Amplitude of Somatosensory Evoked Potentials (SEPs) Recorded in Short-Latency SEP Condition Is 80% of That in Giant SEP Condition

Abstract

Purpose:

Giant somatosensory evoked potentials (SEPs) with enhanced long-loop reflex (C-reflex) are useful to detect cortical motor hyperexcitability in patients with myoclonic epilepsy. The recording conditions of giant SEPs are different from those of short-latency SEPs (SSEPs). We investigated the waveform characteristics obtained for each condition.

Methods:

Forty-eight upper limbs of 24 adult normal subjects (12 men, age 35.5 ± 9.7 years [mean ± SD]) were investigated. Somatosensory evoked potentials of each subject were recorded in both conditions on the same day. The main differences in recording conditions were reference electrodes (SSEP: Fz vs. giant SEP: the earlobe electrode ipsilateral to the stimulated limb), stimulus rate (5 vs. 1 Hz), and bandpass filter (20 Hz–3 kHz vs. 1 Hz–1 kHz). Somatosensory evoked potentials were elicited by unilateral percutaneous electrical stimulation of the median nerve at the wrist with intensity of 110% of the movement threshold and recoded at C3'/C4'.

Results:

The amplitudes of N20 onset–N20 and N20–P25 were significantly larger in giant SEP condition than in SSEP condition (p < 0.001). The mean + 3SD of N20–P25 amplitude was 10.0 μV in giant SEP condition and 7.8 μV in SSEP condition. The N20–P25 amplitude was significantly correlated between giant SEP condition and SSEP condition (R = 0.64, p < 0.001). C-reflex was not elicited.

Conclusions:

The amplitude of SEPs in SSEP condition is equivalent to 80% of that in giant SEP condition. The information is useful for detecting cortical hyperexcitability in various neurological disorders including myoclonic epilepsy.

Somatosensory evoked potentials (SEPs) are used to evaluate the sensory pathway through the peripheral sensory nerve, spinal cord, and cortical primary somatosensory area.¹ Increased amplitude of cortical components (giant SEPs) with enhanced long-loop reflex (C-reflex) is observed in epilepsy syndrome with cortical myoclonus, reflecting the hyperexcitability of the somatosensory areas.^2–9 In epilepsy syndrome with cortical myoclonus, evaluation of giant SEPs is useful for the diagnosis and management of the disease. The recording condition of giant SEPs^5,7,8 is different from that of short-latency SEPs (SSEPs).¹⁰

We investigated the correlation of waveform characteristics obtained for each condition to provide useful information for clinical practice. Part of this article was presented in the Asian and Oceanian Epilepsy Congress in 2021, Fukuoka, Japan, in an abstract form.

MATERIAL AND METHODS

Participants

A total of 24 healthy adult subjects (12 men, age 35.5 ± 9.7 years [mean ± SD]) without any neurological diseases were included. Their 48 upper limbs were evaluated. All participants provided written informed consent based on the research protocol approved by the Bioethics Committee of National Hospital Organization Utano National Hospital. All experiments of this study were performed in accordance with the Standards for Reporting of Diagnostic Accuracy guidelines and the Declaration of Helsinki. The study population is the same as the control group in our previous publication.¹¹

SEP Recording

In each subject, SEPs were recorded in both SSEP condition and giant SEP condition^5,7,10 in a shielded room on the same day (Fig. 1). In both conditions, SEPs were elicited by unilateral percutaneous electrical stimulation of the median nerve at the wrist with intensity of 110% of the movement threshold. Recording electrodes were Ag/AgCl disc electrodes placed on the scalp contralateral to the stimulus side at 7 cm left/right and 2 cm posterior to Cz (C3′/C4′). Impedance of electrodes was kept below 5 kΩ using conductive paste. The differences in recording conditions were reference electrodes (SSEP: Fz vs. giant SEP: the earlobe electrode ipsilateral to the stimulated limb), bandpass filter (20 Hz–3 kHz vs. 1 Hz–1 kHz), analysis times (100 ms vs. 200 ms), average number (300–500 stim. vs. 50–100 stim), and stimulus rate (5 vs. 1 Hz). In giant SEP condition, the C-reflex was recorded by a pair of disc electrodes placed on the both abductor pollicis brevis muscles, and the subjects were instructed to keep their hands relaxed. The waveform parameters were determined by agreement of board-certified clinical neurophysiologists (A.D., Y.D., and M.K.).

FIG. 1.

Green box: Nomenclature and amplitude measurement of somatosensory evoked potentials (SEPs). N20: Negative (upward) peak with a latency of approximately 20 ms. P25: Positive (downward) peak with a latency of approximately 25 ms. N20o: The onset of N20. Recording conditions differ between short-latency SEPs (SSEPs) (yellow box) and giant SEPs (pink box). Blue box: Recording conditions common to both SEPs.

Statistics

The variables of N20–P25 amplitude on giant SEP condition did not exhibit a normal distribution as per the Kolmogorov–Smirnov test; thus, the results of parametric tests were confirmed by nonparametric tests. To obtain the most suitable model to represent the relationship of waveform measurements between SSEP condition and giant SEP condition, 12 mathematical models were checked by curve-fitting analysis to correlate values of giant SEP condition against values of SSEP condition. All analyses were performed using SPSS statistical software (version 27; IBM Japan, Tokyo, Japan). The significance level was set at p = 0.05 for group comparison and p = 0.01 for correlation analyses.

RESULTS

Comparison of SEP Amplitudes

Nomenclature and amplitude measurements of SEPs are shown in Fig. 1. N20 and P25 are defined as negative and positive peaks of waveform, and N20o is defined as the onset of N20. The amplitudes of N20o–N20 and N20–P25 were significantly larger in giant SEP condition than in SSEP condition (p < 0.001; Table 1, Fig. 2). The mean + 3SD of N20–P25 amplitude was 10.0 μV in giant SEP condition and 7.8 μV in SSEP condition. The C-reflex was not elicited. Women showed longer latency of P25 and N30 and higher amplitudes of P25–N30 than did men (Table 2). The latency of N20o and N20 was shorter with stimulation on the dominant side than on the nondominant side (Table 3).

TABLE 1.

Amplitude Measures of Each Condition (μV)

	Giant SEP Condition	SSEP Condition	Ratio*	P †
N20o–N20
Mean ± SD	4.2 ± 1.8	1.9 ± 0.8	45%	P < 0.001
Mean + 2SD	7.8	3.5	45%
Mean + 3SD	9.6	4.3	45%
N20–P25
Mean ± SD	4.6 ± 1.8	3.5 ± 1.4	76%	P < 0.001‡
Mean + 2SD	8.2	6.4	78%
Mean + 3SD	10.0	7.8	78%
P25–N30
Mean ± SD	3.0 ± 1.8	3.0 ± 1.5	100%	n.s.
Mean + 2SD	6.6	6.0	91%
Mean + 3SD	8.4	7.5	89%

^*Ratio of amplitude (SSEP/giant SEP conditions).

^†Paired t test.

^‡Confirmed by the Wilcoxon signed-rank test (p < 0.001).

SEP, somatosensory evoked potential; SSEP, short-latency SEP; n.s., not significant.

FIG. 2.

Upper row: Representative waveforms of a 23-year-old woman, recorded by stimulating the left median nerve at the wrist. The amplitudes of N20o–N20 and N20–P25 are larger on giant somatosensory evoked potential (SEP) condition than on short-latency SEP (SSEP) condition. Lower row: Scatter plots of SEP amplitudes recorded in giant SEP condition (ordinate) against that in SSEP condition (abscissa), showing significant linear correlation of N20–P25 amplitude. Significance was confirmed by using the Spearman rank correlation (ρ = 0.666, p < 0.001).

TABLE 2.

Group Comparison of Waveform Measures Between Men and Women (Mean ± SD)

	Male	Female	P *
Age, years	37.0 ± 11.4	33.9 ± 7.4	n.s.
SSEP condition, latency/height, msec/m
N20o	9.08 ± 0.44	9.12 ± 0.54	n.s.
N20†	11.12 ± 0.35	11.22 ± 0.45	n.s.
P25‡	13.53 ± 1.20	13.88 ± 1.54	n.s.
N30	18.62 ± 1.30	19.87 ± 1.77	0.009
SSEP condition, amplitude, μV
N20o–N20	1.88 ± 0.90	1.95 ± 0.71	n.s.
N20–P25	3.38 ± 1.47	3.63 ± 1.41	n.s.
P25–N30	2.57 ± 1.51	3.48 ± 1.31	0.033
Giant SEP condition, latency/height, msec/m
N20o	8.30 ± 0.69	8.00 ± 0.53	n.s.
N20	11.18 ± 0.52	11.21 ± 0.76	n.s.
P25	14.26 ± 1.11	15.05 ± 1.10	0.020
N30	18.59 ± 2.11	19.67 ± 1.36	0.044
Giant SEP condition, amplitude, μV
N20o–N20	4.05 ± 1.54	4.44 ± 1.92	n.s.
N20–P25	4.42 ± 1.79	4.83 ± 1.80	n.s.§
P25–N30	2.89 ± 1.58	3.09 ± 2.01	n.s.

^*Unpaired t test.

^§Confirmed by the Mann–Whitney U test (n.s.).

^†Significantly correlated with age (b: R = 0.543, p < 0.001).

^‡Significantly correlated with age (c: R = 0.556, p < 0.001).

SEP, somatosensory evoked potential; SSEP, short-latency SEP; n.s., not significant.

TABLE 3.

Pairwise Comparison of Waveform Measures Between Dominant and Nondominant Side Stimulation (Mean ± SD)^*

	Dominant	Nondominant	P †
SSEP condition, latency/height, msec/m
N20o	8.96 ± 0.42	9.24 ± 0.52	0.016
N20	11.18 ± 0.45	11.16 ± 0.36	n.s.
P25	13.83 ± 1.42	13.57 ± 1.35	n.s.
N30	19.34 ± 1.78	19.15 ± 1.54	n.s.
SSEP condition, amplitude, μV
N20o–N20	2.01 ± 0.89	1.82 ± 0.71	n.s.
N20–P25	3.46 ± 1.28	3.55 ± 1.60	n.s.
P25–N30	2.90 ± 1.29	3.15 ± 1.65	n.s.
Giant SEP condition, latency/height, msec/m
N20o	8.15 ± 0.54	8.15 ± 0.71	n.s.
N20	11.02 ± 0.66	11.37 ± 0.60	0.019
P25	14.64 ± 1.19	14.67 ± 1.15	n.s.
N30	19.49 ± 2.12	18.77 ± 1.47	n.s.
Giant SEP condition, amplitude, μV
N20o–N20	4.14 ± 1.71	4.35 ± 1.79	n.s.
N20–P25	4.89 ± 1.96	4.37 ± 1.59	n.s.‡
P25–N30	3.11 ± 1.64	2.87 ± 1.96	n.s.

^†Paired t test.

^*There was one left-handed male subject, and his data were adjusted; there were no ambidextrous subjects.

^‡Confirmed by the Wilcoxon signed-rank test (n.s.).

SEP, somatosensory evoked potential; SSEP, short-latency SEP; n.s., not significant.

Correlation Analysis of N20–P25 Amplitude Between Recording Conditions

The N20–P25 amplitude showed a significant linear correlation between giant SEP condition and SSEP condition (R = 0.64, p < 0.001). As for curve-fit analysis, the models used were linear, logarithmic, inverse, quadratic, cubic, compound, power, sigmoid, growth, exponential, and logistic. The growth model (Y = e^{0.795 + 0.188X}) was more suitable than the linear model (Y = 1.827 + 0.799X) (R = 0.66 vs. R = 0.64). The growth model is the same as the linear model with natural logarithmic transformation of dependent variables (ln(Y) = 0.795 + 0.188X). About 10 µV for giant SEP condition was equivalent to 8.0 µV for SSEP condition (Fig. 3). The power model (Y = 2.298 × X^0.542) also showed the highest correlation coefficient (R = 0.69, p < 0.001).

FIG. 3.

Scatter plots of N20–P25 amplitude recorded in giant somatosensory evoked potential (SEP) condition (ordinate) against that in short-latency SEP (SSEP) condition (abscissa). Left: Curve-fit analysis, showing that growth (green) and power (red) models are more suitable than linear model (black). Right: Natural logarithmic transformation of N20–P25 amplitude on giant SEP condition (ordinate). According to linear model (blue), 10 µV on giant SEP condition, i.e., ln(10) = 2.3, is equivalent to 8.0 µV on SSEP condition (dotted lines).

DISCUSSION

Comparison of SEP Amplitudes Between SSEP Condition and Giant SEP Condition

To the best of our knowledge, this is the first study to compare waveform measurements of SEPs between SSEP condition and giant SEP condition. The current study showed that N20o–N20 and N20–P25 amplitudes were significantly different between SSEP condition and giant SEP condition. This is most likely due to the fact that Fz, the reference electrode of SSEP condition, is influenced by cortical SEPs.^12,13 In addition, a previous report described that the amplitude of the middle latency cortical component after N20 decreases with use of a high bandpass filter and high stimulus rate.^11,14

Correlation Analysis of N20–P25 Amplitude Between SSEP Condition and Giant SEP Condition

In this study, the N20–P25 amplitude recorded in SSEP condition was equivalent to 78% of that in giant SEP condition. The initial prototype of giant SEP recording condition was described in 1977, which used a logarithmic scale and considered deviation beyond 2.38SD from the mean value of the normal control group as abnormal.² Based on the data, the same group determined an SEP as “giant” when N20–P25 amplitude was larger than 8.6 μV or P25–N33 amplitude was larger than 8.4 μV.^2,3 In 1990s, the upper normal limit was set at the mean + 3SD of the logarithmic values recorded from normal subjects, i.e., N20–P25 amplitude larger than 6.3 μV and P25–N35 amplitude larger than 9.8 μV.^5,6 More recently, the mean + 3SD of amplitudes obtained from the control subjects without logarithmic transformation are used, N20–P25 > 10.0 μV or P25–N35 > 8.1 μV for the younger subgroup and N20–P25 > 20.0 μV or P25–N35 > 14.8 μV for the older subgroup.⁸ In general, it is currently accepted that an amplitude more than 10 μV has high diagnostic significance in cortical myoclonus and epilepsy.⁹ Our findings suggest that N20–P25 amplitude larger than 8 μV in SSEP condition could be considered as giant SEPs. In the posterior tibial nerve stimulation, the upper limit was set at the mean + 2SD⁴ or +3SD⁷ of normal subjects. We previously evaluated a diagnostic validity of giant evoked potentials using different upper limits, i.e., the mean + 2SD or the mean +3SD, in genetically proven benign adult familial myoclonic epilepsy.¹¹ Cumulative data would enable us to determine the optimum reference range in accordance with the purpose of evaluation and the tentative diagnosis.

Clinical Implication of Increased Amplitude of SEPs

The increased amplitude of SEPs can be seen in patients with other central nervous system disorders, such as cerebrovascular disease, demyelinating disease, spinal cord disease, and hydrocephalus.¹⁵ Thus, it can be useful in patients with suspected central nervous system pathologies. In patients with amyotrophic lateral sclerosis, a report showed that an SEP amplitude of more than 8 μV in SSEP condition is a predictive factor for poor prognosis.¹⁶ The amplitude corresponds to 10 μV for giant SEP condition, as shown in our study. Abnormal enhancement of cortical excitability can be associated with worsening of neurodegeneration, and evoked potential amplitude can be a surrogate marker for disease progression. A recent study of conventional SEP recording which set the upper limit at the mean + 2.5SD, giant SEPs were found in 6.6% of patients evaluated in neurology department.¹⁷ It should be emphasized that a tentative diagnosis of functional disorders would need to be reconsidered if amplitude of SEPs was enlarged.

Utility of Analyses of Relationship Among Recording Conditions

More than 40 years have passed since clinical relevance of giant SEPs was described,² and recent literatures are showing that enlarged SEPs reflect abnormal cortical excitability in various neurological disorders in addition to epilepsy syndromes with cortical myoclonus.^15,16 Our data could allow clinical neurophysiologists to begin with screening SEP evaluation using SSEP condition and switch to giant SEP condition only when indicated. For further investigation including C-reflex, giant SEP condition would be necessary.

In this study, we used Fz as a reference electrode in SSEP condition in accord with the official guidelines on evoked potentials by the Japanese Society of Clinical Neurophysiology.¹⁸ The derivation is robust to noise and suitable for inexperienced technologists or uncooperative patients. The guidelines by the American Clinical Neurophysiology Society recommend the derivation of contralateral CP (i.e., midway between C and P) to ipsilateral CP to the stimulated limb.¹⁹ This derivation is also robust to noise and free from cortical potential generated in areas other than the primary somatosensory cortex, but can be affected by giant SEPs on the other hemisphere with delayed latency in patients with progressive myoclonus epilepsy.²⁰ Recording electrodes of SSEPs are conventionally placed at 7 cm left/right and 2 cm posterior to Cz (C3′/C4′) in Japan following Shagass and Schwartz.²¹ The same placement was used in giant SEP condition of our study. By contrast, in previous studies, C3/C4 and CP3/CP4 are used and analyzed as a whole.²² Effect of these minute differences on SEP waveforms should be assessed in the future.

Limitations

The most important limitation of this study is small sample size. However, in previous studies, control data were determined from approximately 20 normal subjects.^2,5,8 Large studies across a wide range of age groups would be useful to confirm the correlation between the two conditions.

CONCLUSIONS

In conclusion, our study showed that SEP amplitude recorded in SSEP condition was equivalent to approximately 80% of that in giant SEP condition. Further studies, including various neurological disorders, are warranted to evaluate the diagnostic significance and the predictive value for prognosis.