The effects of spontaneous swallows
on breathing before, during, and after solitary swallows were
investigated in 13 awake goats. Inspiratory (TI) and
expiratory (TE) time and respiratory output were determined
from inspiratory airflow [tidal volume (VT)] and
peak diaphragmatic activity (Diapeak).
The onset time for 1,128 swallows was determined from pharyngeal
muscle electrical activity. During inspiration, the later the
swallowing onset, the greater increase in TI and
VT, whereas there was no significant effect on
TE and Diapeak. Swallows in early expiration
increased the preceding TI and reduced TE,
whereas later in expiration swallows increased TE. After
expiratory swallows, TI and VT were reduced whereas minimal changes in Diapeak were observed. Phase
response analysis revealed a within-breath, phase-dependent effect of
swallowing on breathing, resulting in a resetting of the respiratory
oscillator. However, the shift in timing in the breaths after a swallow
was not parallel, further demonstrating a respiratory phase-dependent effect on breathing. We conclude that, in the awake state, within- and
multiple-breath effects on respiratory timing and output are induced
and/or required in the coordination of breathing and swallowing.
 |
INTRODUCTION |
BOTH BREATHING AND
SWALLOWING are continual ongoing events in mammals, although at
different frequencies. As a result of the shared use of the upper
airway, it is extremely important that the motor pattern generators
associated with breathing and swallowing are tightly coupled to provide
effective coordination for cleansing and removing secretions from the
lower airways (3, 8, 16). A loss of coordination between
these pattern generators is associated with dysphasia, weight loss,
coughing, and pulmonary aspiration leading to pneumonia. However, the
exact neural and physiological nature of the interconnections between
the swallowing and breathing pattern generators is unknown.
Nevertheless, the functional relationships between these generators
have been investigated for decades.
Although the frequency of spontaneous swallows during the different
phases of respiration varies among species, swallowing has been
generally reported to inhibit breathing in human, dogs, rabbits, and
cats (1, 2, 5, 7, 8, 22, 23, 25, 29). For example,
swallows during inspiration or expiration are reported to increase the
phase of respiration in which it occurred. In the few studies that
measured respiratory output [e.g., tidal volume (VT) and
electrical activity of the diaphragm and laryngeal abductors],
swallowing reduced inspiratory output (11, 23, 24). In
addition, the observations of swallows during the transition between
inspiration-expiration or expiration-inspiration are infrequent
(18, 22), but if observed they disrupt respiratory timing
in humans (27).
The neural substrates that generate and coordinate the motor patterns
for swallowing and breathing are located in the dorsomedial and
ventrolateral brain stem (1, 4, 16, 14). To understand the
interaction between these pattern generators, several interpretative models have been applied, mostly using anesthetized or decerebrate models with superior laryngeal nerve stimulation (7, 19, 27). Few studies have examined this interaction on the
swallowing and respiratory pattern generators in unanesthetized animals
and humans (20, 23-25).
The aim of this study was to investigate in unanesthetized, awake goats
the effect of spontaneous swallows on respiratory output and timing. We
hypothesized that a swallow in either phase of respiration would
produce a within-breath effect to increase respiratory timing, enough
to cause a phase delay but not a resetting of the respiratory rhythm.
We also hypothesized that swallows would not have an effect on the
timing or VT of the breath before the swallow but would
reduce the VT if the swallow occurred during inspiration.
| |
METHODS |
Animals.
Thirteen adult goats of various breeds that were used in this study
received humane care in compliance with the Guide for the Care
and Use of Laboratory Animals formulated by the National Research
Council, 1996. The study protocol was approved by the Institutional
Animal Care Committee of the Medical College of Wisconsin.
Surgical procedures.
Surgery was performed to implant chronic electromyographic (EMG)
electrodes in the diaphragm (EMGDia), thyropharyngeus
(EMGTP), posterior cricoarytenoid (EMGPCA), and
thyroarytenoid (EMGTYA) muscles. In four goats, a
tracheostoma was created at the level of the eighth tracheal ring to
isolate the upper airway. The animals received an intravenous injection
of ketamide (Ketaset) and xylazine (12:1 ratio, 15 mg/kg) for induction
of anesthesia before intubation. After intubation for mechanical
ventilation, general anesthesia was maintained with 1.5-2.5%
halothane in oxygen (sufficient to eliminate the withdrawal reflex and
any signs of pain.)
EMG electrode placement has been previously described
(10). Briefly, Teflon-coated 32-gauge stainless steel
bipolar microelectrodes were inserted into the muscles defined above.
For implants into airway muscles, a midline incision was made on the
ventral surface of the neck from the hyoid bone to 4 cm below the
thyroid cartilage, exposing the lateral aspect of the pharynx. The
EMGTP electrode was sewn into the thyropharyngeus midway
between the posterior midline of the pharynx and the insertion on the
thyroid cartilage 0.5 cm below the cranial laryngeal nerve. To
implant the EMGPCA, a small window was made at the inferior
edge of the thyroid cartilage, and the tissue was dissected between the
anterior wall of the esophagus and the posterior wall of the cricoid
cartilage to expose the posterior cricoarytenoid. For placement of the
EMGTYA electrode, a U-shaped window was made in the lateral
wall of the thyroid cartilage to expose the thyroarytenoid. The
EMGSP electrode was sewn into the stylopharyngeus muscle
close to its disappearance under the hypopharyngeus. The wires were
looped in the subcutaneous layer of connective tissue and exited on the
lateral ventral surface of the neck. For the EMGDia, a
lateral thoracotomy was performed between the ninth and tenth ribs
midway between the sternum and spine. The diaphragm electrodes were
implanted in the costal portion of the diaphragm and exteriorized next
to the incision. After the above instrumentation, the carotid arteries
were elevated bilaterally so that they could be easily catheterized for
monitoring of arterial blood pressure during the experimental studies.
A 5-cm segment of each carotid artery was elevated subcutaneously and
sutured in place.
For at least 24 h after surgery, laboratory personnel frequently
inspected the animals. The animals received daily intramuscular antibiotics (ceftiofur sodium, 2 mg/kg), and their rectal temperature, eating habits, and behavior were monitored daily. These measures indicated that the goats were in good health and fitness during recovery from surgery and the subsequent experimental period.
Methods of measurement and experimental design.
For measurements of airflow in the animals without a tracheostoma, a
tight-fitting, custom facemask was connected to a one-way breathing
circuit. In the animals with a tracheostoma, airflow measurements were
made via an 8-Fr cuffed tracheostomy tube that was inserted into the
trachea, cuff inflated, and connected to a one-way breathing circuit
with a two-way nonrebreathing valve (model 2600, Hans Rudolph). A
pneumotachograph was connected in-line on the inspiratory side of the
nonrebreathing valve and connected to a differential pressure
transducer to measure inspiratory airflow only. The proximal ends of
the EMG wires were connected via microclips to a Grass recorder for
signal processing and recording. The EMG signals were filtered at a
band pass of 3-500 Hz. A carotid artery was catheterized at least
2 days before initiation of experimental studies. Arterial blood
pressure was monitored during all studies by connecting the arterial
catheter via a Statham blood pressure transducer to a Grass recorder.
The airflow, raw EMG signals, and arterial blood pressure were sent to
a CODAS computer data acquisition system at a sampling rate of 250 Hz
for display, digital recording, and analysis.
Before initiation of data collection, the goats were acclimated for
several hours a day for a week in a stanchion where they were connected
to the recording system. On the days the goats were studied,
respiration, upper airway EMG activity, and blood pressure were
continuously monitored for a minimum of 2 h. Chewing and other
movement artifacts were eliminated from the analysis. All goats were
studied in the awake state in the prone recumbent position. Respiratory
rate and arterial blood pressure were normal throughout the
experimental period, indicating that the goats were in good health,
relaxed, and free of pain.
Data analysis.
Respiratory airflow and EMG signals were processed and analyzed
(WinDaq, DATAQ Instruments). Raw EMG data were full-wave rectified and
passed through a moving time averager (time constant of 0.1 s) to
obtain an integrated EMG signal. The integrated EMGTP,
EMGPCA, and EMGTYA signals were analyzed to
obtain the occurrence of swallowing. The absolute time of a swallow was
determined at the peak of a 10-fold increase in the integrated
EMGTP and EMGTYA activity with a 200-ms
duration and without signs of movement artifacts. The start of a
swallow was then set at 0.15 s before peak EMGTP
activity. Even though the TP muscle is activated in the later part of
the recruitment order of pharyngeal muscles, the onset of activation is
within 0.2 s of the initiation of the oropharyngeal phase of swallowing (8, 9, 18). In this study, spontaneous swallows were considered as nonfeeding swallows that were initiated reflexively by the accumulation of oropharyngeal secretions. These secretions were
presumably sensed by afferent receptors in the laryngeal mucosa
innervated by the trigeminal, glossopharyngeal, and vagal nerves
(1, 8, 15, 16). None of the swallows analyzed were related
to mastication.
For the analysis of ventilation, a two-state computer algorithm
(inspiration-expiration) was used for the automatic detection of the
breath-by-breath calculations for VT, duration of
inspiration (TI), and duration of expiration
(TE) from the airflow signal. Peak EMGDia
activity (Diapeak) was obtained for each breath. In this
two-state computer algorithm, inspiration was determined when flow
exceeded zero respiratory flow by 0.01 l/s for a minimum duration of
0.2 s. The start of expiration was determined when a pause in the
EMGDia activity and in the inspiratory flow longer than
0.35 s was detected. Although this period was shorter than the
reported deglutition apnea in animals and humans (5, 17, 22-24, 27, 30), we felt that it was a conservative estimate of the establishment of expiratory cycle. In addition, our observations in goats led us to the use of these criteria, because deglutition apneas, if present during inspiration, were very short. Therefore, we
believe that the determination of the phase of breathing was accurately
and reliably determined. The absolute time for the start of inspiration
for each breath was recorded along with the time of each swallow. The
total duration of a breath (Ttot) was calculated as the sum of
TI and TE. The inspiratory duty cycle (TI/Ttot) and the ratio of VT to TI
(VT/TI) were calculated for each breath.
Three types of analysis of the effects of swallowing on breathing were
made. First, to examine the general effects of swallowing on breathing,
swallows were categorized into one of four phases of ventilation (Fig.
1): expiration (SwE), the
transition from expiration to inspiration (SwEI),
inspiration (SwI), and the transition from inspiration to
expiration (SwIE). The effects of SwE,
SwEI, SwI, and SwIE on
TI, TE, Ttot, VT,
TI/Ttot, VT/TI, and
Diapeak (expressed as a percentage of control) on the
breath before (n 
1), during (n), and
after (n + 1) the swallows were evaluated. Second, the gamma
analysis (
) evaluated whether the time of occurrence of a
swallow, within either inspiration (I) or expiration
(E), had an effect on the breath before or during a
swallow (23, 24). For this analysis, five consecutive
breaths were evaluated, with a solitary swallow in the fourth breath
(Fig. 2, A and B). In this analysis, a breath was considered to start with expiration followed by inspiration. Swallows with movement or augmented breaths within the set of five consecutive breaths were eliminated from this
analysis. Averages for TI, TE, VT,
and Diapeak were calculated from the first, second, and
fifth breaths of the set of five breaths and used as control values.
The value for I was calculated as the time from the
beginning of inspiration to the onset of the solitary swallow,
expressed as percentage of the control TI. The value for
E was calculated as the time from the beginning of expiration to the onset of the solitary swallow, expressed as percentage of the control TE. Values for TI,
TE, VT, and Diapeak for the
previous breath (n 1) and the breath during
(n) the swallow were expressed as a percentage of its
control value and were plotted against I or
E.
View larger version (25K):
[in this window]
[in a new window]
|
Fig. 1.
Categorization and examples of swallows. A: solitary
swallows with no swallows in the prior or following breaths were
classified into 1 of 4 categories on the basis of when they occurred
during the breath: SwE, from the end of the previous
inspiration (Insp) to 0.1 s before the start of the next
inspiration; SwEI, from 0.1 s before and 0.15 s
after the start of inspiration; SwI, from 0.15 s after
the start of inspiration to 0.1 s before the end of inspiration;
SwIE, from 0.1 s before the end of inspiration. Exp,
expiration;  , inspiratory flow. The percentages of control
respiratory values evaluated in this study were compared for
SwE, SwEI, SwI, and
SwIE on the breath before (n 1), during
(n), and after (n + 1) the swallows (see
METHODS). B: examples of solitary
SwE and SwI. BP, blood pressure; TP,
thyropharyngeus; TPMTA, moving time average of TP;
SP, stylopharyngeus; SPMTA, moving time average of SP; Dia,
diaphragm; DiaMTA, moving time average of Dia. Dotted lines
(a and c) mark the beginning of the swallows.
Notice the electrical activity in Dia during expiration just before and
after the start of the swallow. This activity in the Dia is often
associated with a brief inspiratory flow and is termed
Schluckatmung or a "swallow breath." Arrows indicate a
short absence of Dia activity either during a Schluckatmung
(b) or an inspiration (d). C: examples
of solitary SwEI and SwIE. Dotted lines
indicate the beginning of the swallows.
|
|
View larger version (12K):
[in this window]
[in a new window]
|
Fig. 2.
Illustration of inspiratory (I; A) and
expiratory (E; B) gamma analysis.
I was calculated as the period of time from the
beginning of inspiration to the start of the swallow during
inspiration. E was calculated as the period of time from
the beginning of expiration to the start of a swallow during
expiration. Inspiratory time (TI), expiratory time
(TE), tidal volume (VT), and peak diaphragmatic
activity (Diapeak) for the breaths before and during the
swallow were calculated. Control values were calculated from the 2 breaths before and 1 breath after the breaths used in this analysis.
|
|
Finally, to evaluate the effect of swallows on the oscillations of the
respiratory pattern generator, a phase response analysis (similar to
the analysis of biological oscillators) was performed (26, 27,
31). In this analysis, six consecutive breaths were evaluated,
with a solitary swallow in the fourth breath. The old phase (
) was
defined as the time from the beginning of inspiratory flow of the
fourth breath to the onset of the swallow (Fig.
3A). The subsequent cophases
(
) were defined as the time from the onset of swallow to the
beginning of inspiratory flow of breaths 5 (n1) and 6 (n2) and to the end of expiration
of breath 6 (n3). The
cophase for the breath before the swallow (n
1) was calculated as the time from the beginning of the
swallow backward to the beginning of inspiratory flow of breath
3. Movement or augmented breaths in any of these breaths
eliminated the set from analysis. Ventilatory values from breaths
1, 2, and 6 of the set were averaged and
used as controls. All values of and were expressed as a
fraction of control values. Values of VT and
Diapeak in this analysis were calculated as a percent of
control values.
View larger version (14K):
[in this window]
[in a new window]
|
Fig. 3.
Phase response analysis. In this analysis, a set of 6 consecutive breaths (without movements and augmented breaths) was
evaluated, with a solitary swallow in the 4th breath. A: the
old phase () was calculated from the start of inspiration to the
start of the swallow. Cophases () were calculated from the start of
the swallow to the start of inspiration of the previous breath
(n 1) and the start of the 3 subsequent breaths (n1,
n2,
n3, respectively). In a
hypothetical model in which swallowing had no effect on respiration
rhythm, the normalized measures of and (expressed as a fraction
of control) would be equal to 1 ( + = 1).
B: plot of subsequent n vs. in
which a family of swallows that occurred during respiration had no
effect on respiratory rhythm. In this hypothetical example, the slope
of n1 vs. would be 1, and
the effect on any subsequent breaths (n) would also be
anticipated by the equation n = + n. As a result, the slope of the relationship between the
onset of swallowing and respiration rhythm would be represented by a
series of linear parallel lines with a slope of 1.
|
|
Following the above definitions of and , the effect of
swallowing on the respiratory rhythm can be assessed, as previously presented by Paydarfar et al. (27). As defined in Fig.
3A, the normalized measures of and would equal 1 ( + = 1) if swallowing had no effect on rhythm at any
time during the respiratory cycle, and the slope of
n1 vs. would be 1.
Furthermore, as demonstrated in Fig. 3B, the effect on any subsequent breaths (n) would be anticipated by the equation,
n = + n with a similar
slope. A phase advance or delay in the onset of breathing would be
anticipated if swallowing has an effect on the respiratory rhythm
generator to shorten or lengthen inspiration or expiration. If
swallowing had more than a transient effect on the respiratory rhythm,
then a permanent shift or resetting of the subsequent start of
ventilation would occur. To evaluate the possible phase shift in the
phase response curves, the difference in the onset from the previous
breath (n1 n 1,
n2 n1,
n3 n2) for each swallow in all the animals was calculated and averaged every 0.05 interval of . Given
sequential breaths and the above equation for
n, a parallel shift in the
n curves would result in a difference of 1. A
phase delay would result in an increase in the difference between
n curves relative to the previous swallow,
whereas a phase advance would result in a decrease in the difference. To estimate the potential phase shift of the breath before the swallow,
the difference between n 1 and
a nonaffected old phase (ideal) with a slope
of 1 was calculated and averaged every 0.05 interval of . In this
later comparison (n 1 ideal), the expected difference
would be zero if there were no difference between the curves for any
value of .
The phase transition from inspiration to expiration (IE) for the
breaths with swallows was estimated by calculating the half-cycle period relative to a specific old phase that began at the start of
inspiration. The cophase (
IE) in this analysis was
calculated as the period of time from the start of the swallow to the
start of expiration. The average phase transition was then estimated for the breaths with swallows by solving for the relationship of vs. IE by using a third-order polynomial equation for an intercept of zero.
Statistical analysis.
The average rate of swallowing and its standard deviation were
calculated for each goat, and a Kolmogorov-Smirnov test was performed
to test whether the rate of swallowing significantly deviated from
Gaussian distribution (P < 0.05). For the
categorization of swallows, a Student's t-test was
performed on the means of the percent change of TI,
TE, Ttot, VT, TI/Ttot,
VT/TI, and Diapeak for the breaths
before, during, and after the swallows, compared with no change (100%)
within SwE, SwEI, SwI, and
SwIE (P < 0.02). A comparison was also
made using a one-way ANOVA between the means of the percent change for
the breaths before, during, and after the swallows within
SwE, SwEI, SwI, and
SwIE (P < 0.02).
In the gamma analysis, a linear regression analysis was performed on
TE, TI, VT, and Diapeak
(expressed as a fraction of its control) on the previous breath, and
the breath in which the swallow occurred was evaluated for
I and E (P < 0.02). The
slopes and intercepts were compared to determine whether the time of
occurrence of the swallow (I or E) during
inspiration or expiration had an effect on respiratory parameters from
the breath before and during the swallow.
In the evaluation of the phase shift, the mean value of
n1 n 1, n2 n1, and
n3 n2 at every 0.05 interval of was tested to determine whether it differed significantly from a
hypothetical value of 1 using a Student's t-test
(P < 0.02). Similarly, the mean of
n 1 ideal was compared with a value of 0 by using a
Student's t-test (P < 0.02).
| |
RESULTS |
During expiration, the normal phasic respiratory activity of
EMGTP was inhibited just before its recruitment for a
swallow (Fig. 1B, a). Commonly associated with
the swallows during expiration was a small burst of activity in the
EMGDia (Fig. 1B, b). During late
expiration, swallows were associated with a burst of inspiratory airflow, usually termed a "swallow breath" or
Schluckatmung. However, if a swallow occurred during the
first third of expiration, inspiratory flow was not seen. During
inspiration, phasic EMGSP activity was inhibited at the
onset of EMGTP swallow activity (Fig. 1B,
c). A short inhibition, or pause, in EMGDia
activity occurred midway through EMGTP swallow activity
(Fig. 1B, d). In all goats (with an intact upper
airway or with a tracheostomy), a short pause (<0.1 s) and a
significant decrease in inspiratory airflow were seen during swallows
after increases in EMGTP activity and decreases in
EMGDia.
The average rate of swallowing for all the goats was 2.88 ± 1.22 swallows/min (Table 1). Goats with
tracheotomies had an average rate of 1.86 ± 0.31 swallows/min,
whereas goats with intact airways presented with an average rate of
3.3 ± 1.9 swallows/min. On average, 12.4% of the swallows from
each goat were not used in the analysis because of movement artifacts
or the presence of more than one swallow within a breath. The average
duration of the moving time average of a swallow in the TP signal was
0.25 ± 0.03 s. No pharyngolaryngeal or swallowing
dysfunction was observed in any goat in this study.
Effects of spontaneous SwE, SwEI,
SwI, and SwIE on breathing.
In this analysis, shown in Fig.
4, a total of 1,128 solitary swallows met the criteria for evaluation during the
observation of 7,573 breaths. In the SwE analysis (317 swallows, 28% of total), TE was significantly increased in
the breaths before, during, and after the swallow. In contrast,
TI decreased during the inspiration after the
SwE. As a result, Ttot was increased in the breaths before,
during, and after the swallows despite the decrease in TI
after the swallow. The inspiratory duty cycle, TI/Ttot, was decreased in the subsequent breath after SwE because of the
decrease in TI and increase in TE. The
VT just before and after SwE was significantly
reduced, as well as VT/TI. Similarly,
Diapeak activity was decreased in the inspirations before
and after the SwE. Therefore, swallows during early and mid
expiration had an immediate effect on timing during expiration
(increasing) and the subsequent inspiration (decreasing), while having
a wider effect over time by decreasing respiratory output, as indicated
by VT and Diapeak on the breaths before and
after the swallow.
View larger version (48K):
[in this window]
[in a new window]
|
Fig. 4.
Categorical analysis of respiratory data for different
phases of respiration for the breath before (n 1),
during (n), and after (n + 1) the swallow.
Phase of respiration in which the swallows occurred is defined in
METHODS and Fig. 1A. Number of swallows analyzed
for each category: SwE = 317, SwEI = 296, SwI = 507, and SwIE = 8. A: TI. B: TE.
C: total duration of the breath (Ttot). D:
VT. E: TI/Ttot. F:
VT/TI. G: Diapeak. All
values are expressed as a percentage of control and are presented as
means with standard deviation bars. Control values for each animal were
estimated from breaths without artifacts or swallows as defined in
METHODS and were used to normalize each animal's
individual observation. * Means are significantly different from
control (P < 0.01).  Means are significantly
different from n breath (P < 0.01).
|
|
Swallows during the EI transition (296 swallows, 26% of total)
provided a slightly different pattern (Fig. 4). Immediately after
SwEI, TI was significantly decreased, whereas
no significant changes were found in TE or Ttot in the
breaths before, during, or after the swallow. A small increase in
TI/Ttot was observed in the breath after SwEI.
After SwEI, VT was significantly decreased and
then increased in the subsequent breath. No changes in
VT/TI were seen, demonstrating that
VT and TI decreased proportionally in the
immediate inspiration but then increased in the subsequent breath. In
contrast, Diapeak was not significantly altered in the
breaths before, during, or after SwEI. Therefore, swallows during the EI transition solely affected the subsequent inspiratory timing and, consequently, VT, but not total output, as
suggested by Diapeak.
In the SwI analysis (507 swallows, 45%), TI
was greatly increased during SwI, whereas no effect was
seen in TI in the breaths before or after (Fig. 4).
Swallows during SwI had no effect on TE. As a
consequence, TI/Ttot was significantly increased.
VT was slightly increased in the breath with a swallow and
in the subsequent breath, whereas a significant decrease in
Diapeak was observed during SwI. However,
VT/TI was significantly decreased during the
inspiratory swallow, primarily as a result of the increases in
TI. These results suggest that swallows during inspiration affected respiratory timing but not output.
Because so few swallows were observed during the IE transition (8 swallows; 0.01%), great variability and lack of significant changes were observed in these data.
Within-breath effects of swallows: gamma inspiratory and expiratory
analysis.
In the gamma inspiratory analysis, the slope of the
TI-vs.-I relationship for the n
breaths was significantly different (P < 0.01) from
zero and from the n 1 breaths (Fig.
5B, Table
2). In addition, the estimated
y-intercept of the TI-vs.-I
relationship was significantly above control (P < 0.02). However, the observed TI did not show a linear
increase from the estimated intercept but, rather, an abrupt increase
around 0 to 15 I. In contrast, no differences were found
in the TE-vs.-I relationship either for the
breaths before (n 1) or during (n) the
swallow (Fig. 5A and Table 2). Swallows during early
inspiration tended to reduce VT as demonstrated by the
significantly reduced y-intercept of the
VT-vs.-I relationship for n
breaths (P < 0.02), whereas swallows in the later part
of inspiration increased VT, as indicated by the
significant positive slope of the VT-vs.-I
relationship (P < 0.01; Fig. 5C, Table 2).
In contrast, the slope of Diapeak vs. I was
not significantly different from zero for both the n and
n 1 relationships. However, a significant reduction
in the y-intercept of the Diapeak vs.
I was seen for the n breaths (Fig.
5D, Table 2). Primarily because of the lengthening of
TI above control in the n breaths (Fig.
5B), values greater than control (100%) were observed for
I.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 5.
I Analysis of respiratory data. A:
TE. B: TI. C:
VT. D: Diapeak. In total, 240 swallows qualified as defined in METHODS and Fig.
2A. Fewer events with adequate Diapeak data were
observed and presented (n = 121). Each value is
presented as a percentage of control of the individual animal control
value.
|
|
In the gamma expiratory analysis (E), swallows during
expiration had a different effect on the respiratory pattern generator. In the n 1 breaths, the slope of the
TE-vs.-E relationship, although
significantly greater than zero, was small (Fig.
6A, Table
3). However, in the n
breaths, swallowing during expiration had a significant effect on
TE, although there was considerable variation in the data.
The slope of the TE-vs.-E relationship (0.42) was significantly greater than zero and significantly different from the slope of the n 1 breaths (P < 0.01), whereas the y-intercept was significantly less
than the control (71.42%; P < 0.01; Fig. 6A, Table 3). In contrast, the slope of
TI-vs.-E relationship for n 1 breaths was significantly less than zero (0.10,
P < 0.01), with a y-intercept that was
significantly greater than control (109.8, P < 0.02).
However, the y-intercept for the
TI-vs.-E relationship for the n
breaths was significantly less than control (93.8%; P < 0.01; Fig. 6B, Table 3), whereas the slope was not significantly different from zero. Therefore, although swallows during
expiration demonstrated a moderate effect on TE in the cycle before the swallow, these results demonstrate a strong
within-phase dependency of TE on the timing of the onset of
a swallow (not shown in the categorical analysis). For the effect of
expiratory swallows on TI, the data suggest the later the
swallow occurs in expiration, the less effect it has on the previous
TI, whereas, in contrast, the following inspiratory period
is attenuated irrespective of when the swallow occurred in the previous
expiration. Similar to what was observed in I analysis
above, the increases in TI and TE produced
E values >100%.
View larger version (33K):
[in this window]
[in a new window]
|
Fig. 6.
E Analysis of respiratory data. A:
TE. B: TI. C:
VT. D: Diapeak. Number of swallows
as defined for this analysis in METHODS and Fig.
2B in the graphs for TE, TI, and
VT was 259. Number of observations for Diapeak
was 194. Each value is presented as a percentage of control of the
individual animal's control value.
|
|
The slopes of VT vs. E for both the
n 1 and n breaths were not significantly
different from zero. However, only the y-intercept for the
n breaths relationship was significantly less (88%) than control (P < 0.02; Table 3). In contrast, neither the
slopes nor the intercepts for either
Diapeak-vs.-E relationships were significant
(Fig. 6, Table 3). These results suggest that total respiratory output
was not altered, but the overall attenuation of VT was
similar to the decrease in inspiratory flow.
Effect of swallows on the respiratory phase response and output.
The occurrence of spontaneous swallows that meet the criteria for the
phase response analysis (427) demonstrated a bimodal distribution (Fig. 7A), with
~53% occurring during inspiration (average IE transition 0.47 ± 0.006; see arrow in Fig. 7A) and 47% during
expiration. The n 1 curve in
Fig. 7B illustrates the effect of swallowing on respiratory timing in the breath before the swallow. A linear regression analysis of the n 1 curve found a slope
of 1.01 ± 0.01, with a y-intercept of 0.004 ± 0.001 and a correlation coefficient (R2) of
0.97, indicating that swallowing had no effect on the timing of the
previous cycle regardless of the phase of onset in the respiratory
cycle. In contrast, the flattening of the
n1 curve between the values
of 0.1 and 0.3 reflects a phase delay in the respiratory timing
produced by increases in TI with swallows during
inspiration. The slope of the n1 curve after the IE transition (
0.4) to a value of 0.7 was
approximately 1.23. This slope was more negative than that found for
the n 1 curve over the same
period, indicating that swallows during this period may have produced a
phase advance in the respiratory rhythm. This latter finding is
supported by the tendency of TE to decrease early in the
TE-vs.-E relationship (Fig. 6A).
After values of 0.7, the n1
curve again flattens and runs parallel with its expected intersection
with zero. This flattening of the
n1 curve and its extension beyond
values of 1.0 is a result of both an increase in TE,
delay in the onset of the next cycle, and a comparable decrease in the
subsequent TI. These later observations are also seen in
the E analysis in the TE above 90% and in
the generally depressed TI-vs.-E
relationship in the breaths with swallows (Fig. 6 and Table 3). There
was a parallel shift between the n1 and
n2 and
n3 curves, except between of 0.9 and 1.2, where an upturn in
n2 and
n3 suggests a slight phase
delay. These results clearly support a type 1 or weak resetting
of the respiratory rhythm generator.
View larger version (30K):
[in this window]
[in a new window]
|
Fig. 7.
Distribution of swallows throughout normalized
respiration and the phase response analysis of the effect of the onset
of swallowing respiratory timing. A total of 427 swallows meet the
criteria for this analysis as presented in METHODS.
A: swallows as a percentage of total swallows presented
relative to their onset during a normalized respiratory cycle ().
Average transition from inspiration to expiration (IE) was ~0.47 ± 0.006 of . B: individual values for phase response
analysis of the cophase for breaths before
(n 1), during
(n1), and 2 (n2,
n3) after a swallow. Solid line
for n 1,
n1,
n2, and
n3 represents the best-fit curves
for the nonlinear regression analysis. Formulas for the calculations of
, n 1,
n1,
n2, and
n3 are presented in
METHODS and Fig. 3A. C: changes in
TI/Ttot for swallows at various and changes in breaths
before (n 1), during (n1), and 2 after
(n2 and n3). TI/Ttot values are
expressed as a fraction of control breaths (c) as defined in
METHODS. Because of the effect of swallowing to cause phase
delay in the respiratory rhythm, values >1.0 were observed for
.
|
|
The changes in TI relative to TE during a
respiratory cycle due to swallowing is shown in the graph of
TI/Ttot(n)-TI/Ttot(c), where
(c) is the control breath (Fig. 7C). In this figure,
the TI/Ttot(n1)-TI/Ttot(c)
curve increases above control during the inspiratory section of the
phase response plot. This increase is explained by progressive
increases in TI, with swallows during inspiration having no
effect on TE (Fig. 5). However, after the IE transition
(0.4), the cause of the increased TI/Ttot is due to an
overall shortening of TE relative to TI, as
shown in the E analysis (Fig. 6). Thereafter, the
eventual return to control levels of the
TI/Ttot(n1)-TI/Ttot(c) curve is due
to the relative lengthening of TE compared with a
general shortening of TI (Fig. 6). The effect of swallows
on TI/Ttot is limited to the n1 breaths, as
n2 and n3 curves are observed at control levels.
In Fig. 8, the effect of swallows on
VT is seen during expiration ( ~0.8-1.5), where
the VT(n)/VT(c) curve is lowered
during the inspiration after the swallow (n2 curve).
Subsequent compensation due to a reduction in VT was not
seen in the n3 curve. A modest decrease in
Diapeak(n)/Diapeak(c) is observed at
different sections of the n1 and n2 curves. The
n1 curve is slightly depressed near the IE transition
( 0.35), whereas the n2 curve is depressed in
the later part of expiration ( 0.8; Fig. 8B). In
contrast to the categorical data (Fig. 4) and the within-breath data
(Figs. 5 and 6), these latter data show that swallowing has a limited, temporal perturbation effect on respiratory output.
View larger version (42K):
[in this window]
[in a new window]
|
Fig. 8.
Changes in measures of respiratory output for swallows at
various in the breath before (n 1), during
(n1), and 2 after (n2 and n3).
A:
VT(n) · VT(c)1.
B:
Diapeak(n) · Diapeak(c)1.
Values are expressed as a fraction of control breaths as defined
in METHODS. Because of the effect of swallowing to cause
phase delay in the respiratory rhythm, values >1.0 were observed for
.
|
|
Figure 9 presents the differences in
n 1 ideal, n1 n 1,
n2 n1, and
n3 n2 every 0.05 interval. No
significant difference was found for n 1 ideal at any value of (P < 0.02), thus supporting the observation that the
timing of the previous breath was unaltered. In contrast, n1 n 1 differed significantly
from the predicted difference of 1 at several points along the axis
(P < 0.02). The increase in
n1 n 1 from 0.15 to 0.40 of reflects a phase delay in the n1 curve due solely to increases in TI. Thereafter, from a value of 0.40 to 0.60, the
n1 n 1 curve progressively
decreased. Although fewer swallows were observed during this phase of
ventilation, this decrease was due to the mixed and varied effects of
swallows to 1) increase TI just before the
swallow (phase delay) and 2) decrease TE (phase
advance) in early expiration. From a value of 0.65 to 0.85, the
difference between n1 and
n 1 was not significantly
different either from 1 or from the expected difference if swallowing
had no effect on respiration. Thereafter, n1 n 1 significantly increased
because of the effect of swallows late in expiration to increase
TE or phase delay the onset of the next breath
(P < 0.02). A significant decrease (P < 0.2) in the n2 n1 curve between a of 0.9 and 1.1 is a result of the effect of swallows during the later part of
expiration to also decrease the subsequent TI and produce a phase advance of the next respiratory cycle.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 9.
Differences in cophases (n 1, n1,
n2, and
n3) from the previous cophase or
a cophase (ideal) in which swallowing would not have an
effect on respiratory rhythm. All values are expressed as a fraction of
percent change. As described in METHODS, the expected
difference between n 1 ideal is 0 (dotted lines). However, the expected
difference between the measured cophases if swallowing had no effect on
respiratory timing would be 1 (dotted lines). Individual differences
were calculated and averaged over 0.05 periods of . Means and
standard deviation bars are presented for each 0.05 periods of and
compared with the expected. As mentioned in Figs. 7 and 8, values >1.0
were observed for because of the effect of swallowing to cause
phase delay in the respiratory rhythm. * Means are significantly
different from control (P < 0.01).
|
|
| |
DISCUSSION |
Our data suggest that spontaneous swallows have a phase-dependent
effect on respiratory timing and output in the awake state. Changes in
breathing before and after a swallow further suggest that the
interaction between the respiratory and swallowing pattern generators
exists beyond the apparent, immediate, all-or-nothing event of swallowing.
The interaction between swallowing and breathing.
The distribution of swallows during the phases of ventilation differs
among animals, infants, and adults. In unanesthetized (18,
20) and anesthetized (18, 19, 20) animals,
80-95% of the spontaneous swallows occur during inspiration. In
human infants, spontaneous swallows are reported to be equally
distributed during expiration and inspiration (30). In
unanesthetized and anesthetized adults, spontaneous swallows are
primarily produced during expiration (5, 23, 24, 27).
These studies suggest that swallowing may be coupled to specific phases
of ventilation. In goats, a biphasic occurrence of swallows during the
phases of ventilation was observed, with the least number of swallows occurring during the IE transition and early expiration. That over half
of the swallows occur during inspiration suggests a well-coordinated,
anatomically efficient means to minimize aspirations during a phase of
ventilation that is normally at high risk, at least in human adults. In
goats, the epiglottis overlaps the soft palate (10), which
has been shown to allow a bolus to circumvent the glottis during an
inspiratory swallow and pass into the esophagus (22). On
the basis of this anatomy, Negus (22) proposed that airflow would continue during a swallow. A similar anatomic mechanism has been proposed to occur in infants during inspiratory swallows that
not only minimizes aspirations but also allows for an uninterrupted airflow (6, 28). However, in goats, we observed
an interruption of airflow with inspiratory swallows similar to that
found in infants by Wilson et al. (30). We attribute this
interruption to a decrease in diaphragm activity (observed in this
study) and upper airway closure due to constrictor activity
(10). Therefore, we hypothesize that in goats the
overlapping of the soft palate and epiglottis allows for inspiratory
swallows with a low risk for aspirations but does not allow for
separate air and liquid channels.
Although both the presence of swallows during EI transition and
inspiration and the absence of pharyngolaryngeal problems suggest a
benign interaction between the respiratory and swallowing functions,
the lack of spontaneous swallows during the IE transition and early
expiration is curious. In awake humans, Paydarfar et al.
(27) observed the highest occurrence of spontaneous
swallows during the IE transition and the lowest during EI transition, suggesting a very different relationship between the swallowing and
respiratory pattern generators in humans, dogs, monkeys, and goats.
Several studies (19, 27) have shown that stimulation of
peripheral afferents during IE transition has a disruptive effect on
respiratory timing, which may be the neurophysiological basis for the
reduced occurrence of spontaneous swallows during transition in these animals.
Less well investigated has been the effect of spontaneous swallows on
respiratory timing and total output. Studies in decerebrate cats showed
that swallows increased the duration of the current and subsequent
respiratory cycle (21). In awake rabbits, swallows during
expiration significantly increased TE and the preceding TI, whereas during inspiration swallows significantly
increased TI and the subsequent TE
(20). These swallows did not significantly affect
TI or TE in the breaths that followed. In
unanesthetized humans, both spontaneous and water-induced swallows
during expiration increased TE and Ttot, as well as
VT, immediately after the swallow (23).
Spontaneous swallows during inspiration reduced TI,
VT, and the following TE. In the subsequent
breath, VT was increased. A similar response was found in
stimulated swallows that occurred during inspiration, except that Ttot
was decreased (23). In goats, swallows during expiration
also had a similar effect on TE and Ttot. However, in
contrast to data in humans (23), the ensuing
TI, VT, and Diapeak were
significantly reduced, suggesting that an inhibitory effect of
swallowing persisted in the following inspiratory timing and output.
The increase in TI with swallows during inspiration in
goats was similar to that found in rabbits (20) but is in
contrast to the findings in humans (23). Similarly, VT was slightly increased, which is opposite of that found
in humans (23), and the findings of Diapeak in
goats study further support the idea that total respiratory output is
unaltered whereas respiratory timing increases with swallows during inspiration.
Although the information in the literature on the effect of swallows on
the present and subsequent breaths reveals general trends in the
interaction between respiratory and swallowing centers, it does not
investigate or eliminate a within-breath and/or multibreath interaction. Most studies (20, 23) utilized the breath(s) before the swallows as controls. This latter point assumes that stimulated and spontaneous swallows only have an effect on breathing during and after the swallows. Although our results support these findings, they also show that both timing and the output to the diaphragm are altered before a swallow. The neural substrate for this
interaction is unknown. However, it is presumed that a build-up of
secretions in the upper airway activates receptors that travel to the
nucleus tractus solitarius via the superior laryngeal nerve (15). Jean (12-14) reported that some
neurons in the nucleus tractus solitarius exhibit a "preswallowing
activity" that may act as trigger neurons for swallowing whose
activity is increased with stimulation of the superior laryngeal nerve.
Dick et al. (7) also observed that stimulation of the
superior laryngeal nerve below the threshold for eliciting a swallow
resulted in a prolongation of TE. Together, these results
suggest a peripheral feedback mechanism to the medullary neurons that
may elicit changes before a swallow and function to prepare the animal
for a swallow.
The within-breath analysis of the effects of swallowing on the pattern
of breathing provides an insight into the moment-to-moment relationship
between the swallowing and respiratory centers. In infants, there is a
linear relationship between the duration of inspiration and the
TI at the onset of airway closure due to a swallow
(30). The earlier the swallow occurred (within the first half of a breath), the shorter the TI. Thereafter, swallows
produced a longer period of inspiration. Swallows during expiration
also had the greatest effect in increasing Ttot. This relationship produced a negative correlation between Ttot and the occurrence of
swallow during expiration. This finding is in contrast to the observation in anesthetized and unanesthetized adult humans, in whom
spontaneous swallows occurred primarily during expiration, producing a
positive relationship between TE and the occurrence of
swallows during expiration, thereby increasing Ttot (23, 24). However, in anesthetized adult humans, swallows during inspiration abruptly interrupted inspiration and were followed by a
short expiratory period, thereby producing a significant positive
correlation between TI and the onset time of the swallows (24).
Similarly, there is a positive correlation between VT and
the onset of swallows during inspiration in anesthetized and
unanesthetized adults (23, 24). In contrast, an abrupt
cessation of inspiration followed by expiration was not observed in
goats. Instead, swallows progressively lengthened TI the
later they occurred in inspiration (Table 2, Fig. 5B). In
addition, although swallows at the very beginning of inspiration were
related to a reduced VT, thereafter a positive correlation
(above control) was observed for VT and the onset of the
swallow during inspiration. In contrast, no change in
Diapeak was observed. These observations suggest that not
only did swallowing insert a pause into the inspiratory period while not widening the swallow complex and increasing the interruption in
flow, it also had a mild progressive inhibition on ventilation (decreasing) while not altering total activity of the diaphragm.
During expiration in goats, the relationships between the onset of a
swallow and the values of TE, TI, and
VT are considerably different from previously reported
data. In contrast to those for humans, our data show that the earlier
the onset of a swallow during expiration, the shorter the
TE observed, until ~75% E, at which point
swallowing increased TE above control. This temporal effect
of swallowing during expiration is also seen in the negative slope of
n 1 TI data, whereas a generalized
overall inhibitory effect of swallowing was found in the relationship
of the TI vs. E for the n
breaths. Also consistent with this latter finding was the significant
effect of expiratory swallows on the subsequent VT, whereas
Diapeak was not altered. These findings suggest that output
to the diaphragm is not reduced during inspiration after a swallow;
only the timing and the subsequent VT are reduced.
Phase response analysis in the interaction of breathing and
swallowing.
A major advantage of the phase response analysis is the capability to
characterize the relationship between pattern generators without
knowing a priori the specific neuroanatomic basis of the interaction
(29, 31). The phase response of stimulated and spontaneous
swallows has only been investigated in humans by Paydarfar et al.
(27), who found that the later a swallow occurred during inspiration, the later the onset of the subsequent breath. The peak of
this response was near IE transition, after which no effect on the
onset of the next cycle was seen until the swallow occurred at the very
end of the respiratory cycle. This nonlinear effect of swallowing on
breathing is very similar to the results of Nishino and colleagues
(23, 24). In addition, Paydarfar et al. (27) demonstrated that swallows produced a "true" resetting of the respiratory rhythm. In other words, instead of an immediate, short-term shift in the bulbospinal output produced by swallowing, a parallel shift in the subsequent breaths occurred. In goats, resetting of the
respiratory rhythm generator also occurs after swallows at different
onsets within the respiratory cycle, with a parallel shift in the
starting of subsequent breaths. The continuity of the calculated
cophases () further supports the concept of a continual oscillation
of the respiratory pattern generator and true phase resetting.
However, the results from the phase response analysis in our study
differ from those found in humans by Paydarfar et al.
(27). The basis for this difference lies in the
within-breath effects of swallows on breathing. The type of resetting
observed by Paydarfar et al. was a type 0, in which the net change of
the cophase was zero over a cycle of . Except for the phase delay
and advance observed in the n1
and n2 curves, the net slope of
the n1,
n2, and
n3 curves in our study
approximated a negative slope of 1, demonstrating a type 1 (weak)
resetting effect of swallowing on breathing in goats (29,
31). This is also in contrast to a type 0 resetting observed in
the breathing of anesthetized cats by different levels of superior
laryngeal nerve stimulation (26).
EI and IE transitions and swallows.
TOP
ABSTRACT
INTRODUCTION
