ILSA SCHWARTZ, PhD Experimental Study Editor
Effect of superior laryngeal nerve on vocal fold function: An in vivo canine model DAVID H. SLAVIT, MD, THOMAS V. MCCAFFREY, MD, PHD, and ERIK0 YANAGI, MD, Rochester, Minnesota
Assessment of laryngeal framework surgery requires an awarenessof the effect of vocal fold mass, stiffness, and position on voice production. The vibratory pattern of the vocal folds during phonation depends on the subglottic pressure as well as the mass and stiffness of the folds. To assess the effect of variations in vocal fold tension with contraction of the cricothyroid muscle on phonation, eight mongrel dog larynges were studied in vivo. Photoglottography, electroglottography, and subglottic pressure were simultaneously recorded as airflow rate and superior laryngeal nerve (SLN) stimulation were varied. Stimulation of the SLN was modified by varying the frequency and voltage of the stimulating electrical signal. Multiple regression analysis of the data revealed a direct relationship between the voltage of SLN stimulation and frequency of vibration (p< 0.001) at constant subglottic pressure. Increases in the stimulating voltage to the SLN also led to an increase in open quotient (p< O.OOI), but no statistically significant change in speed quotient, subglottic pressure, or sound intensity. Changing the frequency of SLN stimulation had only a modest effect on the frequency of vibration. These results are consistent with the reported findings of an increase in frequency and open quotient with increased tension in an in vitro canine model. The glottographic measurement open quotient appears to be an estimator of cricothyroid contraction and longitudinal vocal fold tension, and may be clinically applicable to the assessment of superior laryngeal nerve injuries and laryngeal framework procedures. [OTOLARVNGOL HEAD NECK SURG 1991;105:857,)
A n awareness of the effect of vocal fold mass, tension, and position on voice production is necessary for the objective assessment of laryngeal framework phonosurgery. The vibratory pattern of the vocal folds depends on the mass and stiffness of the vocal folds, as well as the subglottic pressure. To assess the effect of changes in vocal fold tension and mass on phonation, the alteration in the vibratory pattern of the vocal folds with variations in superior laryngeal nerve stimulation (SLNS) and recurrent laryngeal nerve stimulation (RLNS) was examined in an in vivo canine model. The canine larynx is similar to the human larynx, although there are some differences in histology, size, and configuration of the vocal folds. ’.’ In spite of these differ-
From the Department of Otorhinolaryngology, Mayo Clinic and Mayo Foundation. Submitted for publication Sept. 19, 1990; revision received May 3, 1991; accepted May 15, 1991. Reprint requests: David H. Slavit, MD, Department of Otorhinolaryngology, Mayo Clinic, 200 First St., SW, Rochester, MN 55905. 23 I 1I31228
ences, the canine larynx has been an important model for the study of laryngeal phy~iology.’-~ Photoglottography (PGG) and electroglottography (EGG) can be used to study the vibratory pattern of the vocal folds. EGG measures the impedance to a lowcurrent flow across the neck at the level of the vocal folds, and the signal reflects changes in lateral vocal fold contact areas during the glottal PGG requires the use of a photosensor placed on the neck below the vocal folds to measure the transillumination of light through the glottis during phonation. The PGG signal reflects the variations in the cross-sectional area of the glottal aperture during phonation.’.’ Analysis of the PGG and EGG waveforms allows identification of events related to opening and closing of the glottis, and subdivision of the vibratory cycle into the opening phase, closing phase, and closed phase.’ The open quotient (OQ) and speed quotient (SQ), two objective measures of the glottal cycle with potential clinical applications, can then be calculated (Fig. l).9 Electrical stimulation of the laryngeal nerves leads to contraction of the intrinsic laryngeal muscles. In857
OtolaryngologyHead and Neck Surgery
858 SLAVIT et al.
I I I Open phase I
A' -B A'-A
C-B A' -C
Fig. 1. The period of the vibratoly cycle can be divided into an opening phase, closing phase, and closed phase using photoglottography and electroglottography. The opening quotient (OQ),which is the proportion of time the glottis is open during each cycle. and the speed quotient [SQ), which is the ratio of time the glottis is opening to closing, can then be calculated.
Table 1. Means and standard deviations of open quotient, speed quotient, and fundamental frequency
0 54 1 06 268 Hz
0 11 0 40 55 Hz
Open quotient Speed quotient Fundamental frequency
creased stimulation of the recurrent laryngeal nerve (RLN) has been shown by Slavit et al." to increase sound intensity, fundamental frequency, and SQ, and to decrease OQ. These changes in the vibratory pattern of the vocal folds with increased RLNS are associated with a rise in subglottic pressure, reflecting changes in vocal fold tension, mass, and glottic aperture resulting from contraction of the vocalis muscle and the lateral cricoarytenoid. With further increases in RLNS, the subglottic pressure decreased, reflecting stimulation of the posterior cricoarytenoid. Thus, as RLNS changed, the individual intrinsic laryngeal muscles innervated by the RLN were variably activated, based on the specific electrical stimulation to the RLN.
Contrary to the RLN, the SLN innervates only one intrinsic laryngeal muscle, allowing a more reliable analysis of the effect of SLN stimulation on the laryngeal configuration and subsequent change in vibratory pattern, than can be done for RLN stimulation. Studies of humans using high-speed photography during phonation have demonstrated a lengthening and thinning of the vocal folds with an increase in electromyographic activity of the cricothyroid muscle. "," In both canine and human studies, an increase in cricothyroid activity has been shown to increase the fundamental frequenc~.'.~,''-'~ This investigation involved analysis of EGG and PGG waveforms, using an in vivo canine model to examine the effect of variation in SLNS on vocal fold vibrations during phonation. METHODS
Subjects. Eight adult male mongrel dogs, weighing 20 to 25 kg each, were selected. The dogs had not undergone any previous experimental studies and had normal larynges on direct examination. Preparationand Experimental Design Figure 2 illustrates the experimental method, which is similar to that of previous in vivo canine st~dies.',~
Volume 105 Number 6 December 1991
Effect of superior laryngeal nerve on vocal fold function 859
to sup. laryngeal n. \
Humidified warm air source
Recurrent n. 16
Fig. 2. Schematic representation of the set-up for experimental preparation.
Table 2. Multiple regression results of superior laryngeal nerve stimulation amplitude and frequency on fundamental frequency and open quotient Regression weights
Fundamental frequency F (2302) = 65 p < 0.001 Open quotient F (2302) = 133 p < 0.001
Standard error of beta
SLNS amplitude SLNS frequency
SLNS amplitude SLNS frequency
0.21266 0.00002 133.8295 - 0.0268
Standard error of B
' p < 0.001 tNot statistically significant.
Each animal was anesthetized with an intravenous injection of phenobarbital, titrated until loss of the corneal reflex. Direct laryngoscopy was performed to confirm normal laryngeal anatomy. A midline incision from the mandible to the sternum was then performed. The strap and sternocleidomastoid muscles were retracted laterally, exposing the larynx and trachea. A distal tracheotomy was performed and an endotracheal
tube was placed pointed inferiorly to permit the dog to breathe spontaneously. A more proximal tracheotomy was performed through which a cuffed tracheotomy tube was positioned with the tip 10 cm below the vocal folds and pointed rostrally. With meticulous dissection the RLN and external branch of the SLN were identified on both sides. The nerves were isolated and electrodes were applied to the SLN, just proximal to the crico-
OtolaryngologyHead and Neck Surgery
860 SLAVlT et al.
320 310 300 N
Flow 1,000 rnl/sec
0 SNLS freq 200 Hz
Flow 800 rnl/sec I SNLS freq 100 Hz
Flow 800 mllsec SNLS freq 50 Hz
SNLS amplitude, volts Fig. 3. Fundamental frequency (Fo) as a function of the amplitude of superior laryngeal nervestimulating (SLNS) for four configurations of frequency of SLNS and airflow rate. The pulse width of the SLNS stimulus was constant at 1.5 msec, whereas the recurrent laryngeal nerve was stimulated at a frequency of 80 Hz with 0.5 volt pulses of 1.5msec duration. Note the direct significant relationship ( p < 0.001) between SLNS amplitude and Fo.There was no significant change in Fo as SLNS frequency was changed.
thyroid muscle, and to the RLN, approximately 4 cm inferior to the cricoid cartilage. A gauze silver electrode was applied to the nerve and insulated from the surrounding tissue. Direct observation verified electrical isolation between RLNS and SLNS, with no arytenoid adduction or phonation elicited by maximal SLNS. The cephalid tracheotomy tube was inflated to form a seal and air supplied from a compressed air canister was passed through the tube, up the trachea, and through the larynx. The air was heated to 37" C and humidified to more than 95% relative humidity by being bubbled through a Bennett Cascade unit before being delivered to the larynx. Airflow rate was controlled by a valve and measured by a pneumotachograph. A Lab Crest Flowrater was used before every experiment to calibrate the airflow meter. To measure subglottic pressure, a catheter tip Celestone pressure transducer was placed into the trachea between the cricoid cartilage and the first tracheal ring, with its tip 2 cm below the vocal folds. Calibration of the pressure transducer to 60, 120, and 180 mm of H,O was performed with a pressure gauge before every experiment. Recording electrodes for the EGG (Synchrovoice) were sutured on each side of the thyroid cartilage, with the ground electrode secured to the adjacent strap mus-
cles. A Centronics OSD-2 phototransducer was fixed against the trachea, approximately 2 cm below the cricoid cartilage. A direct current light source supplied through a rigid bronchoscope positioned above the larynx provided the supraglottic illumination for the PGG. The microphone for the Larson-Davis 800B sound intensity meter was mounted 30 cm from the vocal folds. To provide a flat frequency response in the phonation spectrum with suppression of low-frequency background noise, the C-scale of the sound intensity meter was used. A Grass model S44 nerve stimulator with stimulus isolation unit Grass SIU5 was used to provide variable voltage stimulation to the SLN. Voltage levels ranged from 0.5 volts to 2.0 volts and were confirmed using a Tektronix oscilloscope. The frequency of stimulation was also systematically varied over a range from 50 to 400 Hz, while the pulse width was held constant at 1.5 milliseconds. The RLN was stimulated with a Grass model S 11 nerve stimulator, with stimulus isolation unit Grass SIU5 providing a voltage of 0.5 volts and a frequency of 80 Hz with a 1.5 milliseconds pulse duration. Eight dogs were stimulated to phonate. While a constant stimulus was delivered to the RLN, the SLNS voltage, SLNS frequency, and airflow rate were systematically varied while the glottographic signals, sub-
Volume 105 Number 6 December 1991
Effect of superior laryngeal nerve on vocal fold function 861
SNLS amplitude, volts Fig. 4. The effect of amplitude of superior laryngeal nerve stimulation [SLNS) on open quotient [OQ) for four configurations of frequency of SLNS and airflow rate. There was a significant direct relationship ( p < 0.001) between SLNS amplitude and OQ and no significant relationship between SLNS frequency and OQ. The pulse width of SLNS stimulus was 1.5 msec. Recurrent laryngeal nerve stimulation was a 0.5 pulse at 1.5 msec duration at 80 Hz.
glottic pressure, and sound intensity were simultaneously recorded. Certain configurations were retested every 60 minutes to determine if there were changes in the nerve conduction or vibratory nature of the vocal folds over the time of the experiment.
The PGG and EGG signals recorded were amplified to the 2 5 volt range and band-pass-filtered with fivepole Butterworth filters at 1.0 Hz and 2.5 kHz to stabilize the baseline and anti-alias for digital signal processing. The EGG and PGG signals were then lowpass-filtered at 4000 Hz and digitized at 10,000 Hz. An AST 20868 microcomputer with an 80287 mathematical co-processor and ASYSTANT software for high-speed data acquisition and analysis were used. A one-second sample of stable phonation was stored on the hard disk of the microcomputer for later analysis. The airflow rate, subglottic pressure, and sound intensity were recorded simultaneously on a Beckman R6 11 recorder. Analysis of data. For each trial of stimulated phonation, a stable 30-second portion was chosen from the second of waveform data that was recorded and stored on the hard disk. Within this segment, ten consecutive waveforms were analyzed and averaged to calculate a
mean value for OQ, SQ, and fundamental frequency (Fo). Points of glottal opening and closing were determined as previously described by Berke et a1.’ and Childers et al.I4 Data analysis and processing was performed with ASYSTANT software. Multiple regression analysis was performed to determine any statistically significant ( p < 0.05) changes in OQ, SQ, and Fo with alteration of the independently controlled variables of airflow rate, SLNS amplitude, and SLNS frequency. Analysis included the study of the effect of SLNS frequency and amplitude on subglottic pressure and sound intensity. Multiple regression was performed using CSS statistical software. RESULTS
Eight in vivo canine laryngeal models were studied under 50 different combinations of airflow rate, SLNS amplitude, and SLNS frequency, producing phonation in the canine modal register. On the basis of a consistency of the vibratory pattern for the specific conditions retested every 60 minutes, the nerve conduction and vibratory nature of the vocal folds appeared to be constant throughout the experiment. Table 1 lists the mean and standard deviation for Fo, OQ, and SQ. Multiple regression analysis revealed that with increasing SLNS amplitude there was a significant
862 SLAWet 01.
OtolaryngologyHead and Neck Surgery
Fig. 5. The photoglottographic (PGG) waveform is shown for four different amplitudes of superior laryngeal nerve stimulation (SLNS) with SLNS frequency and width held constant. When the SLNS amplitude was 0.7 volts (upper leftPGG), the fundamental frequency (Fo) was 294 Hz and the open quotient (OQ]was 0.61. With an increase in SLNS amplitude to 0.8 volts [upper righlPGG),Fo increased to 323 HZ and OQ rose to 0.66. Further increases in SLNS amplitude to 0.9 volts (/ewer /eft PGG] and 1.0 volts (/ewer right PGG) led to a further increase in both Fo and OQ, from 0.67 at 339 Hz to 0.69 at 345 Hz.
( p < 0.001) increase in Fo (Table 2, Fig. 3). On the contrary, SLNS frequency did not have a significant relationship with Fo. SLNS amplitude was also directly related with the OQ ( p < 0.001). whereas SLNS frequency was shown to have no significant effect on OQ (Table 2, Fig. 4). Neither the SLNS amplitude nor frequency was shown to have a significant relationship with SQ, subglottic pressure, or sound intensity. Figure 5 shows representative waveforms plotted for one subject at four different SLNS amplitudes. As the SLNS amplitude increased, the Fo rose along with the ratio of the open portion to glottic period (OQ). In this particular subject, the OQ was 0.61 at 294 Hz, 0.66 at 323 Hz, 0.67 and 339 Hz, and 0.69 at 345 Hz. The speed quotient was not significantly altered as SLNS amplitude increased from 0.7 volts to 1.0 volts. DISCUSSION
This investigation studied the effect of SLNS on the laryngeal resistance and vibratory nature of the vocal folds within the canine modal register. The results of this study are limited because the larynx is not that of the human and phonation is not produced under nomial physiologic conditions. These are important factors to
consider when applying results from models of vocal fold vibration to human phonation. The experimental finding that Fo increased with increased SLNS amplitude confirms other reports that increased contraction of the cricothyroid muscle raises the pitch.3,'.'5 Changes in the frequency of SLNS appeared to have no significant effect on Fo. When the frequency of SLNS was below 50 Hz, noncontinuous phonation was produced. With an increase in SLNS frequency to 50 Hz, phonation became smooth and continuous, probably as a result of placement of the muscle in tetany. Increases in the frequency of SLNS above 50 Hz had no significant effect on Fo for the SLNS amplitude range studied. However. an increase in Fo was obtained with increased SLNS amplitude. as a result of recruitment of more muscle fibers. The electrical stimulation of the SLN in this in vivo model was not physiologic, in that all neurons discharged simultaneously rather than irregularly and in rotation. In normal physiologic situation, the neurons do not discharge synchronously, and the frequency of stimulation can modulate cricothyroid muscle contraction and Fo. Increasing Fo and OQ with increased cricothyroid contraction confimis the work of others.'" The direct
Volume 105 Number 6 December 1991
relationship of increased cricothyroid contraction to Fo and OQ agrees with studies in the excised canine larynx.I2 The increased longitudinal stress of the vocal folds in the anteroposterior direction with increased cricothyroid contraction lengthened, stiffened, and thinned the vocal folds.I6 The decrease in mass per unit area of the length, decrease in compliance, and increase in elasticity of the vocal folds resulted in a greater rate of vibration and higher pitch. Because of the decreased compliance, the compression of the vocal folds as they collided during the vibratory cycle decreased and the folds remained approximated for a shorter portion within each cycle. There was no significant relationship between SLNS and the SQ, although the opening phase decreased in absolute time because of a smaller excursion and the closing phase decreased because of increased elasticity. Increased SLNS lengthened, stiffened, and thinned the vocal fold as well as changed the vibratory rate and pattern, yet there was no significant effect on subglottic pressure. Thus for a constant airflow, SLNS did not alter subglottic pressure, indicating that contraction of the cricothyroid muscle had no effect on the glottic resistance. Neither alteration in amplitude nor frequency of SLNS effected subglottic pressure. Similarly, SLNS had no significant effect on sound intensity, although a trend existed of decreased sound intensity, with strong contraction of the cricothyroid muscle resulting from a decrease in the amplitude of vocal fold vibration. The results obtained using this in vivo canine model were similar to those obtained from previous work with an excised canine model. l 3 Increased cricothyroid contraction, stimulated in the in vivo model and simulated in the excised model led to a statistically significant increase in Fo and OQ in both models. Cricothyroid contraction had no significant effect on SQ, subglottic pressure, and sound intensity. Although the intrinsic laryngeal muscles cannot be activated by motor nerves in an excised larynx, the actions of these muscles can be simulated. The use of an excised larynx model provides a setting to test new concepts while enabling manipulation of variables not easily controlled in in vivo studies. Clinical Implications
Our results showed that increased longitudinal tension, caused by increased SLNS and contraction of the
Effect of superior laryngeal netve on vocal fold function 863
cricothyroid muscle, increases the Fo and OQ, with no significant effect on SQ. The glottographic measure of OQ appears to be an estimator of longitudinal vocal fold tension. The increased tension of the vocal folds after type 2 and 4 thyroplasties should also cause an increase in Fo and OQ REFERENCES
1. Bradley OC, Graham T. Topographical anatomy of the dog. New York: MacMillian Co, 1959. 2. Berke GS, Moore DM, Hantke DR, Hanson DG,Gerratt BR, Brustein F. Laryngeal modeling: theoretical, in vitro, in vivo. Laryngoscope 1987;99:871-81. 3. Hast MH. Physiological mechanisms of phonation: tension of the vocal fold muscle. Acta Otolaryngol 1962;62:309-18. 4. Rubin HJ. Experimental studies on vocal pitch and intensity in phonation. Laryngoscope I963;73:973- 1015. 5. Childers DG. Krishnamurthy AK. A critical review of electroglottography. CRC Crit Rev Biomed Eng 1985;12:131-61. 6. Fourcin AHJ. Laryngoscopic examination of vocal fold vibration. In: Wyke B, ed. Ventilatory and Phonatory Control Mechanisms. Oxford: Oxford University Press 1975:315-33. 7. Sonesson B. On the anatomy and vibratory pattern of the human vocal folds: with special reference to a photo-electrical method for studying the vibratory movements. Acta Otolaryngol 1960;156(~uppl): 1-75. 8. Baer T, Loefquist A, McGarr NS. Laryngeal vibrations: a comparison between high-speed filming and glottographic techniques. J Acoust SOCAm 1982:73:1304-80. 9. Kitzing PL. Clinical application of combined electro- and photoglottography. I.A.L.P. Conference Proceedings. Copenhagen 1977;1:528-39. 10. Slavit DH, McCaffrey TV, Yanagi E. Effect of recurrent laryngeal nerve on vocal fold functions: an in-vivo canine model. Trans Coll Surg (In press). I . Arnold GE. Physiology and pathology of the cricothyroid muscle. Laryngoscope 1961;71:687-753. 2. Hirano M, Ohala J , Vennard W. The function of laryngeal muscles in regulating fundamental frequency and intensity of phonation. J Speech Hear Res 1969;12:616-28. 3. Slavit DH, Lipton RJ, McCaffrey TV. Glottographic analysis of phonation in the excised canine larynx. Ann Otol Rhino1 Laryngol 1990:396-402. 4. Childers DG, Hicks DM, Moore GP. A model for vocal fold vibratory motion, contact area, and the electroglottogram. J Acoust SOCAm 1986;80(5): 1309-21. 5. Berke GS, Moore DM, Gerratt BR, et al. Effect of superior laryngeal nerve stimulation on phonation in an in-vivo canine model. Am J Otolaryngol 1989;10:181-7. 16. Hirano M. Morphological structure of the vocal cord as a vibrator and its variations. Folia Phoniatr 1974;26:89-94.