Characterization of Vertex Sharp Waves: Sleep Stage Distribution Pattern and Implications for Scoring

Highlights

Introduction

Scoring of sleep stages has evolved over the last decades. The establishment of the Rechtschaffen and Kales (R&K) manual in 1968 provided the first standardized criteria for sleep stage scoring, characterizing sleep into distinct stages based on electroencephalography (EEG) patterns. ¹ It wasn’t until 2007 that the American Academy of Sleep Medicine (AASM) updated the scoring rules into the International Classification of Sleep Disorders scoring manual. This manual is updated periodically, and presents detailed rules, terminology and technical specifications for scoring sleep and sleep-related events in polysomnography (PSG) recordings. ² Sleep staging rules have been defined in detail with demonstrative schematic examples, separately for wakefulness and each sleep stage, including non-rapid eye movement (NREM) sleep stages (N1, N2, and N3), and REM sleep. In adults, who generate alpha rhythm, stage N1 is scored when the alpha rhythm is replaced by low amplitude and mixed-frequency (LAMF) EEG activity constituting more than 50% of the epoch.

Vertex sharp waves (VSW) are sharply contoured waves maximal over the central regions with a duration of less than 500 milli-seconds that have been identified since the R&K manual to appear during N1, helping distinguishing N1 from REM sleep, the latter lacking vertex sharp waves and instead developing sawtooth waves [“page 7” in Reference 1]. Although VSW are not required to score stage N1, they are defined as the epiphenomena of the stage N1, and used as the predictive criteria in adults, who do not generate the alpha rhythm. Although VSW are identified as an important EEG marker of a transition between wakefulness and N1, it's use in other sleep stage transition has not been explored. Instead, the presence of LAMF and slow eye movements are used among the rules defining the transitions among sleep stages. For transition into REM sleep, the epochs containing slow eye movements following an arousal interrupting stage R, should be scored as stage N1 sleep, even if chin EMG tone remains equal to that of stage R sleep [“F. Scoring Stage N1” in Reference 2]. In contradistinction to the use of slow eye movements as evidence for stage N1 following an arousal interrupting stage R, the presence of VSW is not a criterion for scoring stage N1. This paradigm seems inconsistent with the equivalency given to slow eye movements and vertex sharp waves in identifying stage N1 in persons without an alpha rhythm. Vertex sharp waves may emerge following an arousal from stage R sleep, even in the absence of any increase in the chin EMG tonus, or slow eye movements. Based on the current criteria, such an epoch is scored as stage R, because the presence of vertex sharp waves is not a criterion for identification of an epoch as stage N1. In this research we analyzed the occurrence of vertex sharp waves in different sleep stages, presenting examples of vertex sharp waves that emerged following an arousal reaction in REM sleep without any change in the chin EMG tonus or slow eye movements in PSG recordings. We hypothesized that vertex sharp waves are sleep stage-specific EEG activities of superficial NREM sleep, and therefore, they may be an indicator of a sleep stage shift to NREM sleep from REM sleep.

Methods

Results

A total of 20 patients (12 men, 60.0%) were analyzed in this study, and the median age of the study population was calculated as 37 years (95% CI 31.3—46.2 years) with a mean of 38.7 ± 14.4 years (ranging between 20–65 years). PSG data of the patients are presented in Table 1. Of these, only 17 patients (85%) had VSW in both N1 and N2 sleep stage, while only 6 patients (30%) had VSW in N3 sleep, and 10 patients (50%) in REM sleep. The number of epochs containing VSW in each sleep stage, the number of EEG channels with VSW, and the index of epochs containing VSW (per total number of epochs in that specific sleep stage) are presented in Table 2.

OPEN IN VIEWER

Table 1.

Polysomnographic Parameters of the Patients.

Polysomnographic Parameters	Median (95% CI)	Mean ± sd
Total recording time (min)	474.4 (463.2–480.4)	471.8 ± 16.7
Total sleep time (min)	413.5 (384.1–422.6)	403.3 ± 37.4
Sleep latency (min)	2.1 (1.8–10.9)	6.4 ± 8.7
REM sleep latency (min)	112.5 (82.0–165.6)	123.8 ± 81.3
Sleep efficiency (%)	86.8 (81.4–89.6)	85.5 ± 7.9
Wakefulness after sleep onset (WASO, min)	63.5 (48.8–88.7)	68.7 ± 38.8
N1 sleep (min)	33.5 (28.9–43.4)	36.2 ± 14.0
N1 sleep (%)	6.9 (6.1–9.1)	7.6 ± 2.8
N2 sleep (min)	227.0 (199.3–257.3)	228.3 ± 56.3
N2 sleep (%)	47.5 (42.2–54.4)	48.3 ± 11.7
N3 sleep (min)	51.5 (43.2–72.8)	58.0 ± 28.7
N3 sleep (%)	11.7 (9.1–15.4)	12.3 ± 6.1
R sleep (min)	80.5 (70.8–90.5)	80.7 ± 19.1
R sleep (%)	18.0 (15.3–19.7)	17.5 ± 4.2
Apnea-hypopnea index (per hour)	3.0 (2.4–7.5)	4.9 ± 5.0
Mean oxygen saturation (%)	96.3 (94.9–96.7)	95.8 ± 1.7
Minimum oxygen saturation (%)	93.0 (86.5–93.3)	89.9 ± 6.6
Index of periodic leg movements in sleep (per hour)	4.8 (3.3–8.7)	6.0 ± 5.2

Data presented as median (95% Confidence interval, [CI]), as well as mean ± standard deviation.

OPEN IN VIEWER

Table 2.

The Number of Epochs Containing vertex Sharp Waves (VSW) in Each Sleep Stage, the Number of Electroencephalography Channels with VSW, and the index of Epochs Containing VSW (per Total Number of Epochs in That Specific Sleep Stage).

Vertex Sharp Wave Analysis	N1 Sleep	N2 Sleep	N3 Sleep	R Sleep
				At onset	Post-arousal	Spontaneous	Total
Duration (min)	33.5 (29.2–43.4) [36.2 ± 14.0]	227.0 (199.3–257.3) [228.3 ± 56.3]	68.7 (42.6–108.9) [58.0 ± 28.7]	80.5 (70.8–90.5) [80.7 ± 19.1]
Total number of epochs	67.0 (58.4–86.9) [72.7 ± 27.7]	454.0 (398.7–514.7) [456.7 ± 112.7]	137.5 (85.3–217.9) [116.1 ± 57.4]	161.0 (141.7–181.1) [161.4 ± 38.2]
Number of epochs with vertex sharps	11.0 (8.8–21.1) [15.0 ± 11.9]	26.0 (22.7–33.9) [28.3 ± 10.9]	1.0 (0.6–2.3) [1.5 ± 0.8]	1.0 (1.0–1.0) [1.0 ± 0.0]	3 (1.0–6.0) [2.6 ± 1.6]	1.0 (1.0–4.0) [1.7 ± 1.1]	3.5 (1.9–5.4) [3.7 ± 2.4]
Number of EEG channels with vertex sharps	130.0 (90.5–239.1) [164.8 ± 144.5]	216.0 (173.3–275.4) [224.3 ± 99.2]	5.0 (2.2–13.0) [7.6 ± 5.1]	12.0 (10–14) [12.0 ± 1.6]	28.0 (8.0–42.0) [27.8 ± 20.2]	16.0 (8.0–38.0) [15.4 ± 11.0]	24.0 (16.5–53.0) [34.8 ± 25.5]
Index of epochs containing vertex sharps*	21.0 (13.2–35.1) [24.1 ± 21.3]	5.3 (4.9–8.3) [6.6 ± 3.2]	1.0 (0.5–1.6) [0.3 ± 0.6]	0.2 [0.2–0.8] [0.2 ± 0.4]	0.8 (0.4–2.1) [0.7 ± 1.0]	0.5 (0.4–1.8) [0.4 ± 0.6]	2.3 (1.2–3.2) [1.3 ± 1.5]

Data presented as median (95% Confidence interval, [CI]), as well as mean ± standard deviation in brackets.

* (Number of epochs with vertex sharps / Number of epochs) × 100.

The mean number of epochs containing vertex sharps was expectedly the highest in N1 sleep stage, being significantly higher than those in N2 (t = 3.622, p < 0.001), N3 (t = 2.344, p = 0.019) and REM sleep (t = 3.527, p < 0.001). The mean number of the EEG-channels that vertex sharps were observed was also highest in N1 sleep in compared to those in N2 (t = 2.628, p = 0.009), N3 (t = 2.201, p = 0.028) and REM sleep (t = 2.803, p = 0.005). Index of epochs containing vertex sharps was highly significantly higher in N1 sleep stage in compared to all other sleep stages (p ≤ 0.001). Index of epochs containing vertex sharps in REM sleep was also significantly lower than those observed in N2 sleep (p = 0.005). Although index of epochs containing vertex sharps in REM sleep was higher than those in N3 sleep, it was not statistically significant (t = 1.753, p = 0.080) (Figure 1).

OPEN IN VIEWER

Figure 1.

The index of epochs containing vertex sharps in N1, N2, N3 and REM sleep stages, calculated as the epochs containing VSW in a particular sleep stage divided by the total number of the epochs of that particular sleep stage (a). The distribution of VSW in REM sleep stage was also shown, depending on the circumstances they arise during REM sleep: at REM onset, following an arousal reaction in REM sleep stage or spontaneously (b).

The total number of the REM sleep epochs containing vertex sharps were 37 epochs; 4 of them (10.8%) were present at the first epoch of REM sleep period. In twelve epochs (32.4%), vertex sharps emerged spontaneously within REM sleep period; while in 21 epochs (56.8%), they emerged following an arousal in the REM sleep. The median number of EEG-channels where vertex sharps were present spontaneously within REM sleep period was 16.0 (95% CI 8.0–38.0) with a mean of 15.4 ± 11.0; and those emerged following an arousal in REM sleep was 28.0 (95% CI 8.0–42.0) with a mean of 27.8 ± 20.2. The number of the EEG-channels containing vertex sharp waves when they occur in REM sleep was higher if they emerged following an arousal in compared to those emerge spontaneously, though the difference didn’t reach to the statistically significant level (t = 1.572, p = 0.116).

The regression analysis showed that the occurrence of VSW in N2 sleep (r² = 0.747, F = 44.366; p < 0.001; Figure 2a) and REM sleep (r² = 0.473, F = 7.174; p = 0.028; Figure 2c) was positively correlated with the presence of VSW in N1 sleep stage. The occurrence of VSW in N3 sleep (r² = 0.076, F = 0.331; p = 0.596; Figure 2b), also show a positive correlation with the VSW in N1 sleep stage, but not significantly. On the other side, the analysis of VSW in REM sleep stage at different time points demonstrated that the positive correlation was only spared for the VSW appearing following an arousal reaction during REM sleep (r² = 0.432, F = 5.329; p = 0.05; Figure 2f), though the correlation was not positive nor significant for the VSW occurring at the onset of REM sleep (r² = 0.002, F = 0.005; p = 0.952; Figure 2d) or for those appearing spontaneously during REM sleep (r² = 0.132, F = 0.759; p = 0.423; Figure 2e).

OPEN IN VIEWER

Figure 2.

The regression analysis of the occurrence of vertex sharp waves (VSW) given in N2 sleep (a), in N3 sleep (b), and in REM sleep (c) in correlation with the presence of VSW in N1 sleep stage. The occurrence of VSW at the onset of REM sleep (d), in REM sleep following an arousal reaction (e), and spontaneously during REM sleep (f) were also demonstrated in correlation with those in N1 sleep stage.

Discussion

Our study has clearly demonstrated that vertex sharp waves are a phenomenon that predominantly occurs in N1 sleep and when occurring in REM it is mostly associated with an arousal as understood by the current arousal scoring rules. The few isolated vertex sharp waves during REM sleep are worth discussion in the setting of arousal rules. Recently with the development of newer technologies, the understanding of cortical and sympathetic arousals has evolved. K complexes have been known for a long time to result from stimuli, while in fact both K-complex and vertex sharp waves have generated as responses to stimuli.^4,5

The importance of VSW was previously identified by Sekine et al, ⁶ who subdivided N1 into N1a (no VSW but still evidence of alpha) and N1b (no alpha and presence of VSW), and postulated that the presence of VSW during N1b as well as evoked during stimuli, can represent a marker of a different awareness level. Furthermore, Stern et al ⁷ studied VSW emergence during drowsiness with Simultaneous EEG and fMRI from seven individuals in drowsiness and light sleep, estimating the neuroanatomical correlation with VSW generation to be the primary sensorimotor cortex. The study showed that despite similar EEG localization, the fMRI regions differ significantly; K-complexes corresponded to decreased signal in the thalamus and increased in the anterior neocortex while sleep spindles corresponded to increased signal in the thalamus. The findings on the cortex may signify that vertex sharp waves may be a sensory response to an external stimulus. Contrary to these findings, fMRI studies during REM sleep have shown activation in regions of the pontine tegmentum, thalamic nuclei, limbic areas (amygdaloidal complexes, hippocampal formation, anterior cingulate cortex) and posterior cortices (temporo-occipital areas). ⁸

The cyclic alternating pattern (CAP), a neurophysiologic sleep pattern including phasic events followed by background EEG activity during NREM sleep, has been considered a marker of sleep instability. ⁹ CAPs are electrocortical events of abrupt frequency shifts or amplitude changes that are clearly distinguishable from the background EEG rhythm. It consists of cyclic alternation of high-voltage slow waves (phase A) followed by low-voltage irregular activity (phase B); and vertex sharp transients have been identified as a part of phase A, as K-complex sequences, delta and polyphasic bursts. The understanding of CAP emphasizes the presence of different patterns of arousal, like cortical or subcortical arousals, electroencephalographic, autonomic or behavioral arousals or partial arousals. Considering the link between VSWs and CAP, the results of our study support that epochs with vertex sharp waves occurring in stage R movements should also be scored as stage N1 regardless of chin EMG tonus or slow eye movement.

Further support to this proposal is found in studies with event related evoked potentials. Averaged EEG responses to specific stimuli given during sleep were demonstrated to be unique to NREM sleep. The vast majority of studies have focused on the P2, N350, N500 and P900 potentials in response to auditory stimuli or respiratory related evoked potentials, and characteristic waveforms or components were defined to be related to evoked K-complexes and vertex sharp waves.^4,5,10 The relationship between the generation of the N300 and N550 responses and K-complexes was established, while these two potentials were shown to be independent with difference in scalp topographies—N550 was maximum over the fronto-central regions, and N300 was maximum over the vertex. As the localization of N300 was often across the midline, and maximal at the vertex region, it was suggested that N300 was closely related to the vertex sharp waves.^5,6,11 This data was later replicated in other studies using N300 responses to not only auditory but also respiratory stimuli during sleep,^12–14 which produced a sleep specific potential maximal at the vertex. The amplitude of N300 was shown to increase dramatically at sleep onset, and suggested to be related to the emergence of theta activity in N1 stage. It was therefore suggested that N300 potential was a multi-model response (to different modalities of stimuli) specific to early sleep onset during theta activity.

N300 potential, together with N550 potential, were linked to neural generators functional during NREM sleep.^5,6,13,14 At the initiation of sleep, theta activity in stage N1 was suggested to enable processes that were responsible from both N300 and vertex sharp waves. With evolving sleep in stage N2, following the activation of N300, processes that were responsible from N550 and K-complexes were generated. As N300 was best elicited in studies with back-averaging EEG to produce vertex sharp waves, which were also maximal on central regions and could be elicited by external stimuli, they were suggested to reflect the synchronized activation of the neural circuits from the same generator.^5,6 Although the functional significance of these sleep-specific potentials is still under discussion, they were linked to the sensory processing of psychological or biological events and arousal mechanisms with a decrease in thalamocortical synchronization.^4,15 Considering scalp topographies, timing and components of the waves, N300 was hypothesized to be analogue of the vertex sharp waves, which may emerge as isolated phasic events or as bursts of repetitive waves in the presence of theta activity.^4,6

In summary, neuroimaging, electrophysiology, and our current polysomnography study consistently indicate that vertex sharp waves are markers of NREM sleep and when appearing in REM they indicate a sleep stage shift. By revealing distinct neural activity patterns typically associated with NREM generation, these waves could be crucial to identify microarchitectural changes in sleep, otherwise not identified by the current scoring rules. Although the results of this study are compelling and intriguing, further studies with a larger sample is warranted to support and/or delineate the role of vertex sharp waves emerging in different sleep stages.

Supplemental Material

Supplemental Material

Supplemental Material

Supplemental Material