Skip to main content
NCBI home page
The Journal of Physiology logoLink to The Journal of Physiology
. 1999 Nov 1;520(Pt 3):897–908. doi: 10.1111/j.1469-7793.1999.00897.x

Effect of wake–sleep transitions and rapid eye movement sleep on pharyngeal muscle response to negative pressure in humans

Steven A Shea 1, Jill K Edwards 1, David P White 1
PMCID: PMC2269629  PMID: 10545152

Abstract

  1. Genioglossus (GG) activation in response to upper airway negative pressure may be an important mechanism in the maintenance of airway patency. This reflex occurs during wakefulness but is diminished during stable non-rapid eye movement (NREM) sleep. Since obstructive events occur more commonly at wake–sleep transitions and during rapid eye movement (REM) sleep than during stable NREM sleep, we assessed the GG reflex during these two vulnerable states.

  2. Seventeen healthy adults were studied throughout one evening and overnight. Electroencephalograms (EEGs), electro-oculograms (EOGs), submental electromyogram (EMG), GG EMG (intramuscular electrodes), and choanal plus epiglottic pressures were recorded. The GG reflex response to pulses of −8 cmH2O choanal pressure applied via nose mask during early inspiration was quantified repeatedly during relaxed wakefulness, within five breaths of wake–sleep transition (EEG alpha-theta transition) and during REM sleep. Only trials without EEG arousal were analysed, resulting in data from 14 subjects during sleep onset and 10 subjects during REM sleep (overall, 174–491 trials per state).

  3. During wakefulness there was brisk GG reflex activation in response to negative pressure (amplitude: +78.5 ± 28.3 % baseline (mean ± s.e.m.); latency to maximal response: 177 ± 16 ms).

  4. At sleep onset, although there was marked variability among individuals, there was no significant reduction in the magnitude of the GG reflex for the group as a whole (amplitude: +33.2 ± 8.2 % baseline; latency: 159 ± 15 ms).

  5. In contrast, during REM sleep there was a reduction of GG reflex (amplitude: −12.6 ± 8.3 % baseline (P = 0.017 vs. awake); latency: 160 ± 10 ms (n.s. vs. awake)) and greater airway collapsibility during the applied pressures (P = 0.043 vs. awake).

  6. We conclude that there was no systematic reduction in the GG reflex to negative pressure at sleep onset. Nonetheless, it remains possible that sleep-deprived normal subjects and patients with sleep apnoea could react differently.

  7. The apparent inhibition of the GG reflex during REM sleep may help explain why the upper airway is vulnerable to collapse during this state.

Obstructive sleep apnoea syndrome (OSA) is a common and debilitating disorder characterised by sleep-induced increases in pharyngeal resistance and upper airway collapse. A combination of inadequate airway anatomy and loss of upper airway dilator muscle activity during sleep are probably the principal causes of OSA (e.g. White, 1995). The genioglossus muscle (GG) is phasically active with the breathing cycle, having greater activity during inspiration which can serve to dilate or stabilise the airway. In humans, imposed upper airway negative pressure reflexively activates the GG during wakefulness, which is thought to be an important mechanism in the maintenance of pharyngeal patency (Horner et al. 1991). It has also been suggested that this negative pressure reflex may drive the augmented upper airway dilator muscle activity observed in apnoea patients during wakefulness, thereby serving to compensate for the diminished upper airway patency in these individuals (Mezzanotte et al. 1992).

This negative pressure reflex is greatly diminished during stable NREM sleep (Wheatley et al. 1993a; Horner et al. 1994). Hence, loss of the reflex during NREM sleep could help explain the diminished muscle activity that leads to obstructive apnoeas. However, it is well known that obstructive apnoeas occur much more commonly at the wake–sleep transition and during REM sleep than during stable NREM sleep (e.g. Krieger & Kurtz, 1978). Indeed, obstructive apnoeas prevent the occurrence of stable NREM sleep as such apnoeas are usually terminated by arousal. Therefore, we designed this protocol to assess the magnitude of the GG response to negative pressure during the first few breaths of sleep onset (alpha-theta transition detected by electroencephalography) and during REM sleep. We have tested the hypotheses that the magnitude of the GG reflex response to negative pressure will be diminished during these two vulnerable states, when compared to the magnitude of the GG reflex that occurs during relaxed wakefulness.

METHODS

Subjects

We studied 17 healthy volunteers with sufficient data being collected in 14 subjects (10 men, 4 women: mean age, 28.2 years; range, 20–37 years: mean body mass index, 23.1; range, 18.5–26.3). None of the subjects had symptoms of neurological, cardiovascular or pulmonary disorders, sleep-related breathing disorders or snoring. The protocol was approved by the Human Subjects Committee at Brigham & Women's Hospital. All subjects provided informed written consent prior to participation in this study.

Measurements and recording

For detection of wakefulness and sleep stages, subjects were instrumented with three channels of electroencephalography (EEG; C3/A2; C4/A1 and C4/O2), two channels of electro-oculography (EOG; LOC/A1 and ROC/A1) and a submental electromyogram (EMG), with sleep being staged according to standard criteria (Rechtschaffen & Kales, 1968).

To assess the function of a representative upper airway dilator muscle, the activity of the GG EMG was measured using two 36 gauge Teflon-coated stainless steel intramuscular wires. Each wire was passed through a 25 gauge needle. Each needle was inserted into the floor of the mouth at a location 3–5 mm on either side of the frenulum and 15–20 mm into the body of the GG near its insertion in the mandible. After insertion, the needles were extracted, leaving the intramuscular electrodes in place. These wires were referred to a ground electrode on the forehead. The EMG signal was amplified (Grass Model 7P122G; filter characteristics 50 Hz-5 kHz), rectified and ‘integrated’ on a moving time average basis with a 100 ms time constant (Model MA-821-4; CWE Incorporated).

To assess airway collapsibility and the magnitude of the applied negative pressure stimulus, pressures were recorded in the airway at the level of the choanae and in the hypoglossal airspace at the level of the epiglottis (Millar catheters). Before insertion of the catheters into the nose, one nostril was decongested (2–3 inhalations of 0.05 % oxymetazoline hydrochloride (Afrin)) and anaesthetised (approximately 1 ml of 4 % lidocaine (lignocaine) hydrochloride topical spray). After placement, both catheters were taped to the nose to ensure stability. In almost all subjects there was occasional loss of an upper airway pressure signal during the study (probably due to build-up of secretions on the Millar catheters). When this occurred the subject was awakened for repositioning of these catheters or removal, cleaning and re-insertion of these catheters.

To assess inspiratory airflow and for application of negative pressure stimuli, subjects breathed through a sealed nose mask (Healthdyne Technologies; dead space approximately 50 ml). To detect expiratory leaks, carbon dioxide was monitored around the periphery of the mask. The subjects also were monitored via closed-circuit television using a low-light camera (Panasonic; Model WV-CU-101) to ensure that the mouth was closed during recordings. Most subjects naturally breathed entirely through the nose during both wakefulness and sleep, although in a few subjects it was necessary to apply a chin strap to ensure nasal breathing. Inspiratory flow was measured with a pneumotachometer (Fleisch no. 2) and pressure transducer (Validyne differential amplifier; ± 2 cmH2O).

All signals were recorded on a 16-channel polygraph (Grass Model 78E; paper speed, 10 mm s−1). Airway pressures plus the rectified, integrated GG EMG were also digitised at 1000 Hz for off-line computer analysis (‘Sigavg’ software and CED 1401 digitiser; Cambridge Electronic Design, UK).

Protocol

Subjects were studied overnight. Measurements were made during stable relaxed wakefulness with eyes open, during repeated sleep onset transitions and during REM sleep (both phasic and tonic). Measurements during wakefulness were taken either immediately prior to sleep (8 subjects) or immediately after sleep (6 subjects). Collection of sleep transition data generally occurred in the first half of the night. To collect sufficient sleep transition data (i.e. within 2–5 breaths of an alpha-theta transition detected by EEG), subjects were awakened by an experimenter after each sleep-onset measurement (usually, an experimenter would enter the subject's room and speak to the subject). Subjects were then allowed to fall back to sleep after approximately 1 min of documented wakefulness. REM data were collected generally in the second half of the night. As data in the three states were commonly collected at different times of the night, wakefulness data were collected before sleep as well as after sleep in 5 subjects to enable an additional assessment of the effect of elapsed time on measurements within a single state. In addition, since subjects had their eyes open only during the wakefulness data, and it is possible that opening the eyes could affect the results (Shea et al. 1987), a further control study was performed in 10 subjects during relaxed wakefulness with eyes closed.

All measurements were made with the subject in the same posture (confirmed by closed-circuit television monitoring). Subjects lay on their preferred side, with a pillow against the back to ensure that they did not change posture during sleep. A lateral position was selected to diminish the chance of snoring or upper airway collapse during sleep.

The GG reflex was assessed during each state by the application of a negative upper airway pressure pulse having characteristics that have previously been shown to reliably activate the GG in healthy awake subjects (Wheatley et al. 1993a; b; White et al. 1998). Each stimulus was applied in early inspiration by manual activation of a solenoid valve (MarcValve Corporation, Tewksbury, MA, USA). When activated, this valve connected the inspiratory tubing to a vacuum source and resulted in a brief pulse of negative pressure to the upper airway. The time from onset of pressure generation to maximal pressure was on average 77 ms. Thereafter the pressure decayed exponentially to baseline within 340 ms. The target pressure for each pressure pulse was −8 cmH2O at the level of the choanae during each state. These stimuli were applied at irregular intervals; every 2–7 breaths during wakefulness and REM sleep, and within 2–5 breaths of an EEG alpha-theta transition in the sleep onset state. Between 20 and 100 negative pressure pulses were applied to each subject within each state. Examples of the GG reflex response to applied pressure pulses are shown for one subject during relaxed wakefulness in Fig. 1.

Figure 1. Examples of reflex GG EMG activation to applied upper airway negative pressure during relaxed wakefulness in Subject 1.

Figure 1

Open in a new tab

Shown from top to bottom are EEG (C4/O2); L-EOG (left eye outer canthus/A1), R-EOG (right eye outer canthus/A1), raw submental EMG from surface electrodes on the chin (EMG (chin)), raw GG EMG (intramuscular), rectified and integrated GG EMG (downwards deflection: increased GG EMG), choanal and epiglottic pressures (upwards deflections: increased negative pressure; these channels have the same gain) and inspiratory airflow (upwards deflection: increased flow). Two pressure pulses were applied, as indicated by the rapid negative pressure deflections occurring at the beginning of inspiration on the second and sixth breaths. During normal inspirations the epiglottic pressure is slightly more negative than the choanal pressure as negative pressure is generated by the respiratory pump muscles. During the applied negative pressure pulses, the choanal pressure is slightly more negative than the epiglottic pressure (index of pharyngeal collapsibility). Both pressure pulses produced rapid increases in the GG EMG which can be seen both on the raw and integrated GG EMG channels.

Data analysis

All unusual breaths, such as swallows and coughs, were excluded from analysis. Any negative pressure pulses applied within two breaths of such a disturbance also were excluded. The polysomnographic chart records of the EEG channels were visually scrutinised during alpha-theta transitions and during REM sleep to detect any signs of arousal in response to the negative pressure pulses (EEG arousal was defined as a noticeable increase in EEG frequency lasting at least 0.5 s). Trials were excluded from analysis if there was an EEG arousal within 1.0 s of the stimulus onset. Data collected during REM sleep were further segregated into tonic REM (no eye movements) and phasic REM (eye movement within 1 s of the breath on which the stimulus occurred).

Time-appropriate signal-averaged waveforms were produced for the integrated GG EMG and for the pressure signals for each individual within each state (‘Sigavg’ software, Cambridge Electronic Design). Time ‘zero’ was considered to be the onset of the negative pressure stimulus at the choanae. Following the onset of the stimulus, the time to the first noticeable change in GG EMG and the time to the maximal GG EMG response were both used as measures of response latency. Only changes within 200 ms of the stimulus onset were analysed in order to avoid volitional responses to the stimulus. Since the baseline level of EMG at time ‘zero’ can be different between individuals, the GG data were normalised to baseline (100 %) before quantifying the amplitude of the GG reflex and before averaging among subjects. From the normalised individuals' data the peak GG EMGs within 200 ms of the onset of the stimulus were measured. For meaningful comparisons of the GG reflex between states it was necessary to compare both the baseline EMG and the percentage change from the baseline during pressure pulse applications. Thus, for comparisons between relaxed wakefulness and the other states, baseline EMG levels were quantified as a percentage of the baseline during relaxed wakefulness, and the size of the reflex was quantified as the percentage change from baseline within each state (i.e. 100 × (peak EMG – baseline EMG)/baseline EMG) (White et al. 1998).

The fall in pressure between the choanae and the epiglottis during the pressure pulse was quantified during each state as an index of airway collapsibility, as previously described (Wheatley et al. 1993a). This was accomplished by subtracting the peak epiglottic negative pressure from the peak choanal negative pressure on the signal-averaged waveform.

Electromyographic and airway collapsibility responses to negative pressure were compared between relaxed wakefulness and the other states using Wilcoxon matched-pairs signed-ranks tests for between-group analyses and Student's unpaired t tests for within-subject comparisons. Since there were different numbers of subjects in each state, each paired comparison between states is presented separately. The mean response of each individual within each state was used in statistical analyses and when averaging data for producing figures, such that each subject provided equal weighting to the analyses.

RESULTS

Number of successful upper airway negative pressure pulse trials per state

The total number of negative pressure trials during each state are presented in Table 1. Almost all trials were acceptable during relaxed wakefulness. Pressure stimuli caused detectable EEG arousal during REM sleep on only 30 % of occasions, and these were excluded from analyses. However, pressure stimuli often caused detectable EEG arousal during alpha-theta transitions, resulting in rejection of 85 % of the trials.

Table 1.

Number of quantifications of GG EMG reflex response to upper airway negative pressure stimuli during different states

Awake (n = 14) α-θ (n = 14) REM (n = 10)
Attempted trials per state (total) 616 492 622
Successful trials per state (total) 484 174 435
Successful trials per subject (median) 34 10 46
Successful trials per subject (range) 23–54 1–44 14–92
Open in a new tab

α–θ: within 2–5 breaths of EEG alpha–theta transition. Criteria for successful trial: not within 2 breaths of a swallow or movement; additional criterion during sleep: no EEG arousal (arousal defined as increased frequency of EEG within 1 s of stimulus onset).

There were similar numbers of trials for each subject in the data collected for comparison of relaxed wakefulness before sleep with relaxed wakefulness after sleep (n = 5 subjects: pre-sleep; range 30–43 (median 34) trials per subject: post-sleep; range 30–43 (median 39) trials per subject), as well as for the comparison between relaxed wakefulness with eyes open and relaxed wakefulness with eyes closed (n = 10 subjects: awake eyes open; range 25–54 (median 36.5) trials per subject: awake eyes closed; range 9–82 (median 37.5) trials per subject).

Comparison between relaxed wakefulness and alpha-theta transitions

Individuals' responses

Overall, the GG response to negative pressure was largely consistent among subjects when awake (GG activation) and during REM sleep (GG inactivation), but these responses were quite variable among subjects during alpha-theta transitions. An example of this pattern of response is shown in Fig. 2 that depicts the mean integrated EMG waveforms during the three states for two individuals. Both subjects had a brisk GG EMG activation induced by negative pressure during wakefulness, and both showed a reduction in the response during REM sleep. However, during alpha-theta transitions the GG reflex was diminished in one subject and increased in the other subject in this example. This variability among subjects was further investigated by an analysis of the changes in reflex responsiveness between wakefulness and sleep onset for a number of individuals. For this post hoc analysis, we selected those individuals with at least ten successful trials in the sleep onset state (without EEG arousal). This resulted in analysis in 7 subjects of 400 individual pressure pulses (262 during wakefulness, range 30–54 (median 35) trials; and 138 during sleep, range 10–44 (median 14) trials). In these seven subjects, the peak GG EMG during the pressure stimulus was significantly increased at sleep onset in 3 subjects (mean changes: 26 %, 77 % and 120 %), significantly decreased at sleep onset in 3 subjects (mean changes: −14 %, −29 % and −58 %) and unchanged at sleep onset in one subject (mean change 6 %). Identical results were obtained if the magnitude of the GG reflex was expressed as the absolute change from baseline within each state. These results show that individuals can have significant differences in GG reflex between wakefulness and sleep onset.

Figure 2. Example of variability among subjects in GG EMG reflex response to negative pressure.

Figure 2

Open in a new tab

Two individuals' mean integrated GG EMG waveforms during relaxed wakefulness, sleep onset (2–5 breaths after EEG alpha-theta transitions) and REM sleep (tonic plus phasic REM combined). There were between 11 and 54 breaths averaged within each subject within each state. For each individual the GG EMG is normalised to the baseline activity before pressure stimuli during wakefulness (100 %). The vertical dashed line on each plot represents the onset of the pressure stimulus at the choanae. Subject 1 (top) had a brisk and large-amplitude GG EMG activation induced by negative pressure during wakefulness, a diminished GG reflex during sleep onset and a further decrement during REM sleep. Subject 2 (bottom) also had a brisk GG EMG activation during wakefulness but this reflex had a smaller amplitude relative to baseline. At sleep onset Subject 2 had an increased baseline GG EMG and no decrement in the GG reflex (compared to wakefulness), and a ‘reversal’ of the reflex during REM sleep. Data were most variable among subjects at sleep onset.

Group response

The group mean integrated EMG and choanal plus epiglottic pressure waveforms during relaxed wakefulness and alpha-theta transitions are shown in Fig. 3 and the statistical comparisons are presented in Table 2. Note, small differences occur between the values in Table 2 and the impression from Fig. 3 because the table was derived from the analysis of individual subjects' waveforms, whereas the group average waveform (Fig. 3) loses definition as peaks and troughs occur at slightly different times among individuals. It can be seen that during relaxed wakefulness, −8 cmH2O in the upper airway caused a large increase in GG activity that occurred quickly enough to rule out the possibility that it was a voluntary reaction.

Figure 3. Effect of sleep onset on the GG EMG reflex response to negative pressure.

Figure 3

Open in a new tab

The panels show the group (n = 14) mean integrated GG EMG waveform (top) and choanal and epiglottic pressures (bottom) during relaxed wakefulness and immediately following sleep onset (within 2–5 breaths of EEG alpha-theta transition). The GG EMG signal is normalised during both states to the baseline activity before pressure stimuli during wakefulness (100 %). The vertical dashed line on each plot represents the onset of the pressure stimulus in the choanae. The fall in pressure between the choanae and the epiglottis during the pressure pulse was used as an index of airway collapsibility. It can be seen that during wakefulness there was a brisk genioglossus EMG activation induced by negative pressure. During alpha-theta transition the baseline GG EMG increased by approximately 40 % relative to the baseline during wakefulness, and the change in GG EMG induced by the pressure stimulus appeared smaller during alpha-theta transition. However, due to variability among subjects(see Fig 2) none of these differences reached statistical significance (Table 2). There was a similar degree of upper airway collapsibility during wakefulness and alpha-theta transition.

Table 2.

Comparison of negative pressure stimuli, genioglossal reflex response to negative pressure stimuli and airway collapsibility during wakefulness and sleep onset

Awake α–θ Probability
Maximal pressure stimulus in choanae (cmH2O) −8.44 ± 0.37 −8.08 ± 0.35 P = 0.178
Latency from stimulus onset to maximal stimulus (ms) 81 ± 5 82 ± 6 P = 0.855
Latency from stimulus onset to initial GG response (ms) 43 ± 6 47 ± 9 P = 0.524
Latency from stimulus onset to maximal GG response (ms) 177 ± 16 159 ± 15 P = 0.510
Baseline EMG at time ‘zero’ (normalised to % of awake state) 100 ± 0 139.1 ± 23.7 P = 0.510
Maximal GG response (% change from baseline in each state) 62.4 ± 25.8 33.2 ± 8.2 P = 0.433
Collapsibility (choanal minus epiglottic pressure; cmH2O) 1.66 ± 0.37 2.28 ± 0.63 P = 0.859
Open in a new tab

Means ±s.e.m.; n = 14 subjects. α–θ: within 2–5 breaths of alpha-theta transition. Probability: chance that the null hypothesis is false from Wilcoxon matched-pairs signed-ranks test of Awake vs. α–θ.

We were not able to demonstrate systematic group changes in the GG EMG reflex or airway collapsibility to applied upper airway negative pressure between relaxed wakefulness and the immediate sleep onset transitions. During alpha-theta transitions the mean baseline GG EMG was increased, and the same pressure stimulus induced a smaller GG activation than occurred during wakefulness. However, this difference was not significant when the size of the reflex was expressed as the peak GG EMG level, as the percentage change from baseline when awake, or even when expressed as the percentage change from baseline within each state. The response during alpha-theta transitions had a similar time course to the response during wakefulness (Table 2). Finally, our index of airway collapsibility (the difference between choanal and epiglottic pressure during the pressure pulse) was not significantly increased during alpha-theta transitions when compared to relaxed wakefulness.

Group comparison between relaxed wakefulness and REM sleep

A representative individual's mean integrated EMG waveforms are depicted in Fig. 4 during relaxed wakefulness and REM sleep (tonic REM, phasic REM and phasic plus tonic REM combined). It can be seen that during wakefulness, this subject had a brisk GG EMG activation induced by negative pressure. However, this response was initially diminished and then reversed during REM sleep. In comparison to the baseline activity during wakefulness, the GG EMG at time ‘zero’ was slightly higher during tonic REM and lower during phasic REM, with baseline GG EMG being similar to wakefulness when tonic and phasic REM were combined. In this example, and in the other subjects, despite slight differences in the baseline EMG during tonic and phasic REM, the responses to negative pressure during these two states were qualitatively similar.

Figure 4. Example of effect of phasic and tonic REM sleep on the GG EMG reflex response to negative pressure.

Figure 4

Open in a new tab

A representative individual's mean integrated GG EMG waveforms during relaxed wakefulness (mean of 35 stimuli) and REM sleep (tonic plus phasic REM combined (37 stimuli), tonic REM (28 stimuli), and phasic REM (9 stimuli)). In each plot the GG EMG signal is normalised to the baseline activity before pressure stimuli during wakefulness (100 %). The vertical dashed line on each plot represents the onset of the pressure stimuli in the choanae. It can be seen that during wakefulness, the subject had a brisk GG EMG activation induced by negative pressure, but this response was initially diminished and then reversed during REM sleep (phasic or tonic).

The group mean integrated EMG and choanal plus epiglottic pressure waveforms during relaxed wakefulness and REM sleep (phasic plus tonic REM combined) are shown in Fig. 5 and the derived mean data are presented in Table 3. Note, small differences occur between the values in Table 3 and the impression from Fig. 5 because the table was derived from analysis of individual subjects' waveforms, whereas the group average waveform loses definition as peaks and troughs occur at slightly different times among individuals. It can be seen that during relaxed wakefulness, −8 cmH2O in the upper airway caused a rapid large increase in GG activity. During REM sleep the same pressure stimulus induced an initial small GG activation (+14.2 ± 4.7 % (s.e.m.) of pre-stimulus baseline) which followed the same time course as the response during wakefulness (latency 42 ms). However, this response peaked before the occurrence of maximum airway pressure. This rapid small activation was immediately followed by a larger reduction of GG activity, which occurred at approximately the same time (160 ms) as the maximal GG reflex activation when awake. For the group, the decrease in GG EMG was statistically significant when compared to the GG activation during wakefulness. In addition, our index of airway collapsibility (the difference between choanal and epiglottic pressure during the pressure pulse) was significantly greater during REM sleep than during wakefulness.

Figure 5. Effect of REM sleep on the GG EMG reflex response to negative pressure.

Figure 5

Open in a new tab

The panels show the group (n = 10) mean integrated GG EMG waveform (top) and choanal and epiglottic pressures (bottom) during relaxed wakefulness and REM sleep (phasic plus tonic REM combined). The GG EMG signal is normalised during both states to the baseline activity before pressure stimuli during wakefulness (100 %). The vertical dashed line on each plot represents the onset of the pressure stimuli in the choanae. The fall in pressure between the choanae and the epiglottis during the pressure pulse was used as an index of airway collapsibility. It can be seen that during wakefulness there was a brisk GG EMG activation induced by negative pressure, and that this response was reversed during REM sleep. There was also greater upper airway collapsibility during REM sleep.

Table 3.

Comparison of negative pressure stimuli, genioglossal reflex response to negative pressure stimuli and airway collapsibility during wakefulness and REM sleep

Awake REM Probability
Maximal pressure stimulus in choanae (cmH2O) −8.13 ± 0.51 −8.17 ± 0.76 P = 0.879
Latency from stimulus onset to maximal stimulus (ms) 78 ± 6 76 ± 8 P = 0.647
Latency from stimulus onset to initial GG response (ms) 41 ± 4 42 ± 6 P = 0.999
Latency from stimulus onset to maximal GG response (ms) 152 ± 10 160 ± 10 P = 0.576
Baseline EMG at time ‘zero’ (normalised to % of awake state) 100 ± 0 103 ± 32 P = 0.333
Maximal GG response (% change from baseline in each state) 57.6 ± 33.3 −12.6 ± 8.3 P = 0.017*
Collapsibility (choanal minus epiglottic pressure; cmH2O) 1.83 ± 0.45 5.46 ± 0.96 P = 0.043*
Open in a new tab

Means ±s.e.m.; n = 10 subjects. REM = tonic + phasic data. Probability: chance that the null hypothesis is false from Wilcoxon matched-pairs signed-ranks test of Awake vs. REM.

*

Statistical significance (P < 0.05).

Effect of time of night and effect of closing the eyes upon measurements

There were no statistically significant differences in either GG activation or airway collapsibility responses to applied negative pressure pulses during relaxed wakefulness when recorded at the beginning or at the end of the night (n = 5; Table 4) or when relaxed wakefulness data collected with the eyes open and with eyes closed were compared (n = 10; Table 5). In addition, all comparisons between relaxed wakefulness with eyes closed and the two sleep states yielded exactly the same results (using the statistical cut-off of α = 0.05) as those comparisons between relaxed wakefulness with eyes open and the two sleep states presented above.

Table 4.

Comparison of negative pressure stimuli, genioglossal reflex response to negative pressure stimuli and airway collapsibility during wakefulness recorded at the beginning and at the end of the night

Pre-sleep Post-sleep Probability
Maximal pressure stimulus in choanae (cmH2O) −9.22 ± 0.34 −9.22 ± 0.38 P = 0.746
Latency from stimulus onset to maximal stimulus (ms) 90 ± 4 98 ± 5 P = 0.545
Latency from stimulus onset to initial GG response (ms) 42 ± 18 46 ± 13 P = 0.500
Latency from stimulus onset to maximal GG response (ms) 163 ± 16 172 ± 18 P = 0.625
Baseline EMG at time ‘zero’ (normalised to % of pre-sleep) 100 ± 0 108 ± 17 P = 0.625
Maximal GG response (% change from baseline in each state) 101 ± 23 84 ± 32 P = 0.625
Collapsibility (choanal minus epiglottic pressure; cmH2O) 1.42 ± 0.37 1.78 ± 0.27 P = 0.438
Open in a new tab

Means ±s.e.m.; n = 5 subjects. Probability: chance that the null hypothesis is false from Wilcoxon matched-pairs signed-ranks test of Awake pre-sleep vs. Awake post-sleep.

Table 5.

Comparison of negative pressure stimuli, genioglossal reflex response to negative pressure stimuli and airway collapsibility during wakefulness recorded with eyes open and with eyes closed

Awake: eyes open Awake: eyes closed Probability
Maximal pressure stimulus in choanae (cmH2O) −8.12 ± 0.51 −7.57 ± 0.44 P = 0.102
Latency from stimulus onset to maximal stimulus (ms) 77 ± 6 84 ± 6 P = 0.154
Latency from stimulus onset to initial GG response (ms) 41 ± 4 44 ± 6 P = 0.959
Latency from stimulus onset to maximal GG response (ms) 163 ± 19 153 ± 18 P = 0.519
Baseline EMG at time ‘zero’ (% awake eyes open value) 100 ± 0 126 ± 14 P = 0.106
Maximal GG response (% change from baseline in each state) 159 ± 34 155 ± 24 P = 0.879
Collapsibility (choanal minus epiglottic pressure; cmH2O) 1.83 ± 0.45 1.68 ± 0.60 P = 0.499
Open in a new tab

Means ±s.e.m.; n = 10 subjects. Probability: chance that the null hypothesis is false from Wilcoxon matched-pairs signed-ranks test of Awake with eyes open vs. Awake with eyes closed.

Effect of order of presentation of pressure stimuli on the GG reflex

To test whether each GG response to a pressure stimulus is an independent effect, unaffected by the occurrence of prior stimuli, we determined whether there was a systematic change in the magnitude of the GG reflex to sequential pressure stimuli in 7 subjects during wakefulness. We found no significant trend in the data with sequential trials in 5 of 7 subjects. There was a significant increase over time in one subject, and a significant decrease over time in the other subject. The average Spearman rank correlation coefficient between trial number and magnitude of the GG reflex was −0.04 (range −0.72 to +0.47). Overall, in a paired analysis, there was no significant difference between the magnitude of the GG reflex from the first 5 trials to the last 5 trials during wakefulness (Wilcoxon matched-pairs signed-ranks test, P = 0.74). These analyses indicate that a stable measurement of the magnitude of the reflex could be obtained using our technique.

DISCUSSION

Our results demonstrate that in healthy humans, GG activation in response to upper airway negative pressure is quantitatively and qualitatively different during REM sleep when compared to wakefulness. During wakefulness there was a brisk GG activation in response to negative airway pressure, but the same pressure stimulus caused a reduction of GG EMG during both tonic and phasic REM sleep. There was also a greater degree of upper airway collapse (i.e. presumed narrowing of the airway) during application of the pressure stimulus in REM sleep than in wakefulness. These results were consistent among subjects. However, we were not able to demonstrate changes in the GG EMG reflex or airway collapsibility response to applied upper airway negative pressure between relaxed wakefulness and immediate sleep onset transitions.

Relevance to obstructive sleep apnoea syndrome

We believe that understanding the changes in upper airway reflexes during sleep is germane to understanding the mechanism that initiates obstructive apnoeas. Awake patients with OSA have relatively increased GG activity, possibly driven by the negative pressure reflex to counteract the effects of an innately smaller upper airway (Mezzanotte et al. 1992). In these susceptible patients, loss of this reflex during sleep would increase the likelihood of upper airway collapse. It has previously been shown that the reflex activation of the GG is significantly reduced during stable NREM sleep in healthy subjects (Wheatley et al. 1993a; Horner et al. 1994). In both of these studies, stage I sleep was specifically excluded from analyses because it is a transitory state. But stable NREM sleep and obstructive apnoeas are generally mutually exclusive events. Hence, we felt it pertinent to study the immediate sleep onset period, as well as REM sleep, when apnoeas most commonly occur (e.g. Krieger & Kurtz, 1978). Our finding of an apparent inhibition of the GG EMG in response to negative airway pressure during REM sleep perhaps explains why the upper airway is most vulnerable to collapse during this stage of sleep.

Lack of reduction in GG EMG reflex to negative airway pressure during sleep onset

Other workers have found evidence that an upper airway reflex to negative pressure persists to some extent during NREM sleep. For example, Horner et al. (1994) still found a detectable reflex response of the GG during stage II NREM sleep when a relatively large pressure of −25 cmH2O was applied to a face mask worn by the subjects (n = 4). The magnitudes of these responses during sleep were less than half of that observed when awake. Wheatley et al. (1993a) used −8 cmH2O (as used in the current study) and found that the magnitudes of the GG responses during stages II/III/IV NREM sleep were less than one-sixth of the waking response (n = 6). Similar findings were observed by the same authors for the tensor palatini muscle, a tonically active airway dilator (Wheatley et al. 1993b) (n = 6). In the current study, there was a trend for the GG EMG reflex amplitude to decrease with sleep onset but this was not statistically significant. The mean difference was −29 % and the standard deviation of this difference was 80.5 %. Given this variability among subjects, we have calculated that it would require studying 123 subjects to find the same observed mean difference to be significant at the 0.05 probability level (2-tailed test) using a statistical power of 80 %. Thus, there is clearly the possibility of a Type II statistical error. However, the results indicate that the actual group mean physiological effect is either small or non-existent. Taken together, our data suggest that there is not a systematic decrease in the magnitude of the GG EMG reflex to negative pressure at sleep onset. Instead, the amplitude of the reflex probably gradually declines as sleep progresses from light sleep to deeper stages of NREM sleep, although it can still be activated by sufficiently high pressures during deeper NREM sleep. Perhaps the most interesting finding is the variability among subjects in this response at sleep onset (e.g. Fig. 2) because it seems possible that individuals who preserve their reflex response during sleep onset may be less susceptible to the development of OSA. The reasons for the differences among subjects were not studied in the current experiment.

These data do not necessarily indicate that there is no reduction in the magnitude of this reflex at sleep onset in patients with OSA. Most patients with untreated OSA are chronically sleep deprived and sleep-deprived people ‘descend’ more quickly into sleep than non-sleep-deprived people (Borbely, 1994). Therefore, it remains possible that there occurs a reduction in the GG reflex at sleep onset in sleep-deprived OSA patients. Indeed, Mezzanotte et al. (1996) found that patients with untreated OSA had a much larger decrease in GG EMG and tensor palatini EMG during the first few breaths after an alpha-theta transition than occurred in non-sleep-deprived normal controls, which could be due to the loss of this reflex. On the other hand, it is noteworthy that Berry et al. (1997) found the phasic GG EMG to be highly correlated to negative oesophageal pressure (and therefore airway pressure) during obstructed apnoeas during NREM sleep in severe OSA patients, and that this correlation was reduced by upper airway anaesthesia. This latter study suggests that GG EMG reflex may be present to some extent in patients with severe OSA, although it was not discussed whether these patients were sleep deprived or receiving therapy prior to that study. Further studies that quantify the rate of change of EEG during sleep onset are needed to define more clearly the role of this reflex in apnoea generation.

Reduced GG EMG reflex to negative airway pressure during REM sleep

We found a consistent reduction of the GG EMG response to the negative airway pressure pulses during REM sleep. Similarly, decreases in GG EMG activation during REM sleep compared to NREM sleep have been observed during experimentally imposed upper airway occlusions in normal subjects (Kuna & Smickley, 1988) and during obstructive apnoeas in a patient population (Okabe et al. 1994). Likewise, a REM-related abolition of GG reflexes has been reported in dogs in which negative pressure stimulation of the airway caused augmented tonic GG activity during wakefulness but not during REM sleep (Harms et al. 1996). This decreased or absent GG reflex activation in response to negative upper airway pressure observed in the current study and in other studies could help explain why the upper airway is vulnerable to collapse during this state.

The different response during wakefulness and REM sleep could be attributable to differences at any site along the reflex pathway, i.e. the afferent information, the central processing, or the tonic level of facilitation versus inhibition of hypoglossal motoneurons. For example, superior laryngeal nerve (SLN) inputs to the hypoglossal motoneuron cause a complex excitatory-inhibitory hypoglossal response whose characteristics change as the intensity and frequency of the electrical stimulation of the SLN is altered, presumably reflecting a balance between inhibitory and excitatory afferent inputs from the same nerve (Mifflin, 1997). Even if the afferent information reaching the brainstem is identical between states, there are substantial differences in the neurophysiological state of the brainstem between wakefulness and REM sleep that could affect the GG response (e.g. Chase, 1983; Chase & Morales, 1994; Rampon et al. 1996; Horner, 1996; Kubin et al. 1998). It is unclear whether the reversal of the GG reflex during REM sleep that we observed in the current study represents an alteration in afferent neural traffic, inhibition of hypoglossal motoneurons or disfacilitation of hypoglossal motoneurons. Further studies will be needed to delineate the underlying mechanisms.

Critique of methods

The importance of our findings is dependent upon whether or not the results can be extrapolated to the normal physiological situation or, perhaps more importantly, whether the results can be extrapolated to the obstructive apnoeic patient. These concerns fall into four general areas. (i) Is the activity of the GG representative of the activity of the ensemble of all upper airway muscles? (ii) Is the applied stimulus representative of a normal physiological or even a pathophysiological stimulus? (iii) Are the reflex mechanisms activated by rapid upper airway pressure pulses representative of what occurs during normal sleep onset and during normal REM sleep? (iv) Is the protocol adequate to assess sleep-state effects on GG EMG reflexes?

First, these data must be viewed in the context that they assess the effect of sleep on the responsiveness of only one of the many upper airway muscles. The GG was selected because it is considered to be one of the principal upper airway dilators, is easily accessible to the experimenter, is large enough to enable reliable electrode placement and because we wished to compare these results with previous studies investigating GG activity during NREM sleep (Wheatley et al. 1993a; Horner et al. 1994). However, in OSA patients collapse of the airway occurs more commonly at the level of the soft palate than at the level of the oropharynx (Horner et al. 1989). Hence, state-related alterations in the responsiveness of palatal muscles, or of the whole ensemble of upper airway muscles, may prove to be more relevant to the pathogenesis of OSA than an assessment of GG activity alone. However, the net effect of the action of all of the upper airway muscles is reflected in our measurement of collapsibility (i.e. pressure drop between choanal and epiglottic regions during negative pressure stimulus application). Our data indicate that a reduction in GG reflex activation was associated with increased upper airway collapsibility, suggesting that the GG activity may reflect the net activity of all upper airway muscles quite well in these states.

A related concern is that the level of tonic GG EMG prior to assessing the magnitude of the reflex activation by negative pressure may differ between states and be affected by variables other than sleep state, including muscle length and position, level of respiratory drive (including arterial blood gas tensions) and compensatory changes related to alterations in upper airway resistance or compliance (ultimately a reflection of the activity of the ensemble of all upper airway muscles). Therefore, we assessed the magnitude of the reflex both as a peak GG level and as a percentage change from baseline tonic GG level. The results were similar with both approaches. We found no significant difference in tonic GG activity during sleep onset when compared to wakefulness. This was not surprising since it has previously been shown that there is an initial decrease followed by an increase in GG activity over the first five breaths of an EEG alpha-theta transition (Worsnop et al. 1998). It should also be noted that while skeletal muscle hypotonia is a characteristic of REM sleep, we found no difference in tonic GG activity during REM sleep when compared to wakefulness. This result is consistent with our previous observations, with intramuscular electrodes, that tonic GG activity is similar during tonic REM and non-REM sleep (Wiegland et al. 1991). Second, in our experiment it was necessary to impose a rapid and relatively large pressure stimulus (approximately −8 cmH2O over 77 ms) to distinguish the GG EMG reflex response from underlying changes in GG EMG that occur during a normal breath. The imposed pressure stimuli that we used were larger than would occur in a normal individual during tidal breathing. Nonetheless, the stimulus amplitude utilised is certainly within the range experienced by snorers and apnoeics, and in that context the applied pressure pulse appears to be a useful investigative probe of the pathological state.

A third concern is the possibility that these reflex data are affected by arousal from sleep, such that the results do not reflect reflex mechanisms during either sleep onset or REM sleep. To address this, we excluded any negative pressure pulses associated with a noticeable EEG arousal. Although it was relatively easy to detect an EEG arousal during sleep onset, it was more difficult during REM sleep as the EEG is characteristically desynchronised. We attempted to be conservative by deleting any trial where a noticeable quickening of the EEG occurred within 1 s of the pressure stimulus. Furthermore, we assessed the reflex only within 200 ms of the initiation of the stimulus, thereby reducing the chance that data are contaminated by arousal artifacts not seen on the EEG channel. Finally, even if an occasional trial was associated with undetected EEG microarousal, we do not believe that this would have affected our results as there was no qualitative difference detectable when data with and without visible arousal were compared (current study; data not presented).

A related concern is the possibility that the upper airway pressure stimulus activated a startle response in our subjects that affected GG activity, and that this startle mechanism is not representative of a normal sleep onset or REM effect. Certainly, it is known that auditory tones can activate a startle reflex during wakefulness, NREM and REM sleep that results in inhibition of the diaphragm (Kline et al. 1990). However, in contrast to the qualitatively different GG reflex between wakefulness and REM sleep that we observed in the current study, the overall response to auditory tones in the study of Kline et al. (1990) was qualitatively similar during wakefulness and REM sleep. Hence, we do not believe that our results are representative of a startle mechanism.

A fourth concern is whether or not the protocol itself is adequate to assess sleep-state effects on GG EMG reflexes. One important factor is the definition of relaxed wakefulness used for comparisons with the sleep stages, as well as the definition and method of collection of sleep onset data. We did not compare the GG EMG reflex just before sleep onset with just after sleep onset. We assumed that application of the negative pressure stimuli during relaxed wakefulness would arouse the subject thereby preventing sleep. Therefore, we chose to assess the reflex during stable relaxed wakefulness. However, these subjects were probably less drowsy when wakefulness data were collected than they would have been immediately before the sleep onset. If drowsiness and sleep onset are both associated with a reduction in the GG reflex, separating the collection of awake and sleep onset data could bias the data towards finding a significant difference between the two states. However, as we did not detect a significant difference between wakefulness and sleep onset, we have more confidence that a real physiological effect of sleep onset was not missed.

Another possible concern about the protocol is that repeated pressure pulse stimuli could feasibly result in different GG effects than single pressure pulse stimuli. For instance, it has been shown that a brief pulse of negative pressure has a prolonged effect on expiratory duration in dogs (Harms et al. 1996), and a maintained upper airway negative pressure stimulus results in the adaptation of the activity of some upper airway muscles in rabbits (Mathew, 1984). The fact that we found no systematic trend in the magnitude of the GG reflex data with sequential trials during wakefulness partially addresses this concern. Our stimuli were separated by 2–7 breaths but we did not investigate whether there were differences in the GG reflex when the duration between stimuli was varied. It seems likely that investigation of repeated pressure stimuli is more germane to the normal physiological condition of breath-by-breath fluctuations in airway pressure than investigation of isolated stimuli. Finally, over the longer time scale, the similarity in GG activation and airway collapsibility observed during relaxed wakefulness at the beginning and end of the night suggest that the changes in these variables between wakefulness and REM sleep are genuine state effects, little influenced by the time of the recording.

Summary

We believe that the pressure stimulus employed in this protocol is a useful approach that has potential relevance to studying the pathogenesis of OSA. Our results demonstrate that GG activation in response to upper airway negative pressure is not systematically altered in healthy subjects at sleep onset. Instead, there was marked variability in responses among subjects. Thus, loss of this reflex may not be implicated in upper airway obstruction at sleep onset. However, the situation may be different in sleep-deprived apnoeic patients. Our results also show that the GG reflex to negative pressure is quantitatively and qualitatively different during wakefulness and REM sleep. During wakefulness negative airway pressure activated the GG, whereas the same pressure stimulus during REM sleep caused an apparent inhibition of muscle activity. The upper airway was also more collapsible during REM sleep. Although there are many neurophysiological mechanisms that could explain this reversal of the GG response during REM sleep, we can only speculate on these mechanisms in humans. Regardless of the underlying mechanism, we believe that these results are important because a reversal of this reflex response to negative pressure during REM sleep – whereby respiratory efforts generating negative upper airway pressure would inhibit airway dilators – may explain why the upper airway is most vulnerable to collapse during this stage of sleep.

Acknowledgments

This research was supported by US PHS grants NIH GCRC RR 02635, NIH HL48531 and NIH HL60292.

References

  1. Berry RB, McNellis MI, Kouchi K, Light RW. Upper airway anesthesia reduces phasic genioglossus activity during sleep apnea. American Journal of Respiratory and Critical Care Medicine. 1997;156:127–132. doi: 10.1164/ajrccm.156.1.9608037. [DOI] [PubMed] [Google Scholar]
  2. Borbely AA. Sleep homeostasis and models of sleepregulation. In: Kryger MH, Roth T, Dement WC, editors. Principles and Practice of Sleep Medicine. 2. Philadelphia: W. B. Saunders Company; 1994. pp. 309–320. [Google Scholar]
  3. Chase MH. Synaptic mechanisms and circuitry involved in motoneuron control during sleep. International Review of Neurobiology. 1983;24:213–258. doi: 10.1016/s0074-7742(08)60223-8. [DOI] [PubMed] [Google Scholar]
  4. Chase MH, Morales FR. The control of motoneurons during sleep. In: Kryger MH, Roth T, Dement WC, editors. Principles and Practice of Sleep Medicine. 2. Philadelphia: W. B. Saunders Company; 1994. pp. 163–175. [Google Scholar]
  5. Harms CA, Zeng YJ, Smith CA, Vidruk EH, Dempsey JA. Negative-pressure induced deformation of the upper airway causes central apnea in awake and sleeping dogs. Journal of Applied Physiology. 1996;80:1528–1539. doi: 10.1152/jappl.1996.80.5.1528. [DOI] [PubMed] [Google Scholar]
  6. Horner RL. Motor control of the pharyngeal musculature and implications for the pathogenesis of obstructive sleep apnea. Sleep. 1996;10:827–853. doi: 10.1093/sleep/19.10.827. [DOI] [PubMed] [Google Scholar]
  7. Horner RL, Innes JA, Morrell MJ, Shea SA, Guz A. The effect of sleep on reflex genioglossus muscle activation by stimuli of negative airway pressure in humans. The Journal of Physiology. 1994;476:141–151. [PMC free article] [PubMed] [Google Scholar]
  8. Horner RL, Innes JA, Murphy K, Guz A. Evidence for reflex upper airway dilator muscle activation by sudden negative airway pressure in man. The Journal of Physiology. 1991;436:31–44. doi: 10.1113/jphysiol.1991.sp018536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Horner RL, Shea SA, McIivor J, Guz A. Pharyngeal size and shape during wakefulness and sleep in patients with obstructive sleep apnoea. Quarterly Journal of Medicine. 1989;72:719–735. [PubMed] [Google Scholar]
  10. Kline LR, Hendricks JC, Silage DA, Morrison AR, Davies RO, Pack AI. Startle-evoked changes in diaphragmatic activity during wakefulness and sleep. Journal of Applied Physiology. 1990;68:166–173. doi: 10.1152/jappl.1990.68.1.166. [DOI] [PubMed] [Google Scholar]
  11. Krieger J, Kurtz D. EEG changes before and after apnea. In: Guilleminault C, Dement WC, editors. Sleep Apnea Syndromes. New York, USA: Alan R. Liss, Inc; 1978. pp. 161–176. [Google Scholar]
  12. Kubin L, Davies RO, Pack AI. Control of upper airway motoneurons during REM sleep. News in Physiological Sciences. 1998;13:91–97. doi: 10.1152/physiologyonline.1998.13.2.91. [DOI] [PubMed] [Google Scholar]
  13. Kuna ST, Smickley J. Response of genioglossus muscle activity to nasal airway occlusion in normal sleeping adults. Journal of Applied Physiology. 1988;64:347–353. doi: 10.1152/jappl.1988.64.1.347. [DOI] [PubMed] [Google Scholar]
  14. Mathew OP. Upper airway negative-pressure effects on respiratory activity of upper airway muscles. Journal of Applied Physiology. 1984;56:500–505. doi: 10.1152/jappl.1984.56.2.500. [DOI] [PubMed] [Google Scholar]
  15. Mezzanotte WS, Tangel DJ, White DP. Waking genioglossal EMG in sleep apnea patients versus normal controls (a neuromuscular compensatory mechanism) Journal of Clinical Investigation. 1992;89:1571–1579. doi: 10.1172/JCI115751. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Mezzanotte WS, Tangel DJ, White DP. Influence of sleep onset on upper airway muscle activity in apnea patients versus normal controls. American Journal of Respiratory and Critical Care Medicine. 1996;153:1880–1887. doi: 10.1164/ajrccm.153.6.8665050. [DOI] [PubMed] [Google Scholar]
  17. Mifflin SW. Intensity and frequency dependence of laryngeal afferent inputs to respiratory hypoglossal motoneurons. Journal of Applied Physiology. 1997;83:1890–1899. doi: 10.1152/jappl.1997.83.6.1890. [DOI] [PubMed] [Google Scholar]
  18. Okabe S, Hida W, Kikuchi Y, Taguchi O, Takashima T, Shirato K. Upper airway muscle activity during REM and non-REM sleep of patients with obstructive apnea. Chest. 1994;106:767–773. doi: 10.1378/chest.106.3.767. [DOI] [PubMed] [Google Scholar]
  19. Rampon C, Peyron C, Petit J, Fort P, Gervasoni D, Luppi P. Origin of the glycinergic innervation of the rat trigeminal motor nucleus. NeuroReport. 1996;7:3081–3085. doi: 10.1097/00001756-199611250-00058. [DOI] [PubMed] [Google Scholar]
  20. Rechtshaffen A, Kales A. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep States of Human Subjects. Washington, DC: Public Health Service, US Government Printing Office; 1968. [Google Scholar]
  21. Shea SA, Walter J, Pelley C, Murphy K, Guz A. The effects of auditory and visual stimulation upon the breathing pattern of resting healthy subjects. Respiration Physiology. 1987;68:345–357. doi: 10.1016/s0034-5687(87)80019-1. [DOI] [PubMed] [Google Scholar]
  22. Wheatley JR, Mezzanotte WS, Tangel DJ, White DP. Influence of sleep on genioglossus muscle activation by negative pressure in normal men. American Review of Respiratory Disease. 1993a;148:597–605. doi: 10.1164/ajrccm/148.3.597. [DOI] [PubMed] [Google Scholar]
  23. Wheatley JR, Tangel DJ, Mezzanotte WS, White DP. Influence of sleep on response to negative airway pressure of tensor palatini muscle and retropalatal airway. Journal of Applied Physiology. 1993b;75:2117–2124. doi: 10.1152/jappl.1993.75.5.2117. [DOI] [PubMed] [Google Scholar]
  24. White DP. Pathophysiology of obstructive sleep apnea. Thorax. 1995;50:797–804. doi: 10.1136/thx.50.7.797. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. White DP, Edwards JK, Shea SA. Local reflex mechanisms: influence on basal genioglossal muscle activation in normal subjects. Sleep. 1998;21:719–728. doi: 10.1093/sleep/21.7.719. [DOI] [PubMed] [Google Scholar]
  26. Wiegland L, Zwillich CW, Wiegland D, White DP. Changes in upper airway muscle activation and ventilation during phasic REM sleep in normal men. Journal of Applied Physiology. 1991;71:488–497. doi: 10.1152/jappl.1991.71.2.488. [DOI] [PubMed] [Google Scholar]
  27. Worsnop C, Kay A, Pierce R, Kim Y, Trinder J. Activity of respiratory pump and upper airway muscles during sleep onset. Journal of Applied Physiology. 1998;85:908–920. doi: 10.1152/jappl.1998.85.3.908. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES