By Stephen J. Fried,* Diane M. Smith,* and Alan D. Legatt*†

Summary: Median nerve somatosensory evoked potential (SSEP) monitoring is commonly used during carotid endarterectomy to permit selective shunting in only those patients who are determined to have inadequate collateral flow after carotid cross-clamping. The N20 component is recorded from the CPc (contralateral centroparietal) electrode; either CPi (ipsilateral centroparietal) or Fpz (forehead) can be used as the reference. Because of the distribution of the subcortically generated N18 component, the CPc-Fpz derivation might record both the N20 and the N18 components and might therefore inadequately detect hemispheric ischemia after carotid cross-clamping. Somatosensory evoked potentials (SSEP) recorded were compared using these 2 derivations during 38 carotid endarterectomies to assess their ability to detect neurophysiologic changes after carotid cross-clamping. Although, as expected, the baseline N20 component was significantly larger when recorded with the CPc-Fpz derivation than with the CPc-CPi derivation (3.1 vs. 2.4 mV in the hemisphere ipsilateral to the clamped carotid, P , 0.001), there was no significant difference in the postclamp amplitude decline between the 2 derivations (8.7% vs. 8.6%, P ¼ 0.82). It is concluded that CPc-Fpz is an acceptable derivation for recording postclamp hemispheric somatosensory evoked potential (SSEP) changes during carotid endarterectomy and may be advantageous because it provides a larger amplitude somatosensory evoked potential (SSEP) than the CPc–CPi derivation.

Key Words: N20, Median nerve SEP, Somatosensory evoked potentials Carotid endarterectomy, Intraoperative monitoring, Reference electrode.

(J Clin Neurophysiol 2014;31: 55–57)

Intraoperative neurophysiologic monitoring is commonly used during carotid endarterectomy. In most cases (between 80% and 90%), the collateral circulation is sufficient to provide blood flow to both hemispheres during carotid clamping (Ballotta et al., 2010; Rerkasem and Rothwell, 2009). However, in cases where significant hemispheric ischemia is demonstrated during clamping, a shunt can be placed to restore flow to the distal carotid artery. The role of monitoring is to determine which patients require shunting because routinely placing shunts in patients who do not require them may lead to an increased risk of stroke (Salvian et al., 1997; Woodworth et al., 2007).

Intraoperative monitoring during the procedure may be performed in several ways. One technique is to perform the surgery under local anesthetic block with the patient awake with strength testing performed during the procedure to monitor for hemispheric ischemia. EEG can also be used, with the monitoring team assessing either raw or processed EEG to identify hemispheric asymmetries that are suggestive of ischemia. Another modality is monitoring of somatosensory evoked potentials (SEPs) to median nerve stimulation. In particular, a 50% decline in amplitude of the SEPs generated in the hemisphere ipsilateral to the cross-clamped carotid is used as a marker of ischemia requiring shunting. Median nerve SEPs have several advantages, namely, that they can be performed in patients who cannot tolerate an awake procedure and that their interpretation is less subjective than that of EEG. Other techniques include near infrared spectrometry for cerebral oximetry, transcranial Doppler for measurement of middle cerebral artery blood flow, and measurement of the carotid stump pressure.

The N20 is a negative potential that is recorded over the centroparietal scalp, contralateral to the stimulated median nerve. It is a near-field potential, generated from the hand area of the primary somatosensory cortex. More specifically, the cortical SEPs generated after median nerve stimulation consist of a tangentially oriented N20-P30 component, generated in Brodmann area 3b, on the posterior bank of the central sulcus, and a radially oriented P25- N35 component, generated in Brodmann area 1 (Allison et al., 1991).

The origin of the N18 component is controversial, and it may have more than one source (Sonoo, 2000). Structures ranging from the lower medulla (nucleus cuneatus) to the thalamus have been proposed as N18 generators; however, based on intracranial recordings in humans, the most likely origin is in the rostral pons or the mesencephalon (Philips et al., 1998). Owing to the fact that its generators are deep and close to the midline, the N18 component has a fairly widespread and symmetrical distribution over the dorsal scalp (Desmedt and Cheron, 1981). Thus, in a CPc-CPi (centroparietal contralateral minus centroparietal ipilateral to the stimulated median nerve) derivation, the N18 will be markedly attenuated by in-phase cancellation, and this derivation will record a more “pure” N20 (American Clinical Neurophysiology Society, 2006). However, if the field of the N18 does not extend all the way to the forehead, this component would not be canceled in a CPc-Fpz (centroparietal contralateral to median nerve stimulation minus forehead) derivation and that derivation should therefore yield a combined N20-N18 peak that is larger than the N20 alone.

Because the N18 is generated by regions supplied by the posterior circulation, it should not be affected by clamping of the carotid artery. As such, the CPc-Fpz derivation, which contains an N20 peak that may be augmented by the N18 component, may less clearly demonstrate ischemia caused by carotid cross-clamping (Sonoo, 2000; Sonoo et al., 1992). However, there may be advantages to the CPc-Fpz derivation, in that it might yield a high-amplitude waveform that requires less averaging (Lee and Seyal, 1998).

In this study, we compared the SEPs recorded by the CPc-Fpz and the CPc-CPi derivations during intraoperative monitoring for carotid endarterectomy. We specifically looked at the decline in amplitude occurring after cross-clamping to see if there is a significant difference in the ability to detect hemispheric changes between the two reference choices.


We retrospectively reviewed the intraoperative monitoring studies for carotid endarterectomies that were performed at the Montefiore Medical Center over a 2-year period. In each study, both the CPc-Fpz and the CPc-CPi derivations were included in the recording montage for the median nerve cortical SEPs. In total, 38 carotid endarterectomies were included in the analysis.

The SEP data were recorded using Protektor evoked potential recording systems (XLTek Corporation, Oakville, Ontario, Canada). The median nerves were stimulated at a rate of 6.1 Hz. Data were filtered with a bandpass of 5 to 3,000 Hz for the cortical SEP recording channels. Impedances were maintained less than 2.5 kU. After automatic artifact rejection (raw data rejection threshold set at 95 mV), SEP waveforms with at least 250 sweeps per average were recorded. The left and right median nerves were tested using alternating stimuli. The N20 amplitude was measured as the voltage difference between the N20 and following positive trough.

The N20 component was measured bilaterally using both the CPc-Fpz and the CPc-CPi channels. The P14 components, generated at the level of the cervicomedullary junction, were also monitored during the procedure, mainly to serve as a control. During monitoring, SEP runs were typically performed every 2 to 4 minutes, and surgeons were alerted when there were significant adverse SEP changes, defined as a 50% decrease in the N20 amplitude from baseline. In the analysis for this study, the average N20 amplitudes across all runs performed within the last 15 minutes before carotid cross-clamping, and across those performed within the first 15 minutes after cross-clamping, were calculated for both median nerves and for both recording derivations. The median nerve SEPs recorded in the hemisphere ipsilateral to the clamped carotid artery (i.e., from median nerve stimulation contralateral to the clamped carotid) were designated the “active” side. The SEPs recorded in the hemisphere contralateral to the clamped carotid artery (i.e., from median nerve stimulation ipsilateral to the clamped carotid) were designated the “inactive” side.

The preclamp SEP amplitudes were calculated for both sides and for both reference choices. The percent change in the amplitudes was also calculated for both sides and for both reference choices. These datasets were compared with each other using the Wilcoxon signed rank test.


A total of 38 subjects were included in the analysis. The surgery was on the left carotid artery in 18 subjects and on the right carotid artery in 20 subjects. Twenty-eight of the patients had symptomatic stenosis and 10 had asymptomatic high-grade stenosis. The baseline preclamp N20 amplitudes were significantly higher using the Fpz reference than using the CPi reference on both the active side (3.1 vs. 2.4 mV, P , 0.001) and inactive side (4.1 vs. 3.1 mV, P , 0.001). Additionally, the baseline preclamp amplitudes on the active side were significantly smaller than those on the inactive side, using both the Fpz (3.1 vs. 4.1 mV, P , 0.01) and the CPi (2.4 vs. 3.1 mV, P , 0.01) references.

After clamping, there was typically a decline in the amplitudes of the SEPs on both the active side (in 31 of 38 cases with the Fpz reference and 30 of 38 cases with the CPi reference) and the inactive side (in 28 of 38 cases with the Fpz reference and 27 of 38 cases with the CPi reference). The difference in the postclamp decline between the active and the inactive sides was significant for both the Fpz (8.7% vs. 3.5% decline, P ¼ 0.014) and CPi references (8.6% vs. 2.3% decline, P ¼ 0.013). However, there was no significant difference between the Fpz and CPi references in recording the postclamp decline on the active side (8.7% vs. 8.6%, P ¼ 0.82).

In 2 operations, the surgeons were alerted to a decline in the SEP amplitude to less than 50% of baseline after carotid crossclamping. In both cases, the decline was detectable simultaneously with both the Fpz and CPi references. In one of the cases, the SEP changes resolved with an increase in the patient’s blood pressure. In the other case, the SEPs recovered only after placement of a carotid– carotid shunt (Fig. 1).


The average amplitude of the N20 recorded with the Fpz reference was significantly higher than that recorded with the CPi reference on both sides (29% larger on the active side and 32% larger on the inactive side). This indeed fits the idea that the CPc-CPi derivation serves to subtract out the N18 component and thereby record a more “pure” N20; the superimposed N18 in the CPc-Fpz derivation increases the apparent N20 amplitude. However, it was hypothesized that the SEP recorded with the CPc-Fpz derivation, augmented approximately 30% by the posterior circulation subserved N18, would therefore show less of an amplitude decline after carotid cross-clamping. This was not the case, however, both in general and in the two cases in which the surgeons were alerted to a significant amplitude decline.

Why did the CPc-Fpz waveform show the same attenuation as the CPc-Cpi waveform after anterior circulation ischemia? One possibility is that the field of the N18 extended to the forehead electrode, as suggested by the data of Desmedt et al. (1987), so that N18 was to a large extent canceled in both derivations. But, if that was the case, why was the CPc-Fpz waveform larger? One possibility is the dipole field of the N20 component and its inversion, P20. Since N20 is generated in the posterior bank of the central sulcus, it has a tangential dipole that is at right angles to the central sulcus at the hand area, which usually projects a simultaneous positivity, P20, anteriorly (Legatt and Kader, 2000). The field of the P20 typically extends to the forehead (Desmedt et al., 1987; Okada et al., 1996; Tsuji and Murai, 1986). The presence of an inverted N20 at the Fpz electrode would increase the apparent amplitude of the parietal N20 in the CPc-Fpz derivation.

Because the N20 recorded with the CPc-Fpz derivation is, at baseline, significantly larger, this derivation may, in fact, have some advantages for monitoring carotid endarterectomies. However, it should be pointed out that this may not always be the case because Fpz is more likely than CPi to contain EMG artifact when the anesthetic depth is light. One way to resolve this could be for the anesthesiologists to provide muscle relaxation to minimize such artifact. A CPc-CPi channel is still useful, however, and should be recorded in case an individual patient’s N18 distribution is such that the CPc-Fpz channel does contain a substantial N18 contribution. Because of that possibility, a marked attenuation of the N20 amplitude in the CPc-CPi channel should be regarded as a significant adverse change even if the attenuation of the N20 amplitude in the CPc-Fpz channel does not reach the alarm limits. The increased number of channels in modern evoked potential monitoring systems should permit monitoring of both of these SEP channels simultaneously.

When monitoring upper limb SEPs during spine surgery, it is obvious that either derivation is acceptable, as the N18 and N20 components are both generated rostral to the regions at risk. However, our results are most likely not only applicable for monitoring of carotid endarterectomy but also other operations involving risk to regions subserved by the anterior circulation. In summary, we conclude that CPc-Fpz is an acceptable derivation for recording postclamp hemispheric changes during carotid endarterectomy and may be advantageous because it records a larger amplitude SEP than does the CPc-CPi derivation.


Allison T, Wood CC, McCarthy G, Spencer DD. Cortical somatosensory evoked potentials. II. Effects of excision of somatosensory or motor cortex in humans and monkeys. J Neurophysiol 1991;66:64–82. American Clinical Neurophysiology Society. Guideline 9D: guidelines on shortlatency somatosensory evoked potentials. J Clin Neurophysiol 2006;23:168– 179. Ballotta E, Saladini M, Gruppo M, et al. Predictors of electroencephalographic changes needing shunting during carotid endarterectomy. Ann Vasc Surg 2010;24:1045–1052. Desmedt JE, Cheron G. Non-cephalic reference recording of early somatosensory potentials to finger stimulation in adult or aging normal man: differentiation of widespread N18 and contralateral N20 from the prerolandic P22 and N30 components. Electroencephalogr Clin Neurophysiol 1981;52:553–570. Desmedt JE, Nguyen TH, Bourguet M. Bit-mapped color imaging of human evoked potentials with reference to the N20, P22, P27, and N30 somatosensory responses. Electroencephalogr Clin Neurophysiol 1987;68:1–19. Lee EK, Seyal M. Generators of short latency human somatosensory evoked potentials recorded over the spine and scalp. J Clin Neurophysiol 1998;15: 227–234. Legatt AD, Kader A. Topography of the initial cortical component of the median nerve somatosensory evoked potential. Relationship to central sulcus anatomy. J Clin Neurophysiol 2000;17:321–325. Okada J, Shichijo F, Matsumoto K, Kinouchi Y. Variation of frontal P20 potential due to rotation of the N20-P20 dipole moment of SEPs. Brain Topogr 1996;8:223–228. Philips M, Kotapka M, Patterson T, et al. Brainstem origins of the N18 component of the somatosensory evoked response. Skull Base Surg 1998;8:133–140. Rerkasem K, Rothwell PM. Routine or selective carotid artery shunting for carotid endarterectomy (and different methods of monitoring in selective shunting). Cochrane Database Syst Rev 2009;CD000190. Salvian AJ, Taylor DC, Hsiang YN, et al. Selective shunting with EEG monitoring is safer than routine shunting for carotid endarterectomy. Cardiovasc Surg 1997;5:481–485. Sonoo M. Anatomic origin and clinical application of the widespread N18 potential in median nerve somatosensory evoked potentials. J Clin Neurophysiol 2000;17:258–268. Sonoo M, Genba K, Zai W, et al. Origin of the widespread N18 in median nerve SEP. Electroencephalogr Clin Neurophysiol 1992;84:418–425. Tsuji S, Murai Y. Scalp topography and distribution of cortical somatosensory evoked potentials to median nerve stimulation. Electroencephalogr Clin Neurophysiol 1986;65:429–439. Woodworth GF, McGirt MJ, Than KD, et al. Selective versus routine intraoperative shunting during carotid endarterectomy: a multivariate outcome analysis. Neurosurgery 2007;61:1170–1176.

From the Departments of *Neurology and †Neuroscience, Montefiore Medical Center and the Albert Einstein College of Medicine, Bronx, New York, U.S.A. This study was presented at the 2013 annual meeting of the American Clinical Neurophysiology Society. Address correspondence and reprint requests to Alan D. Legatt, Department of Neurology, Montefiore Medical Center, 111 East 210 Street, Bronx, NY 10467, U.S.A.; e-mail: alegatt@montefiore.org. Copyright  2014 by the American Clinical Neurophysiology Society ISSN: 0736-0258/14/3101-0055 Journal of Clinical Neurophysiology Volume 31, Number 1, February 2014 55.