waves, conduction block and temporal dispersion with relatively preserved amplitudes. Uniform slowing suggests hereditary disease; patchy slowing with block suggests acquired inflammatory disease (e.g. CIDP). [1][2][5]
The key cross-cutting teaching point: a preserved SNAP in an anaesthetic dermatome localises the lesion proximal to the DRG (root) and rules out a post-ganglionic (plexus/nerve) cause for that sensory loss — the cleanest single discriminator between radiculopathy and plexopathy. [1][3][8]
Common pitfalls and caveats¶
- Temperature. The commonest source of spurious "slowing." A cool limb prolongs latencies, slows CV and increases amplitude. Maintain and document limb temperature (≥32 °C upper limb); warm a cold hand before believing a borderline-slow median latency. [1][4][5]
- Anomalous innervation — Martin–Gruber anastomosis (MGA). A median-to-ulnar crossover in the forearm (~15–30% of people) produces confusing patterns: an unexpectedly high ulnar CMAP on proximal stimulation, a "too-fast" or initially-positive median forearm response, and apparent ulnar conduction block that is really a crossover artefact. Recognise it before over-calling an elbow lesion. [1][2]
- Volume conduction / co-stimulation. An initial positive deflection of a CMAP, or an unexpectedly large response, may be a distant muscle picked up by volume conduction or inadvertent co-stimulation of an adjacent nerve — not the target. Check montage and stimulus intensity. [1][2]
- Timing after injury. EDX has a built-in delay and must be timed to the question. After an acute axonal injury, motor Wallerian degeneration takes ~7–11 days (and sensory a few days longer) before distal stimulation reveals the full amplitude loss — a study done too early underestimates the lesion or can falsely suggest a (recoverable) conduction block. Fibrillation potentials on needle EMG do not appear until ~2–3 weeks post-denervation (proximal muscles earlier, distal later). For a traumatic nerve injury, a baseline at ~3–4 weeks and a repeat at ~3 months are usually the most informative. [1][2]
- Numbers versus the clinical picture. A value just outside a population range in an asymptomatic territory, or a normal study in a clearly symptomatic patient, must be reconciled with the examination — not reported in isolation. EDX refines and localises a clinical hypothesis; it does not replace it. [1][2]
See Also¶
- nerve tests and conduction studies
- patient: how nerves work and heal
References¶
- Preston DC, Shapiro BE. Electromyography and Neuromuscular Disorders: Clinical–Electrophysiologic–Ultrasound Correlations. 4th ed. Philadelphia: Elsevier; 2021.
- Kimura J. Electrodiagnosis in Diseases of Nerve and Muscle: Principles and Practice. 4th ed. New York: Oxford University Press; 2013.
- Ramani PK, Lui F, Arya K. Nerve Conduction Studies and Electromyography. In: StatPearls. Treasure Island (FL): StatPearls Publishing; 2024. Available from: https://www.ncbi.nlm.nih.gov/books/NBK611987/
- Chen S, Andary M, Buschbacher R, Del Toro D, Smith B, So Y, et al. Electrodiagnostic reference values for upper and lower limb nerve conduction studies in adult populations. Muscle Nerve. 2016;54(3):371–7.
- Dillingham T, Chen S, Andary M, Buschbacher R, Del Toro D, Smith B, et al. Establishing high-quality reference values for nerve conduction studies: a report from the Normative Data Task Force of the American Association of Neuromuscular & Electrodiagnostic Medicine. Muscle Nerve. 2016;54(3):366–70.
- Buschbacher RM, Prahlow ND. Manual of Nerve Conduction Studies. 2nd ed. New York: Demos Medical Publishing; 2006.
- Rosario NB, De Jesus O. Electrodiagnostic Evaluation of Carpal Tunnel Syndrome. In: StatPearls. Treasure Island (FL): StatPearls Publishing; 2024. Available from: https://www.ncbi.nlm.nih.gov/books/NBK562235/
- Dillingham TR. Electrodiagnostic Evaluation of Cervical Radiculopathy. In: StatPearls. Treasure Island (FL): StatPearls Publishing; 2024. Available from: https://www.ncbi.nlm.nih.gov/books/NBK563152/
- American Association of Neuromuscular & Electrodiagnostic Medicine; American Academy of Neurology; American Academy of Physical Medicine and Rehabilitation. Practice parameter for electrodiagnostic studies in carpal tunnel syndrome: summary statement. Muscle Nerve. 2002;25(6):918–22.
- American Association of Electrodiagnostic Medicine, Campbell WW. Practice parameter for electrodiagnostic studies in ulnar neuropathy at the elbow: summary statement. Muscle Nerve. 1999;22(3):408–11.