Journal of Applied Physiology


Previous studies in humans showed that genioglossal muscle activity is higher when individuals are supine than when they are upright, and prior experiments in anesthetized or decerebrate animals suggested that vestibular inputs might participate in triggering these alterations in muscle firing. The present study determined the effects of whole body tilts in the pitch (nose-up) plane on genioglossal activity in a conscious feline model and compared these responses with those generated by roll (ear-down) tilts. We also ascertained the effects of a bilateral vestibular neurectomy on the alterations in genioglossal activity elicited by changes in body position. Both pitch and roll body tilts produced modifications in muscle firing that were dependent on the amplitude of the rotation; however, the relative effects of ear-down and nose-up tilts on genioglossal activity were variable from animal to animal. The response variability observed might reflect the fact that genioglossus has a complex organization and participates in a variety of tongue movements; in each animal, electromyographic recordings presumably sampled the firing of different proportions of fibers in the various compartments and subcompartments of the muscle. Furthermore, removal of labyrinthine inputs resulted in alterations in genioglossal responses to postural changes that persisted until recordings were discontinued ∼1 mo later, demonstrating that the vestibular system participates in regulating the muscle's activity. Peripheral vestibular lesions were subsequently demonstrated to be complete through the postmortem inspection of temporal bone sections or by observing that vestibular nucleus neurons did not respond to rotations in vertical planes.

  • tongue musculature
  • vestibular system
  • obstructive sleep apnea
  • upper airway

in addition to the pump muscles that move air into and out of the lungs, muscles in the upper airway also contract in a coordinated fashion during respiration. Some of these muscles are active during inspiration and serve to maintain airway patency. The genioglossal muscle, which serves, in part, to protrude the tongue, is typically regarded as an inspiratory upper airway muscle (23). Genioglossal muscle activity has also been reported to increase during certain postural alterations, presumably to reduce pharyngeal resistance. In particular, studies in humans have revealed that genioglossal activity is higher when subjects are supine than in the upright position (3, 6, 15, 19, 22, 25, 29, 30). Dorsal flexion of the head in humans was also reported to elicit an increase in genioglossal firing (3), although in conscious cats the same movements were described as producing a decrease in muscle activity (2). Limited data are available concerning the effects of other postural changes on genioglossal activity, whereas it has been shown that muscle firing is lower when human subjects are prone than when they are supine (22).

Data from experimental animals have indicated that the vestibular system participates in eliciting alterations in genioglossal activity during changes in posture. Electrical stimulation of the vestibular nerve (7, 11, 14, 28) or caloric stimulation of the ear (11, 13) produced responses in the hypoglossal nerve or hypoglossal motoneurons, which innervate genioglossus and other muscles that move the tongue. Electrical stimulation of the labyrinth (vestibular portion of the inner ear) has also been directly demonstrated to produce contractions of the genioglossal muscle (1). Furthermore, nose-up rotations of the head in cats with extensive denervations to eliminate nonlabyrinthine inputs produced by the movement elicited increases in the firing of the hypoglossal nerve (24). Ear-down tilts also produced increases in hypoglossal nerve activity in one-fourth of the animals (24), and static ear-down rotations in rabbits have been shown to elicit increased discharges of some hypoglossal motoneurons (12). Changes in hypoglossal nerve activity produced by electrical stimulation of the vestibular nerve (7) or nose-up tilts of the head (24) were abolished by transection of the eighth cranial nerve, demonstrating that these responses were produced by activation of the vestibular system.

It is not surprising that postural changes elicited variable effects on genioglossal activity, as this small muscle comprises fibers with a complex geometry (5, 16, 17, 20, 21). At least two compartments of genioglossus have been distinguished in all mammals that have been considered: one with fibers running horizontally that play a primary role in tongue protrusion, and another with fibers coursing obliquely that depress the tongue (5, 21). In the dog, three functional subdivisions of the oblique compartment have been discovered that presumably are active during different tasks (21). Because conventional electromyographic (EMG) techniques are likely to sample activity in multiple compartments and subcompartments of genioglossus, responses that are recorded from the muscle can presumably vary, depending on the specific fibers whose activity is monitored at a particular time.

The studies in humans that considered the effects of postural changes on genioglossal muscle activity have a number of limitations. In such experiments, it is only possible to monitor genioglossal discharges for a short period of time, which immediately follows the insertion of the recording electrode into the muscle. It is possible that extraneous factors, including discomfort associated with the electrode placement or laboratory environment, could alter the responses. Furthermore, it is difficult to ensure that electrode position does not shift during postural alterations, as the electrodes cannot be firmly secured to the tissues. In addition, determining the role of the vestibular system in producing the postural-related changes in genioglossal discharges is difficult in human subjects, as invasive procedures are required for selective stimulation or ablation of labyrinthine afferents. The goal of the present study was to chronically monitor genioglossal muscle activity in a conscious feline model and to ascertain the effects of static nose-up and ear-down tilts of the whole body on the muscle's activity. The feline was used as an animal model because of the wealth of information regarding both neural regulation of respiration and posture in this species; the only major disadvantage of employing the cat (or another quadrupedal animal) relates to the fact that its airway geometry differs from that in bipeds, such that the effects of postural alterations on genioglossal activity observed in the present experiments may not be equivalent in humans. Whole body tilts were employed to alter head position in space, as opposed to head rotations on a stable body, so that neck afferents were not stimulated and the possibility of activating airway receptors by stretching of the trachea and other structures was diminished. Furthermore, we determined the effects of bilateral removal of vestibular inputs on the spontaneous and tilt-related firing of the muscle. We tested the hypothesis that ear-down and nose-up tilts have variable effects on genioglossal discharges, depending on the particular compartments of the muscle whose activity is sampled in a particular animal, but that, in all cases, these responses are altered by removal of labyrinthine inputs.


All experimental procedures conformed to the National Institutes of Health Guide for the Care and Use of Laboratory Animals and were approved by the University of Pittsburgh's Institutional Animal Care and Use Committee. Data were collected from nine adult female cats obtained from Liberty Research (North Rose, NY). Animals were spayed before being included in this study to eliminate cyclic changes in hormonal levels.

Overview of experimental procedures. Changes in genioglossal muscle activity were recorded during nose-up (9 animals) and ear-down (6 animals) static tilts of the whole body at 20, 40, and 60° amplitude. The tilt table was rotated manually and secured in the tilted position by using a spring device that permitted rotation to one of the three predetermined amplitudes. Rotations from the Earth-horizontal to the nose-up or ear-down position were performed rapidly, at a velocity of ∼60°/s. In six of the animals, labyrinthine inputs were also surgically eliminated, and the effects of these lesions on both the muscle's baseline activity and changes in activity during tilts were ascertained. The design of these experiments was similar to that in a previous study in which the effects of body tilts and vestibular lesions on the activity of the diaphragm and abdominal musculature were considered (4). For the ∼2-mo period preceding data collection, animals were acclimated to being restrained in the prone position on the tilt table. For this purpose, a vinyl bag with attached straps was placed around the animal's body; the straps were secured to the tilt table to prevent the animal's position from shifting during testing. Furthermore, the animal's head was immobilized by inserting a screw into a bolt mounted on the skull. Food was provided at the end of the testing period as a reward, and the experimental session was terminated promptly if the animal attempted to remove itself from restraint or showed any signs of distress (e.g., vocalization that persisted for longer than a few seconds). Genioglossal EMG activity was recorded by using a pair of Teflon-insulated stainless steel wires (Cooner Wire, Chatsworth, CA) that were stripped of insulation for ∼1 mm and inserted directly into the muscle and fixed in place using sutures. The insulated portions of the wires were led subcutaneously and soldered to a connector mounted on the animal's head. Data were collected during recording sessions with a duration of ∼30 min; one to three recording sessions were conducted per day. During each recording session, only one direction of tilt (either nose up or left or right ear down) was typically performed, although table rotations of 20, 40, and 60° amplitudes were randomly dispersed so that the animals could not predict the amplitude of tilt at the onset of each rotation. Tilts persisted for 40–60 s and were separated by at least 1 min. Data were not analyzed from trials in which abrupt transients in genioglossal muscle activity occurred as a result of voluntary tongue movement related to vocalization, swallowing, or other activities.

Genioglossal muscle responses to tilts were recorded in animals with an intact labyrinth over a period of 35–139 days (median of 79 days). For each trial, average genioglossal muscle activity was ascertained for the period during which the animal was tilted maximally and compared with activity averaged over an equivalent time period immediately before the tilt to determine the percent change in activity produced by the tilt. Subsequently, vestibular inputs were removed bilaterally in six animals, and data collection continued in the same manner as before the lesions. The postlesion recordings began the day after the surgery and continued for 4 wk.

Surgical procedures. A recovery surgery was required for each animal to secure a bolt to the skull to permit head fixation and to implant EMG recording electrodes. This surgery was performed using sterile procedures in a dedicated operating suite. Animals were initially anesthetized by an intramuscular injection of ketamine (15 mg/kg) and acepromazine (0.2 mg/kg). Subsequently, an endotracheal tube and intravenous (IV) catheter were inserted. Anesthesia was then maintained using 1–2% isoflurane vaporized in O2 such that areflexia was present and heart rate was stable. Ringer lactate solution was infused IV to replace fluid loss during surgery, and a heating pad was used to maintain rectal temperature near 38°C. To place EMG electrodes for recording genioglossus activity, a small incision was placed on the underside of the jaw, and superficial musculature was retracted. After implantation of electrodes in the genioglossal muscle and closure of the overlying muscles and skin using sutures, the animal's head was secured in a stereotaxic frame, and a head fixation bolt was implanted, as described in a previous publication (9). Wires from the EMG recording electrodes were attached to a connector mounted on the skull behind the fixation plate.

A second recovery surgery was performed in six of the animals following initial data recording to eliminate vestibular inputs. Anesthesia was produced by using the same procedures as employed in the initial surgery. The tympanic bulla on each side was exposed by using a ventrolateral approach and was opened to expose the cochlea. A drill was used to remove temporal bone near the base of the cochlea, thereby producing a labyrinthectomy that rendered the vestibular apparatus dysfunctional. This procedure also provided access to the internal auditory canal. The eighth cranial nerve was then transected under microscopic observation within the internal auditory canal. Thus two independent lesions affecting the vestibular system were made on both sides to assure that vestibular inputs were removed. To ensure that animals received adequate hydration and nutrition during the postsurgical period, an IV injection port remained in place for 2 days after surgery, and 50 ml of 5% dextrose solution were administered IV each day. In addition, feeding was done by hand until the animal's spontaneous consumption of food and water returned to prelesion levels.

Procedures for recording and analyzing EMG activity. During recording sessions, a cable was attached to the head-mounted connector to allow EMG signals to be fed to an amplifier (model 1700, A-M Systems, Carlsberg, WA); activity was amplified by a factor of 104, filtered with a band pass of 10–10,000 Hz, and full-wave rectified with a time constant of 1 ms. Subsequently, signals were recorded digitally using a 1401-plus data collection system (Cambridge Electronic Design, Cambridge, UK) interfaced with Macintosh (Apple Computer, Cupertino, CA) G4 computer; the sampling rate was 1,000 Hz. A potentiometer mounted on the tilt table provided a recording of table position; the voltage from this potentiometer was digitized and sampled at 100 Hz. The Spike-2 software package (Cambridge Electronic Design) was used for data analysis.

Statistical analyses of data were performed with the use of the Prism 3 software package (GraphPad Software, San Diego, CA) running on a Macintosh G4 computer. A nonparametric ANOVA procedure (Friedman test), in combination with a post hoc test (Dunn's multiple-comparison test), was used to compare the changes in genioglossal muscle activity that were elicited by different amplitudes of tilt in a particular direction or different directions of tilt at a fixed amplitude. The same statistical test was employed to compare spontaneous and tilt-elicited changes in muscle activity before and at various periods subsequent to removal of vestibular inputs. A Mann-Whitney test was used to compare overall changes in spontaneous and tilt-related muscle activity before and after vestibular lesions. Statistical significance was set at P < 0.05. Pooled data are presented as means ± SE.

Verification of vestibular lesions. Nystagmus or changes in eye position never occurred following removal of vestibular inputs, suggesting that the lesions were complete bilaterally. Furthermore, we conducted an additional analysis to verify that vestibular inputs were eliminated following the combined labyrinthectomy and vestibular neurectomy. In three animals, anatomic methods were employed for this purpose. At the conclusion of data recording in these cases, the animals were anesthetized with an intraperitoneal injection of pentobarbital sodium (40 mg/kg) and perfused transcardially with phosphate-buffered saline followed by paraformaldehyde-lysine-sodium metaperiodate fixative (18). The head was removed and decalcified by using a solution of EDTA and 5% formic acid. The temporal bone on each side was subsequently removed, embedded in 15% celloidin, cut in the coronal plane (30-μm thickness), and stained with hematoxylin. The temporal bone sections were then inspected histologically to determine the extent of damage to the eighth cranial nerves and vestibular labyrinth.

A physiological analysis was employed in the other three animals that experienced peripheral vestibular lesions to ensure that labyrinthine inputs were eliminated. These animals were rendered decerebrate and prepared for acute recordings from the vestibular nuclei using procedures discussed in previous studies (10, 31). In each case, the spinal cord was transected at C2, as it was previously demonstrated that nonlabyrinthine inputs (presumably including spinal inputs) elicited during whole body tilts could provide for modulation of the activity of vestibular nucleus neurons (31). After this procedure, an IV infusion of metaraminol bitartrate (80 μg/ml; Aramine; Merck) in lactated Ringer solution was used to keep arterial blood pressure >100 mmHg. A bilateral vagotomy was also performed in one animal to prevent visceral inputs from reaching the brain stem. Effects of sinusoidal “wobble” stimuli (26), which incorporate both roll and pitch tilts, on the firing of 95 vestibular nucleus neurons were recorded; 5° stimuli were delivered at 0.5 Hz, and 5–15° stimuli were delivered at 0.05–0.1 Hz. These stimuli would be expected to robustly modulate vestibular nucleus neuronal activity in labyrinth-intact animals (26). After the recording session, animals were anesthetized using pentobarbital sodium and transcardially perfused using the same procedures employed for the anatomic control experiments discussed in the previous paragraph. The brain stem was cut into 100-μm sections, and recording sites were reconstructed and verified to be located in the vestibular nuclei using procedures employed in previous studies (10, 31).


Effects of whole body tilts on genioglossal muscle activity. Examples of genioglossal EMG activity monitored before and during 60° nose-up tilts of two animals are illustrated in Fig. 1. Figure 1 includes responses recorded before and subsequent to the bilateral removal of vestibular inputs for one of the cases. For the animal whose data are illustrated in Fig. 1A, nose-up tilt elicited an increase in muscle activity before removal of vestibular inputs. In contrast, for some other animals, including the case shown in Fig. 1B, genioglossal muscle activity was depressed during nose-up tilts preceding the peripheral vestibular lesions. The variability in responses between animals is further indicated in Fig. 2, which shows the mean percent change in genioglossal firing produced by 60° nose-up, ipsilateral ear-down, and contralateral ear-down tilts for each cat employed in this study. All data included in Fig. 2 were collected when vestibular inputs were intact. In six cases (animals 1, 2, 3, 4, 6, and 7), nose-up tilt elicited an increase in genioglossal discharges; in the other three cases (animals 5, 8, and 9), a decrease in muscle firing was produced. Ipsilateral ear-down tilts produced more consistent effects and elicited an increase in genioglossal activity in every cat tested; contralateral ear-down tilts evoked an increase in muscle firing in five of the six cats that experienced this stimulus. Furthermore, the relative effects of nose-up and ear-down stimuli on genioglossal discharges were variable from animal to animal. For example, nose-up tilt produced a larger (Friedman test) increase in muscle activity than either ipsilateral or contralateral ear-down tilt for animal 4, whereas, in animal 3, tilts in the ear-down direction were more effective.

Fig. 1.

Examples of genioglossal muscle activity recorded immediately before and during 60° nose-up (NU) tilts, both preceding (Pre-Lesion) surgical removal of vestibular inputs in 2 animals, and subsequent to (Post-Lesion) the lesions in 1 case. A: responses obtained from the cat identified as animal 1 in subsequent figures. B: responses from animal 5.

Fig. 2.

Mean percent changes in genioglossal muscle activity produced by 60° tilts in the NU, ipsilateral ear-down (IED), and contralateral ear-down (CED) directions. Data from each animal included in this study are shown in a different panel. In 3 cases (animals 7, 8, and 9), ear-down tilts were not conducted; thus only responses to NU tilts are indicated. Values are means ± SE. Significant difference from the change in activity produced by NU tilt, P < 0.05.

For each animal, some variability in the magnitudes of genioglossal muscle responses to each direction and amplitude of tilt was also noted. For example, Fig. 3 shows the changes in activity elicited by every 60° trial performed in animal 1. Although individual responses differed in amplitude, it was apparent that the effects of nose-up and contralateral ear-down rotations tended to differ (the large majority of nose-up tilts elicited an increase in activity, whereas virtually all contralateral ear-down trials produced a decrease in firing). Furthermore, it did not appear that response magnitudes changed systematically over time, but varied above and below the mean values from trial to trial.

Fig. 3.

Changes in genioglossal muscle activity in 1 animal (animal 1) that were elicited during individual trials that employed 60° stimuli; each symbol designates the effects elicited by a particular tilt. Open symbols, responses to NU, IED, and CED rotations recorded before removal of labyrinthine inputs; solid symbols, responses to NU tilts following peripheral vestibular lesions; shaded bar, mean of responses to a particular stimulus condition ± SE.

To quantitatively compare the mean magnitudes of genioglossal muscle responses to different amplitudes and directions of tilt, we considered the absolute value of the alteration in muscle firing produced by these stimuli. Figure 4 illustrates the mean absolute change in muscle activity elicited by ear-down and nose-up tilts of 20, 40, and 60° amplitudes; data from all animals were pooled for this analysis. For each direction of tilt, the magnitude of the effects was significantly dependent on tilt amplitude (Friedman test), such that 60° stimuli produced larger responses than 20° stimuli. However, we did not uncover any significant differences between the absolute magnitude of responses to nose-up, ipsilateral ear-down, and contralateral ear-down tilt at a particular stimulus amplitude, presumably because of the disparity of the effects between animals.

Fig. 4.

Mean absolute change in muscle activity elicited by ear-down and NU tilts of 20, 40, and 60° amplitudes. Data from all cases were pooled for this analysis, but were weighted such that each animal contributed equally to the averages. Values are means ± SE.

Consequences of removal of vestibular inputs on genioglossal activity and responses to tilts. The effects of peripheral vestibular lesions on the percent change in genioglossal activity produced by 60° nose-up tilts are shown in Fig. 5; data for each of the six cats that experienced the lesions are provided in a separate panel. Figure 3 illustrates the effects of each 60° nose-up tilt performed before and after lesions in one of the animals (animal 1), whereas Fig. 1A shows examples of responses to 60° nose-up tilts collected before and after elimination of labyrinthine signals. In five of the cases, the changes in muscle activity elicited by 60° nose-up tilts were significantly (Mann-Whitney test) altered by removal of vestibular inputs, and in the sixth case (animal 7) the effects approached significance (P = 0.086). The lesions produced a reduction in the change in muscle activity elicited by tilts in four of the cats (animals 1, 4, 5, and 7); whereas, in another case (animal 8), the alteration in muscle activity during tilts became larger following loss of vestibular signals. In the remaining cat (animal 9), genioglossal activity decreased during nose-up tilts before the lesions, but increased by a similar magnitude during nose-up rotations after elimination of labyrinthine signals.

Fig. 5.

Effect of removal of vestibular inputs on the percent change in genioglossal activity that occurred during 60° NU tilts. Data from each animal that experienced vestibular lesions are shown in a different panel. Values are means ± SE. Significantly different from the response elicited before vestibular lesions, P < 0.05.

Prior studies have revealed that, although removal of vestibular inputs compromises the ability of conscious animals to maintain stable blood pressure during nose-up tilts, recovery occurs after a few days (8, 9). To explore whether the effects of vestibular lesions on genioglossal responses persisted for longer than 1 wk, we compared the percent change in genioglossal activity elicited by 60° nose-up tilts before, for the first week after, and for the second to fourth weeks after elimination of labyrinthine signals. This analysis is shown in Fig. 6. There was no indication that the effects of lesions on the responses dissipated over time, and, in some cases (e.g., animal 4), the responses appeared to deviate more from those monitored before the lesions subsequent to 1 wk after the labyrinthectomy. However, for most animals, an ANOVA analysis failed to show a significant difference in the percent change in genioglossal activity during tilt at the different time points relative to the vestibular lesions. This lack of statistical significance might be due to the loss of statistical power as postlesion responses were separated into two different groups for comparisons.

Fig. 6.

Percent change in genioglossal activity elicited by 60° NU tilts preceding, during the 1st week subsequent to (W1), and during the 2nd to 4th weeks subsequent to (W2–4) removal of labyrinthine inputs. Data from each animal that experienced vestibular lesions are shown separately. Values are means ± SE. Significantly different from the response elicited before vestibular lesions, P < 0.05.

The effects of removal of vestibular inputs on the percent change in genioglossal muscle activity elicited by 60° ipsilateral and contralateral ear-down tilt were ascertained for three cats (animals 1, 4, and 5). For two of these cases (animals 1 and 5), the peripheral vestibular lesions produced a significant (Mann-Whitney test) attenuation of the responses to 60° ipsilateral ear-down tilt; for the third cat (animal 4), the changes in muscle firing during these tilts were not significantly different before and after lesions (P = 0.29). However, elimination of vestibular signals attenuated the change in genioglossal discharges that occurred during 60° contralateral ear-down tilts in only one cat (animal 5); the lesions did not significantly alter the magnitude of contralateral tilt-elicited responses in the other two animals (P = 0.27 for animal 1 and P = 0.84 for animal 4).

We also considered the consequences of elimination of labyrinthine inputs on spontaneous genioglossal muscle activity (monitored before each tilt); this analysis is presented in Fig. 7. Muscle firing was reduced in three cats (animals 4, 5, and 7), but increased in another case (animal 8), subsequent to removal of vestibular signals. In the other two animals, no significant (Mann-Whitney test, P = 0.15 and 0.71) alterations in baseline activity occurred following the lesions.

Fig. 7.

Changes in baseline genioglossal muscle activity resulting from the elimination of labyrinthine inputs. Muscle firing is shown as a percentage of the average activity before lesions. Values are means ± SE. Significantly different from the background firing recorded before vestibular lesions, P < 0.05.

Verification of vestibular lesions. Temporal bone sections from three cats (animals 1, 7, and 8) were inspected to verify that the labyrinthectomy was complete, and the VIIIth nerve was transected on both sides. It was evident from this analysis that vestibular inputs had been completely eliminated. In the other three cats that experienced peripheral vestibular lesions (animals 4, 5, and 9), a physiological analysis was employed to demonstrate that labyrinthine signals did not reach the central nervous system. In total, recordings were made from 95 neurons in these animals that were histologically verified to be located in the vestibular nuclei. Effects of “wobble” stimuli (26), whole body sinusoidal tilts that included both pitch and roll rotations, on the firing of these neurons were ascertained. Maximal tilt amplitudes of 5° (32 cells), 10° (41 cells), or 15° (22 cells) were employed for wobble stimuli delivered at 0.05–0.1 Hz, and 0.5-Hz rotations at 5° amplitudes were also typically utilized. The firing of only one of the neurons was modulated by these stimuli, as indicated by responses with a signal-to-noise ratio ≥ 0.5, for which only the first harmonic was prominent (10, 26, 31). For the 63 units that were tested using 10° stimuli, the mean signal-to-noise ratio for neuronal activity recorded during clockwise wobble rotations performed at the highest frequency at which 10° rotations were delivered was 0.22 ± 0.04; this value differed significantly (P < 0.0001, Wilcoxon signed-rank test) from the minimum signal-to-noise ratio (0.5) that indicated the presence of a response to the stimuli. It is possible that the responses recorded from the single neuron that did respond to tilt were elicited by movement of the viscera during rotations of the tilt table, as visceral afferent inputs have been demonstrated to affect the firing of vestibular nucleus neurons (10). Because the stimuli employed during these recordings should have robustly modulated the firing of a large fraction of vestibular nucleus neurons in labyrinth-intact animals (10, 26), the paucity of responses observed suggests that vestibular inputs were at least greatly attenuated by the lesions previously performed on the animals.


The present results showed that genioglossal muscle activity is altered by changes in body position in either the pitch or roll planes; these alterations in activity were dependent on the amplitude of the body tilt. However, the relative effects of ear-down and nose-up tilts on genioglossal firing were variable from animal to animal. Thus this study revealed that changes in posture produce more complex alterations in genioglossus discharges than have been described in previous studies, which typically only considered the effects of one direction of movement (2, 3, 6, 15, 19, 22, 25, 29, 30). Furthermore, the changes in genioglossal activity produced by nose-up body tilts were altered by removal of vestibular inputs; in most cases, the responses were attenuated in magnitude, but more complex effects were sometimes also observed (e.g., an increase in activity or substitution of increases for decreases in discharges). Background muscle firing was also modified in some animals by elimination of labyrinthine signals. These findings suggest that the vestibular system provides tonic influences on the activity of a subset of genioglossal muscle fibers and modulates the discharges of these fibers during some changes in posture. However, a number of other inputs (e.g., from muscle spindles in the tongue musculature) also apparently contribute to regulating muscle firing, such that tilt-elicited changes in activity are retained, albeit in altered form, subsequent to lesions of the inner ear.

Variability in the magnitude of genioglossal muscle responses to tilt was also observed from trial to trial in a particular animal, although these responses did not appear to change systematically over time. Presumably, this response variability was a result of alterations in the excitability of the multiple neural pathways controlling the activity of genioglossal motoneurons. Factors such as changes in attention or cognitive state, as well as the intention to produce a particular voluntary tongue movement, could alter the relative contributions of different sensory inputs to regulating genioglossal firing. These factors are difficult to control for in a conscious animal preparation, and thus variability in posturally related responses is not unexpected for a muscle that participates in many types of movements and whose excitability is influenced by a large number of sensory signals.

It is tempting to speculate that the activity of the horizontally oriented genioglossal muscle fibers that produce tongue protrusion (5, 21) is most directly and powerfully modulated by vestibular signals, and that these inputs contribute to protruding the tongue when the head is oriented nose-up to maintain airway patency. In contrast, each subcompartment of the portion of the genioglossal muscle composed of obliquely oriented fibers could respond to labyrinthine signals elicited by different directions of head tilt (including ear-down rotations); these signals might be highly integrated with a variety of other sensory inputs. This response pattern would account for the disparity between animals with regard to genioglossal responses to tilt and the effects of vestibular lesions on these responses, because, in each case, EMG recordings presumably sampled the activity of different proportions of fibers in the various compartments and subcompartments of the muscle (5, 16, 17, 20, 21). To fully appreciate the relationship between postural changes and genioglossal activity, it will be necessary to develop techniques for recording from individual subcompartments of the muscle. Considering the small size of genioglossus, and thus the limited number of fibers in each subcompartment, this analysis will be challenging.

Removal of labyrinthine inputs produced alterations in genioglossal muscle responses to nose-up tilts that persisted throughout the time course of data collection. In contrast, the effects of peripheral vestibular lesions on posturally related cardiovascular responses resolved after ∼1 wk, at which time the responses were similar to those recorded when vestibular inputs were intact (8, 9). Similarly, the effects of bilateral labyrinthectomies on both the spontaneous and tilt-related activity of respiratory pump muscles diminished within a few days following the lesions (4). It has been suggested that recovery of the ability to rapidly adjust blood pressure and firing of respiratory pump muscles during postural alterations following elimination of vestibular signals reflects multisensory integration by neurons in the caudal vestibular nuclei, such that inputs from muscle, cutaneous, retinal, and perhaps visceral afferents can be substituted for those from the inner ear (10, 31). It thus appears that nonlabyrinthine signals are not substituted for vestibular inputs following loss of the latter signals in the neural pathways that regulate genioglossal firing, as in the circuits that mediate vestibular influences on sympathetic nervous system and respiratory pump muscle activity. Consequently, it seems likely that different populations of brain stem neurons mediate vestibular influences on the discharges of genioglossus and respiratory pump muscles. The results of a recent experiment that used transneuronal tracing techniques also support this notion; this study employed pseudorabies virus recombinants expressing unique reporters to determine whether the same brain stem neurons provided inputs to both genioglossal and diaphragm motoneurons (27). Although overlapping populations of medullary neurons were transneuronally infected by the viral recombinants injected into the diaphragm and genioglossus, only a small proportion of these neurons were infected by both recombinants, showing that the cells mainly independently contributed to regulating the contraction of the two muscles (27). However, further experiments will be required to determine the functional significance of limiting the replacement of vestibular inputs by nonlabyrinthine signals in the pathways that control genioglossal muscle activity, but not other neural circuits that mediate vestibulo-autonomic responses (4, 8, 9).

In summary, the present data collected in conscious felines show that postural changes in multiple directions can elicit alterations in genioglossal muscle activity. These responses are typically modified after removal of vestibular inputs, indicating that the vestibular system contributes to regulating genioglossus firing. Presumably, different compartments and subcompartments within the muscle are activated by changes in head position in particular directions; when mass activity of the muscle is recorded, as was done in the present study, postural alterations can produce variable effects, depending on the particular combination of muscle fibers whose discharges are sampled. Similarly, it seems likely that vestibular signals have more prominent influences on the activity of some genioglossal muscle compartments than others. The relative magnitude of vestibular system contributions to genioglossal firing also may vary over time, depending on an animal's attention, cognitive state, and whether it is preparing to make a particular voluntary tongue movement. Any of these conditions could presumably affect the excitability of the neural pathways that relay labyrinthine signals to genioglossal motoneurons. Further experiments will be required to explore all of these possibilities. In addition, future research should consider whether these conclusions from studies of felines are applicable to humans and other bipeds, which have a different airway geometry than cats. It is possible that a supine posture results in a particularly strong increase in genioglossal discharges in humans who, unlike felines, commonly sleep in the supine position. Nonetheless, the present findings suggest that further considerations of the relationships between postural alterations in multiple planes, vestibular system activation, and genioglossus firing are warranted in both human subjects and experimental animals.


The authors thank Adam Anker, Ryan Mori, Arju Ali, Brian Sadacca, Andrew Maurer, and Anthony Acernese for valuable technical assistance in the completion of these studies.


This work was supported by National Institutes of Health (NIH) Grants R01 DC-03732 and R01 DC-00693 (to B. J. Yates). Core support was provided by NIH grants EY-08098 and DC-05205. B. J. Jian was supported by National Aeronautics and Space Administration Training Grant NGT5–50292 (Graduate Student Researcher Program).


  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract