INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

MECHANICAL VENTILATION produces apnea in humans (10) and animals (11). In humans, there is redundancy of the sensory information required to initiate apnea. Hypocapnea and pulmonary vagal, upper airway, and chest wall mechanoreceptor stimulation can each induce apnea independently (13, 22-24). Although we know much about the peripheral mechanisms involved in initiating the apnea in the unanesthetized state, no systematic studies exist of the central neural mechanisms that must ultimately cause it.

Some respiratory muscles and neurons are not inactive but instead have a residual or tonic discharge during ventilator-induced apnea. Residual activity has been observed during ventilator-induced apnea in the triangularis sterni (expiratory) muscles in dogs (11), in internal intercostal nerves in cats (21), and in expiratory fibers of the recurrent laryngeal nerve in cats (5). Interestingly, it is sometimes reported that the maximal level of this tonic activity is greater than the maximal level during eupnea (5). Tonic activity in inspiratory muscles during apnea has been reported less often than tonic activity in expiratory muscles (Ref. 2, 21; work of Wyss cited in Ref. 7) and may depend on the species studied and on the preparation (1, 2).

Some medullary respiratory neurons are also tonically active during ventilator-induced apnea. Early studies (3, 7, 16) showed examples of both inspiratory and expiratory neurons that were tonically active during this apnea.

Various theories address the source and significance of residual respiratory muscle and neuronal activity during apnea. In early theories, Batsel (3) and Cohen (7) proposed that residual activity of respiratory neurons was the result of inactivity of cells that reciprocally inhibited them. Another theory proposed that tonic activity during hypocapnic apnea was caused by a CO₂-dependent drive (2). Bulbospinal expiratory neurons progressed from inactivity to tonic activity, and then rhythmic activity as CO₂ tensions increased (2). However, other respiratory neurons were tonically active even at very low CO₂ tensions, leading to the theory that residual activity was caused also by non-CO₂-dependent drives, including those related to different states (2, 11, 25). Nesland and Plum (16) suggested that residual activity reflected a global excitatory state out of which the rhythm to breathe is generated.

In this study, single medullary respiratory neurons were recorded in intact unanesthetized cats during spontaneous breathing and during ventilator-induced apnea. We wanted to determine whether tonic or residual activity could be the cause of apnea when an animal is hyperventilated. For example, tonic expiratory activity might inhibit inspiratory neurons and lock the oscillator in expiration.

Residual activity during apnea was of interest for another reason. In the intact unanesthetized animal, respiratory neurons vary in "respiratoriness" (19). Some have discharge patterns that correlate highly with the timing of the respiratory cycle; others have patterns that are more weakly related to the cycle. These differences have been quantified (19), and it has been proposed that the poor correlation with the respiratory cycle shown by some respiratory neurons is caused by tonic inputs to them (20). These inputs may be related to activity of the nervous system that depends on state of consciousness (20) or on behavioral control (18). We hypothesized that, if the residual activity of respiratory neurons during apnea was the result of tonic or nonrhythmic inputs to the cells, then there should be a direct relationship between the amount of residual activity during apnea and the amount of nonrhythmic activity of the cell during eupnea.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Electrodes for recording electroencephalographic (EEG) and electromyographic (EMG) activity were implanted in four adult cats. In addition, tracheal fistulas were created, and a headcap was attached to the skull.

Surgical Procedures

EMG electrodes (Teflon-coated multistranded stainless steel wires; Cooner AS 632) were implanted in the nuchal muscles and in the diaphragm. The animal was placed in a supine position, and an incision was made caudal to the costal margin from the xiphoid process to the midaxillary line. The head and upper body of the cat were elevated to displace the abdominal contents caudally, and the costal margin was elevated to provide access to the diaphragm. Four wires were placed within costal and semitendinous regions of the right diaphragm. The wires were run under the skin to the occipital bone, where they were cemented in place and brought to a connector (Cinch 19 pin). EEG and EMG electrode wires and a prefabricated headcap with a connector and with standoffs for head restraint were fixed to the skull with dental cement. The animals were allowed to recover for at least 1 mo before experimentation. All procedures, preoperative, postoperative, and experimental, were approved by the Institutional Animal Care and Use Committee.

Experimental Procedures

After adaptation and in a second operation under general anesthesia (as above), a small craniotomy (~5 mm diameter) was made in the occipital bone. The craniotomy allowed passage of microelectrodes through the cerebellum into the brain stem.

The animals were sleep deprived mildly before experimental sessions by housing them in a cold (0°C) environment overnight. For recording sessions the trachea was intubated with a shortened endotracheal tube (18-Fr). Needles were inserted into the lumen of the tube to measure intratracheal pressures and to obtain samples of CO₂. The endotracheal tube was connected to a pneumotachograph (Validyne). EEG and EMG activity, instantaneous airflow rates (Validyne CD15 Carver demodulator), intratracheal pressures (Grass PT5A volumetric pressure transducer), and end-tidal CO₂ levels (Beckman LB2 analyzer) were recorded on an analog tape recorder (Hewlett-Packard) and on a chart recorder (Astro-Med MT9500). The EEG was band-pass filtered (1-35 cycles/s) and amplified (Grass wideband alternating current preamplifiers). EMG signals from the diaphragm and nuchal muscles were led to high-impedance probes (Grass HP511) and to amplifiers (Grass P511) set to pass frequencies from 300 to 10,000 cycles/s. A solenoid-operated two-position valve was attached to the distal end of the pneumotachograph. With the valve set in one position, the animal breathed room air. In the other position, the animal was connected to a ventilator that delivered 40-50 ml tidal volumes at rates of 30-40 per minute. No attempt was made to match exactly eupneic tidal volumes and frequencies or to maintain normal CO₂ levels. The levels of ventilation used caused hypocapnia. End-tidal CO₂ levels during ventilation typically dropped from 5 to 3.5-2.5%.

Tungsten microelectrodes (impedances 1-10 M) were used to record single medullary respiratory neurons. The microelectrodes were mounted to a hydraulic microdrive and driven via the craniotomy through the cerebellum and into the medulla. Signals were led to a high-impedance probe (Grass HIP511) and to a preamplifier (Grass P511) and were recorded on the Astro-Med and analog tape recorders.

The recording sessions lasted ~4 h.

Data Analysis

Mean and maximal discharge rates were determined also during eupnea and during ventilator-induced apnea. For these determinations, cycle-triggered histograms were constructed of activity during eupnea, during the early part of mechanical ventilation (first 18-23 cycles), and during a later part of mechanical ventilation (cycles 19-47). The cycle-triggered histograms were analyzed to obtain the mean and maximal level of activity throughout the cycle as well as the SD of activity across the 20 bins of the histogram. The mean value was calculated as the mean of bins of both the active and inactive phases of the neuron (Figs. 1 and 2). The physiological significance of this mean is unclear, but it is a well-defined value that does not require an arbitrary definition of the active and inactive phases of a cell. This definition becomes increasingly problematic with patterns of activity that augment and/or decrement throughout a major portion of the respiratory cycle. Although problematic in some contexts, the mean value of the activity across the bins of the cycle-triggered histogram characterized well the pattern of activity of many of the neurons during apnea (Figs. 1 and 2).

View larger version (19K): [in this window] [in a new window]   Fig. 1.   Cycle-triggered histograms of a high-2-value (0.81) augmenting inspiratory cell during eupnea and during ventilator-induced apnea; 2-value denotes signal strength and consistency of respiratory component of activity of a cell. Solid lines, discharge rates throughout respiratory or ventilator cycle during eupnea or apnea. Lines connect mean values of activity during each 20-ile of respiratory or ventilator cycle that were calculated from 18 breaths and 18 ventilator cycles, respectively. Dashed lines, mean of mean rates over complete cycle during eupnea and apnea. During apnea, cell was not modulated by ventilator. Amount of residual activity is expressed in this study as 1) ratio of mean activity in apnea to mean activity in eupnea and 2) ratio of maximum activity in apnea to maximum activity in eupnea.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Cycle-triggered histograms of a low-2-value inspiratory cell during eupnea and during ventilator-induced apnea. See Fig. 1 for explanation. This activity of cell had an 2 value of 0.49 in eupnea and was not modulated by ventilator during ventilator-induced apnea. Note that ratio of mean activity in apnea to mean activity in eupnea will be larger for this cell than for cell in Fig. 1 because of small difference in mean value of activity of cell in eupnea and mean value of activity of cell in apnea. Therefore, this cell has more residual activity during apnea (a higher ratio of mean apneic to mean eupneic activity) than does cell in Fig. 1. Ratio of maximal activity in apnea to maximal activity in eupnea is also greater for this cell than for cell in Fig. 1, indicating also a greater amount of residual activity. However, ratio of maximum in apnea to maximum in eupnea is <1.0 for both this cell and cell in Fig. 1.

We calculated also the SD of activity from bin to bin in the cycle-triggered histogram and determined the coefficient of variation (SD/mean). The maximal activity in the cycle-triggered histogram was also determined. This was simply the largest bin value in the cycle-triggered histogram. To normalize the amount of residual activity during mechanical ventilation across cells, means, maxima, and coefficients of variation during mechanical ventilation were expressed as fractions of means, maxima, and coefficients of variation during eupnea. For example, if the mean level of activity during eupnea was 50 action potentials/s and the mean level of activity during mechanical ventilation was 25 action potentials/s, then the residual activity based on these means was expressed as the ratio 25/50 or 0.5. Differences in ratios of means and maxima and the coefficients of variation between cell types (inspiratory and expiratory, low and high ² values) were tested by using a two-way analysis of variance with unequal cell frequencies. Ratios based on the coefficients of variation proved misleading because coefficients of variation were often equal when activity of the cell was great and highly modulated and also when activity was minimal, with only occasional action potentials that were unrelated to the respiratory cycle. Therefore, only data from the ratios based on means and maxima are given in RESULTS.

In two figures (see Figs. 8 and 10) in this study, the activity of cells determined from the mean of interspike intervals <300 ms (an arbitrary value chosen to exclude intervals during the inactive phase of the cells) was plotted breath by breath before and during mechanical ventilation to illustrate the effects of this ventilation. These means were used for illustrative purposes only and were not used to calculate the effects of mechanical ventilation.

Histology

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Cell Types and Locations

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Locations of scarring produced by microelectrode penetrations into medulla in 4 cats (DX, DL, HR, LN) studied. RB, restiform body; 5ST, nucleus of spinal tract of the trigeminal; RFN, retrofacial nucleus; FTG, gigantocellular tegmental field; S, tractus solitarii; 12, hypoglossal nucleus; AMB, nucleus ambiguus; DMV, dorsal motor nucleus of vagus; P, pyramidal tract.

Dependence of Residual Activity During Ventilator-Induced Apnea on Cell Type and on ² Value

View larger version (34K): [in this window] [in a new window]   Fig. 4.   Response of an inspiratory-throughout (I) cell to mechanical ventilation. Note rapid inactivation of cell and brief burst of activity on 5th cycle of ventilator. Ptr, intratracheal pressure, with positive pressures indicated by downward deflections

View larger version (49K): [in this window] [in a new window]   Fig. 5.   Response of an augmenting inspiratory cell to mechanical ventilation. A: induction of apnea in non-rapid-eye-movement (NREM) sleep. B: tonic activity of cell in NREM sleep. C: tonic activity of cell during wakefulness. Note in wakefulness both a higher rate of tonic discharge and organized bursts associated with swallows and attempted breaths. Small action potentials were from an expiratory neuron that was phasically modulated by ventilation. EEG, electroencephalogram.

View larger version (55K): [in this window] [in a new window]   Fig. 6.   Response of an augmenting inspiratory cell to mechanical ventilation. In NREM sleep (A) cell was inactivated rapidly and there was low-rate residual activity. In rapid-eye-movement (REM) sleep (B), residual activity was greater. This greater residual activity was tonic, modulated, and phasic activity. Phasic modulation for this augmenting inspiratory cell, as in most others of this type, resulted in greater activity during deflation-inflation transition. EMG, electromyogram.

View larger version (54K): [in this window] [in a new window]   Fig. 7.   Response of an augmenting expiratory (Eaug) cell to mechanical ventilation. Note sustained activity of cell at onset of ventilation, with a gradual decline in activity with ventilation.

View this table: [in this window] [in a new window]   Table 1.   Mean residual activity as a fraction of mean eupneic activity of high- and low-2-value cells during early and late mechanical ventilation

View this table: [in this window] [in a new window]   Table 2.   Maximal residual activity as a fraction of maximal eupneic activity of high- and low-2-value cells during early and late mechanical ventilation

Inspiratory cells showed less residual activity than expiratory cells (Tables 3 and 4). This difference was observed in both high- and low-²-value categories. It was observed for both mean and maximal value ratios.

In addition to inspiratory and expiratory cells, 21 phase-spanning cells were studied. Only one of these cells had an ² value >0.5. The average ² value of these cells during eupnea was 0.20. Data from them are included in Tables 1 and 2. Like other low-²-value cells, they showed large amounts of residual activity during ventilation: mean and maximal ratios during early ventilation were 1.03 and 0.94, respectively. During late ventilation, mean and maximal ratios were 0.918 and 0.967, respectively. Ten of these twenty-one cells showed phasic modulation by the ventilator. The mean ² value of the phasic modulation of these cells was 0.29.

The difference in residual activity between low- and high-²-value cells was not caused by differences in phasic modulation by the ventilator. Forty of seventy-eight (51%) high-²-value cells were modulated by the ventilator, whereas 35 of 59 (59%) low-²-value cells were modulated by the ventilator. This difference was not significant (² analysis). Similarly, low- and high-²-value cells that were modulated to the ventilator did not differ in the strength of that modulation. The mean ² values of the modulated activities of low- and high-²-value cells were 0.21 and 0.24, respectively. Modulation of high-²-value cells by the ventilator is discussed in greater detail below.

Although low-²-value cells, and in particular expiratory cells in this category, had more residual activity than high-²-value cells, none of these cells was intensely activated during mechanical ventilation.

Responses of High-²-Value Inspiratory and Expiratory Cell Subtypes to Ventilation

View this table: [in this window] [in a new window]   Table 5.   Responses of high-2-value cell types to mechanical ventilation

Inspiratory-throughout and decrementing inspiratory cells were generally inactivated within one or two cycles of the ventilator and showed little or no residual activity (Fig. 4). In contrast, 86% of the augmenting inspiratory cells had considerable residual activity during ventilator-induced apnea (Figs. 5 and 6).

Twelve of the thirteen augmenting expiratory cells remained active during the early stages of ventilation and then showed a decline in activity (Figs. 7 and 8). Three of five decrementing expiratory cells showed sustained tonic activity during ventilation that was comparable to the mean discharge rates during eupnea. Although the mean rates of decrementing expiratory cells were comparable during mechanical ventilation and spontaneous breathing, the maximal rates on the ventilator were less than (n = 2) or equivalent to (n = 1) maximal rates during eupnea. All expiratory-throughout cells had mean discharge rates that were as high (n = 3) or higher (n = 3) on the ventilator than during eupnea. Peak discharge rates on the ventilator were less than peak discharge rates during eupnea (an average of 72% of eupneic rates) for four of the six cells. For two cells they were as high as or higher than the rates during eupnea. Figures 9 and 10 illustrate one of these cells, the activations of which were immediate with the onset of ventilation and sustained. As noted below, this intense activation occurred only in wakefulness and NREM sleep. In REM sleep, the cell was less active spontaneously, and ventilation in that state produced a weak and intermittent activation of the cell (Figs. 9 and 10).

View larger version (24K): [in this window] [in a new window]   Fig. 8.   Quantification of mean discharge rate of same cell shown in Fig. 7. A brief activation and then a gradual decline in activity are shown. EEG power plotted below discharge rate curve was relatively unchanged throughout decline in neuronal activity, which indicates decline was not associated with a change in state. Mean discharge rates were calculated from mean of interspike intervals <300 ms. cps, Cycles per second.

View larger version (40K): [in this window] [in a new window]   Fig. 9.   Response of an expiratory (E) cell to mechanical ventilation in NREM and REM sleep. In NREM sleep, ventilation caused activation of cell that was sustained (A). In REM sleep, eupneic activity of cell decreased and only occasional action potentials were observed (B). Ventilation in REM sleep activated cell, but this activation was weak relative to activation in NREM sleep and it was intermittent (C).

View larger version (17K): [in this window] [in a new window]   Fig. 10.   Quantification of activity of same cell shown in Fig. 9. In NREM sleep, ventilation (Apnea in A) caused activation of the cell to discharge at a mean rate of ~40 cps. In REM sleep (B), cell was inactivated and ventilation produced activation of cell to levels less than one-half of those during ventilation in NREM sleep. Mean discharge rates were calculated from mean of interspike intervals <300 ms.

Residual Activity: State Effects

The percentage of cells having residual activity was higher for expiratory than for inspiratory cells. Of augmenting expiratory cells observed during apnea in wakefulness (n = 8), NREM sleep (n = 7), and REM sleep (n = 2), 100%, 57 and 100%, respectively, had residual activity. This was similar to the pattern shown by inspiratory cells. However, of decrementing expiratory cells observed during apnea in wakefulness (n = 4), NREM sleep (n = 4), and REM (n = 3) sleep, 100, 100, and 33%, respectively, had residual activity. This differed from the pattern shown by inspiratory and augmenting expiratory cells, not only in the persistence of residual activity during NREM sleep but also in the reduction in that activity in REM sleep. Only one expiratory-throughout cell was observed in REM sleep, but it too showed less residual activity in that state (Figs. 9 and 10). Expiratory-throughout cells observed during apnea in wakefulness (n = 5) and NREM sleep (n = 4) all showed some residual activity.

Phasic Modulation of High-²-Value Cells by the Ventilator

View larger version (11K): [in this window] [in a new window]   Fig. 11.   Cycle-triggered histograms of an augmenting expiratory (A and B) and an inspiratory-throughout cell (C and D) during eupnea (A and C) and during ventilation (B and D). Both cell types were modulated to discharge at deflation (Defl)-inflation (Infl) transition; that is, they were nonreciprocally related.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

Central respiratory neurons were eventually neither excited (with the exception of the expiratory cells discussed below) nor inhibited during ventilator-induced apnea. They displayed varying amounts of residual activity that depended on the ² values of their activities and on the state of consciousness. The activity, whether tonic and/or modulated by the ventilator, was often nonreciprocal in neurons that discharge reciprocally during eupnea. Expiratory neurons were, in general, more active than inspiratory neurons during apnea, but their activity levels were intermediate between their most active and their most inactive levels during eupnea. Thus mechanical ventilation creates a state in which respiratory neurons are clearly excitable and are often active but in which the patterning among them is lost.

Significance of Residual Activity

The ² value. Low-²-value cells, whether inspiratory or expiratory, had more residual activity during apnea than high-²-value cells. With the assumption that mechanical ventilation eliminates the respiratory component in the activity of a cell and reveals components other than those related to rhythm generation, this finding supports the interpretation that the ² statistic (19) quantifies the amount of nonrespiratory input to a cell (17, 20). The greater amount of residual activity in low-²-value cells was not caused by a greater percentage of modulated cells in that category because there was no difference in the percentages of high- and low-²-value cells that were modulated by the pump or in the strength of that phasic modulation.

Cell type. Expiratory cells had more residual activity than inspiratory cells. This may indicate that expiratory cells receive more nonrespiratory inputs than inspiratory cells, but this is not reflected in the distribution of ² values of inspiratory and expiratory cells. Alternatively, apnea may create a state in which expiratory cells are relatively more depolarized than inspiratory cells. We do not know by what mechanism or for what purpose this might be the case.

State of consciousness. There was the least amount of residual activity in NREM sleep. This is consistent with the idea that tonic or nonrespiratory inputs are least in that state (11, 20). In wakefulness, both tonic activity, activity related to behaviors such as sniffing and swallowing, and behavioral breathing often occurred. In REM sleep, residual activity also occurred. The patterns of this REM-related activity were not recognizable. The activity appeared as erratic bursts perhaps related to the fractionated breathing that occurs in REM sleep in eupnea.

Mechanisms of Apnea

As another possibility, apnea may result from mechanoreceptor reflexes. There is elegant work by Hayashi et al. (12), Feldman and Cohen (9), and Manabe and Ezure (14) showing that decrementing expiratory cells are activated by pulmonary stretch receptors and that they, in turn, inhibit inspiratory cells. It seemed likely that mechanical ventilation would produce excitation of decrementing expiratory cells with each inflation and that this would inhibit inspiratory cells to cause apnea. Our results do not support this idea. Inspiratory cells are not inactivated during the apnea, and decrementing expiratory cells that were modulated in phase with the ventilator discharged at either the transition from inflation to deflation or at that from deflation to inflation rather than throughout inflation, as might be expected from the work of Hayahsi and others (12). The strength of that modulation had an ² value of 0.26 averaged across the three decrementing expiratory cells studied. This is a relatively weak effect. It is comparable to the strength of modulation of other respiratory cells-more than one-half of which had peak discharges during the deflation-inflation transition of the ventilator.

Another mechanism that must be implicated in the apnea is unloading of the central chemoreceptors. Hypocapnea develops slowly with ventilation (4) because of the buffering capacity of the blood, and there is a delay before central chemoreceptors sense a change in CO₂ levels. Changes in CO₂ are sensed by peripheral chemoreceptors within 3-5 s and by central chemoreceptors within ~25 s in awake lambs (6). Apnea sometimes occurred too rapidly in our study to be attributable to chemoreceptor unloading, but the latter must eventually play a role. It is not known how low CO₂ concentrations affect respiratory neurons to cause apnea. Some cells show a progressive decline in neuronal activity during ventilation that parallels decreasing CO₂ levels. The best examples of this were obtained from augmenting expiratory cells.

Eventually, respiratory cells, with the exception of the expiratory cells that may be motoneurons, appear to be in a disinhibited and disfacilitated state during apnea. The mechanisms causing this state are not known.