Somatosensory and motor evoked potentials are not two views of the same thing. They follow different tracts, in different vascular territories, with different alarm criteria — and only one of those criteria is the tidy rule everyone quotes.
SSEPs follow proprioception and vibration up the ipsilateral dorsal column, which sits in the posterior spinal artery territory. MEPs follow the corticospinal tract, which sits in the anterior spinal artery territory covering two-thirds to four-fifths of the cord. A normal SSEP therefore tells you nothing reassuring about the motor tracts, and the alarm criteria differ in kind: sensory potentials have the familiar 50% amplitude or 10% latency rule, while motor potentials have a consensus range for amplitude and no predictive latency criterion at all.
Evoked potentials are used for two different jobs. Mapping stimulates and records to identify where a tract or structure is, or what an unknown structure does. Monitoring repeats a stimulus through a case and watches for change in amplitude or latency. This day is about monitoring, and about the fact that the two common modalities are not interchangeable.
The somatosensory evoked potential follows the pathway of proprioception and vibration, ascending the ipsilateral dorsal column. That is a posterior structure, supplied posteriorly.
The motor evoked potential follows the corticospinal tract, which lies in the territory of the anterior spinal artery — a vascular network supplying the anterior two-thirds to four-fifths of the cord. The anterior spinal artery is fed by radicular arteries, especially those between T8 and T12, which combine to form the artery of Adamkiewicz; that vessel provides about 75% of the perfusion to the anterior cord in the thoracic, lumbar and sacral regions.
Two further asymmetries follow from that geography. The grey matter and anterior horn cells are more sensitive to ischaemia than white matter. And the motor pathways are more sensitive than sensory pathways to insults from stretch, compression, vascular disruption or direct trauma, because of synaptic transmission within the cord and less redundancy of perfusion.
So the modality that watches the larger territory is also the modality watching the more vulnerable tissue, fed by a supply that a single radicular artery can dominate. That is why combining the two improves detection over SSEP alone, and why MEP monitoring is advocated for all axial skeletal deformity procedures.
For sensory evoked potentials the general principle is that an amplitude reduction of 50% or a latency increase of 10% is significant and raises concern — with the caveat, stated in the same sentence, that smaller changes may indicate impending compromise. The order matters when you say it out loud: amplitude first, because that is usually what moves first.
For motor evoked potentials the picture is different in kind, not just in number. Consensus is that MEP loss is of concern and that an amplitude decrement of 50 to 70% is concerning. Other changes that cause concern:
And on latency: there is no predictive data of latency change, because many benign events lengthen it. A 10% increase without a physiologic cause — such as hypothermia or an electrolyte abnormality — is often reported, but reporting a change is not the same as it predicting anything.
For intramedullary spinal tumour surgery, where the D wave is monitored directly from the cord, surgical teams often halt excision when the D wave amplitude falls by 50%. Note that this is a third criterion for a third signal, which is a fair summary of this field.
A 12,373-patient series across 25 years found MEP changes in 3.1% of primary spine operations and 6.1% of revisions. Intervention reversed the change in 87% of them, and those whose signals did not recover had permanent neurologic change. Where MEPs recovered completely by closure, the rate of injury was zero — which is the finding that makes the intraoperative response worth taking seriously rather than treating a change as a prognostic announcement.
On false negatives, systematic review puts MEP alone at 0 to 0.79% and SSEP alone at 0.63 to 2.7%. In one large study, the sensitivity of each modality independently predicting permanent postoperative change was MEP 100%, spontaneous EMG 14%, SSEP 28.6%. The incidence of major motor injury with no SSEP warning has been reported at 0.063% — small, and not zero, and concentrated in exactly the tract SSEPs do not watch.
An American Academy of Neurology and American Clinical Neurophysiology Society review confirmed the efficacy of MEP monitoring in spine surgery, and a separate guideline concluded there is robust class I evidence for combined SSEP and MEP recording as a diagnostic adjunct. None of that is a randomized demonstration that monitoring improves outcome, and the value likely varies with the particular operation and region of spine.
Because evoked potentials cross a wide range of neural tissue and synapse types, inhalational agents have the most profound effect of any drug class. Muscle responses to transcranial stimulation are especially depressed. Among the volatiles, isoflurane is more potent than halothane or enflurane in depressing neurological function, and sevoflurane and desflurane are broadly similar to each other. Nitrous oxide reduces amplitude and increases latency in cortical sensory and motor responses, alone or in combination. In patients with pre-existing neurological dysfunction, particularly older adults, consensus supports avoiding volatile agents altogether.
Intravenous agents suppress MEPs much less, so total intravenous anaesthesia is preferred when MEPs are monitored, with propofol the usual hypnotic component. Two agents run the other way and are worth knowing for that reason: ketamine and etomidate increase the amplitude of cortical sensory and motor responses, which is why they appear as rescue adjuncts when signals are marginal.
Finally, neuromuscular blockade. Partial blockade reduces MEP amplitude, and muscles vary in their sensitivity to it — so the extent of blockade has to be assessed in the muscles being monitored. The corollary from the spine chapter applies here with force: never titrate against a plegic limb, because denervated muscle is resistant to non-depolarizing agents and monitoring it will lead you to overdose the patient.
No randomized trial addresses this.
That monitoring detects injury is well supported. An evidence-based review by the American Academy of Neurology and the American Clinical Neurophysiology Society confirmed the efficacy of motor evoked potentials in spine surgery, and a separate guideline concluded there is robust class I evidence supporting combined SSEP and MEP recording during spinal cord and spinal column surgery as a diagnostic adjunct for assessing cord integrity. The value likely varies with the specific operation and the region of spine involved.
The alarm criteria are consensus, and the two modalities are not symmetrical. For sensory potentials the general principle is that an amplitude reduction of 50% or a latency increase of 10% is significant — with the explicit caveat that smaller changes may indicate impending compromise. For motor potentials, consensus is that loss is of concern and that an amplitude decrement of 50 to 70% is concerning; other concerning changes include reduced complexity, needing to raise stimulation above 50 V, or any change in stimulus settings needed to hold the signal. On latency the source is unambiguous: there is no predictive data of latency change, because many benign events lengthen it. A 10% increase without a physiologic explanation is often reported, and that is a reporting convention rather than a validated threshold.
What has not been established is that monitoring changes outcome in a randomized comparison. The evidence base is large cohorts, systematic reviews and guideline consensus. It is strong enough that monitoring is advocated for all axial skeletal deformity procedures, and it is not a randomized demonstration of benefit.
Strength, in the guideline's own words Described in the source as "robust class I medical evidence supporting monitoring with both SSEP and MEP recordings during spinal cord/spinal column surgery as a diagnostic adjunct to assess spinal cord integrity."
Almost everyone can recite the alarm criterion: 50% amplitude reduction or 10% latency increase. Almost nobody attaches it to the right modality.
It is a sensory evoked potential principle. Applied to motor potentials it is wrong in a specific and traceable way: the source states plainly that there is no predictive data of latency change for MEPs, because many benign events lengthen MEP latency. The 10% figure circulates for motor potentials as a reporting convention — a change of that size without a physiologic explanation is often reported — and it has never been shown to predict anything.
On the amplitude side the mismatch is subtler. Consensus for MEPs is that a decrement of 50 to 70% is concerning. That is a range, and the range is the finding. It is worth resisting the pull to collapse it to a single number, because a resident who has been taught "50%" will treat 55% as an alarm and a resident taught "70%" will not, and neither is what the consensus says.
The failure this prevents is not a missed alarm. It is a confident, quick, wrong answer when someone senior asks what the criterion is — followed by a conversation in which the actual signal, and what to do about it, never gets discussed. When the MEP amplitude falls, the more useful things to say are what the pressure has been doing, whether the stimulus settings have been changed to keep the signal, and whether the complexity of the response has dropped. All three are on the consensus list and none of them is a percentage.
Full text · Preston goes through LibKey to whatever Preston holds — the PDF where there is one — and works on a phone. The doi link always reaches the publisher's record, whether or not we hold a subscription. If you are reading on a computer, LibKey Nomad adds the same one-click PDF button to PubMed and publisher pages everywhere else; it is a desktop browser extension, so it does nothing on a phone or tablet.
During correction of a thoracic deformity, motor evoked potentials fall markedly in both legs while somatosensory evoked potentials remain unchanged. What best explains this?
Which statement about intraoperative alarm criteria is correct?
Ninety minutes into a posterior thoracolumbar fusion, motor evoked potential amplitudes fall by about 60% bilaterally. Blood pressure has drifted down over the preceding fifteen minutes. What is the best first response?
Bilateral proximal, arm-worse-than-leg weakness is the classic border-zone presentation, sitting on the seam between the anterior and middle cerebral territories where flow is pressure-passive once vasodilation is exhausted. Reviewing the actual pressure trace is the first useful step. Note what this does not do: it does not exclude embolism, because a bed with no reserve is also a bed that cannot wash an embolus through.
Under normal conditions pial flow does not cross into the adjacent territory because the pressures on either side of the boundary are equal. Change one side — a stenosis, a fall in pressure, an embolization — and the boundary moves. That is why the anatomic location varies considerably between people and can change during the course of treatment.