Breathing During Sleep in the Postnatal Period of Rats: The Contribution of Active Expiration

Abstract

Breathing is most vulnerable to apneas and other disturbances during sleep in both humans and rodents, especially in the newborn period. We recently demonstrated in adult rats that, in contrast to the atonia typical of skeletal muscles during rapid eye movement sleep, the normally passive expiratory muscles become active, and their activity is associated with stabilization of breathing and increased ventilation. In this study, we investigated the relationship between respiration and expiratory muscle recruitment across sleep states during the first 2 weeks of rat postnatal development. We instrumented rats with electromyography electrodes in neck and abdominal muscles while sleep states (active and quiet sleep) were classified based on nuchal muscle tone and overt behavior inside a whole-body plethysmograph. Our results indicate that breathing was most irregular in active sleep (AS) and that rats displayed frequent recruitment of expiratory muscle activity, which occurred in both active and quiet sleep states. While the occurrence of active expiration in quiet sleep did not affect ventilation and its frequency decreased with age, the recruitment of expiratory muscles during AS was present across development and it was associated with a reduction in respiratory variability across development and an increase in ventilation at most age groups considered. We conclude that the occurrence of active expiration is a common feature of respiration in the postnatal period, and it significantly contributes to ventilation, in particular in AS.

INTRODUCTION

Breathing is an essential physiological process under neural control that must be functional at birth and be able to adjust to a host of physiological challenges in the newborn period, such as postnatal lung and neuronal development, changes in pattern and dynamics of sleep/wake cycles and circadian rhythms, and the progressive development of the hypoxic and the hypercapnic ventilatory responses.

In the first days and weeks of life, both human and rodent infants spend a large amount of time asleep. Active sleep (AS), the precursor of adult rapid eye movement (REM) sleep,^1,2 dominates in the newborn, but it decreases progressively through development as more time is spent in quiet sleep (QS), the precursor of adult non-REM sleep, and wakefulness (W).^1–5

Breathing in premature, and occasionally full-term human, infants is characterized by frequent respiratory disturbances during sleep, in particular during active or REM sleep. Sleep-disordered breathing in infants is characterized by an irregular breathing pattern that includes frequent depressions (hypopneas) or interruptions (apneas) of breathing that are obstructive, central, or mixed, in origin and are often associated with blood oxygen desaturation.⁶ Sleep-disordered breathing not only poses immediate risks for survival, but it is also associated with serious and life-long health consequences, such as poor weight gain, abnormal motor skills, adverse behavioral changes, decreased cognitive function, maladaptive cardiovascular changes, and sympathetic activation.^7–14

In rodents, several studies have investigated developmental changes in respiratory and metabolic rate, respiratory stability, hypoxia-induced arousal from sleep, and development of the ventilatory responses to hypoxia and hypercapnia within the first 3 weeks of postnatal life.^15–27 These developmental analyses of respiratory behavior have established that ventilation becomes more stable with development, and the function of both peripheral and central chemoreceptors and the neuromodulatory systems that control breathing mature over the first postnatal weeks. Semiquantitative immunohistochemical analysis of neurotransmitransmitters and receptors expression have also identified a critical respiratory period at postnatal day (P) 13 where excitability appears to fall briefly.²⁸ Around the same time in development, ventilatory and metabolic responses to hypoxia appear to be the weakest.^15–17

Importantly, the majority of the developmental studies involving physiological measurements were performed in awake or in an unknown brain state animals or alternatively, with limited (mostly observational) evaluation of sleep/wake state. This is understandable given the challenge of recording electroencephalogram (EEG) activity from newborn pups and that EEG alone is not sufficient to define sleep state due to the lack of clear sleep state-defining cortical EEG features prior to P11 in rats.¹ However, given the impact of sleep state on respiratory control and the evidence that virtually all clinical disorders of breathing that involve the CNS manifest primarily in sleep, brain state information in respiratory studies is critical.

Observational methods for defining sleep state in neonatal rats are a significant improvement.^26,27 However, the systematic analysis of nuchal electromyography (EMG) activity in combination with overt behavioral markers has increased our capacity to detect state-specific behaviors/control mechanisms, and it has greatly advanced our understanding of the postnatal development of sleep dynamics in rodents.^1,3,29–31 Further, the ability to record EMG activity directly from respiratory muscles (current study) gives us the opportunity to determine state-dependent activity and modulation of muscles that control ventilation.

The first objective of this study was therefore to apply these tools to characterize how breathing pattern changes in rats across sleep/wake cycles in the first two postnatal weeks. The second objective of this study was based on our recent demonstration in freely moving adult rats that recruitment of expiratory muscles both reduced respiratory variability and increased minute ventilation (VE) in REM sleep.³² We therefore set out to determine if and when during postnatal development expiratory abdominal (ABD) muscles were recruited as well as the impact of this recruitment on ventilation across sleep/wake cycles.

Our developmental analysis of sleep states indicates a progressive consolidation of sleep toward the adult pattern by the end of the second postnatal week; a reduction in the time spent in AS in the second postnatal week was associated with extension of wake and QS time. Similar to adult rats and human infants, postnatal rats displayed increased respiratory variability in AS compared to QS at all age groups examined. Expiratory ABD activity was also present in all age groups, and its occurrence and intensity progressively decreased with age in both QS and AS. Although ABD recruitment was not associated with changes in respiratory frequency or variability in QS, onset of expiratory ABD activity was associated with a reduction in respiratory variability during AS and increased ventilation in P0–1 and P8.

MATERIALS AND METHODS

Experimental procedures were approved by the Health Science Animal Policy and Welfare Committee of the University of Alberta according to the guidelines established by the Canadian Council on Animal Care.

Animals

A total of 41 Sprague-Dawley perinatal rats of postnatal (P) day P0 (n = 6; within 10 hours of birth), P1 (n = 6), P3–4 (n = 10), P7–8 (n = 10), P14–15 (n = 9) were used in this study; for simplicity, the latter groups are indicated as P4, P8, and P15. An additional 12 rats were used to compare changes in respiratory rate and body temperature in noninstrumented conditions (n = 3/age at P1, P4, P8, and P15). Pregnant rats were received from Charles River breeding facility (Montreal, Canada) at gestational day (E) 13–16 and housed at the University of Alberta. Dams provided care for their offspring in standard size cages, in which food and water were available ad libitum under a regular 12-hour light-dark schedule with lights on at 08:00 am.

Surgery

Postnatal rats weighing between 5 (P0) and 55 (P15) grams were removed from the litter and anesthetized in a chamber saturated with isoflurane and maintained under 2% isoflurane anesthesia for the duration of the acute surgical procedure. Rats were maintained warm during instrumentation procedure with an electrical heating pad set at the lowest setting. Five-millimeter to 1-cm long skin incisions were made in the region of the neck, thorax, and abdomen. Paired EMG electrodes comprised of hook needles (Fine Science Tools, Canada), and multistranded, Teflon-coated, stainless steel wires (AM-Systems, USA) were implanted in the nuchal, intercostal (INT; in ≥P4 rats), and ABD muscles to measure postural, inspiratory, and expiratory muscle activity, respectively. The skin was sutured with a 6/0 nonabsorbable nylon monofilament suture (Havel’s, USA). The long-term acting local anesthetic bupivacaine (Marcaine) was applied subcutaneously (50–100 µL at 5 mg/mL) at the incision sites to eliminate discomfort following instrumentation procedures. The surgical procedure on average lasted less than 30 minutes.

Cortical delta (0.5–4 Hz) activity, indicative of slow wave sleep emerges in rats at P11³³, thus, only P15 rats were instrumented with paired EEG electrodes made of stainless steel screws (#00-96 × 1/16, diameter 1.19 mm; Plastics One, USA) connected with Teflon-coated stainless-steel wires (AM-Systems). Screws were inserted in the skull, above the left frontal and parietal cortices as well as on top of the cerebellum to serve as a ground electrode.³³ Screws and wires were then fixed with dental acrylic to the rats’ skull.

Recording of Respiratory Parameters and EMG/EEG Activity

Following EMG/EEG instrumentation, rats were placed in prone position inside a size-adjusted whole body plethysmographs (P0-1 = 50 mL; P4-8 = 500 mL; P15 = 260 mL) directly warmed by a servo-controlled heating pad positioned beneath the plethysmograph and set at temperature of 37 ± 1°C (Harvard Apparatus, Canada). Ambient room temperature was set at 25°C. Body temperature was measured using a noncontact infrared thermometer (BeBetter, model #11–927, Canada) before and after plethysmograph recordings in the interscapular region. Surface body temperatures prior to recording procedures were: 35.5 ± 0.4°C, 34.8 ± 0.3°C, 35.9 ± 0.4°C, 35.6 ± 0.4°C, and 35.1 ± 0.3°C in P0, P1, P4, P8, and P15 rats, respectively. At the end of the experiment, recorded surface body temperatures were 35.5 ± 0.4°C, 35.0 ± 0.2°C, 36.5 ± 0.1°C, 35.9 ± 0.6°C, and 35.4 ± 0.3°C in P0, P1, P4, P8, and P15 rats, respectively. No significant difference in temperature between the two time points was observed for any age group investigated (p > .05), suggesting a limited effect on body temperature during the recording session.

Rat pups at all ages were always in the prone position for the duration of the recording. Electrodes were channeled outside the recording chamber and connected to differential amplifiers (AM-Systems). EMG/ EEG signals were recorded at 1 kHz using the Powerlab 16/30 data acquisition system (AD Instruments, USA). EMG and EEG signals were amplified at a gain of 10,000 and filtered between 100–500 Hz and 0.1–500 Hz, respectively. A constant flow of air was delivered through the plethysmograph chamber, respective to the size of the plethysmograph (80 mL/minute flow in 50-ml plethysmographs, 450 mL/minute flow in 260-mL plethysmographs, and 800 mL/minute flow in 500-mL plethysmographs), and pressure fluctuations were recorded via a differential pressure transducer connected to a carrier demodulator (Validyne, USA). The pressure fluctuations resulted in a signal indicative of inspiratory and expiratory events. The airflow signal was high-pass filtered with a frequency cutoff of 0.01 Hz to eliminate small signal drifts during the recording session. Also, a video camera was setup to simultaneously record overt behavior in order to allow for analysis of sleep/wake states.²⁹ A total recording session of 2 hours was collected for all age groups with the exception of P15 rats, which underwent a 5-hour recording session in order to obtain sufficient periods of AS for analysis.

Data Analysis

The first hour of each recording session was excluded from analysis to ensure adequate recovery from the surgery and from the effects of anesthesia. Sleep states were confirmed using video recording (to observe overt behavior and myoclonic twitching), nuchal EMG (to measure muscle tone and body movements), and EEG activity (to measure cortical activity changes in P15 rats). Through video monitoring we were able to identify occurrence of myoclonic twitches against a background of nuchal muscle atonia, which are a defining feature of AS, whereas, behavioral quiescence and low nuchal activity were indicative of QS (Supplementary Video 1). Periods of coordinated motor movements like yawning, kicking, and stretching were indicative of W.²⁹ In P15 rats, EEG recordings provided further classification for sleep states as the convergence of low nuchal activity and fast low-voltage cortical EEG are typical of AS whereas high power delta waves (0.5–4 Hz bandwidth) are a characteristic feature of QS.³³

Within each sleep state, a breath-by-breath analysis was performed on respiratory flow traces to calculate breaths per minute (bpm) and the coefficient of variation of the respiratory period (CV_RP; obtained by calculating the ratio between the standard deviation and the mean BPM), a measure used to quantify respiratory variability.³² Given the limitation in plethysmograph readings associated with signal drifts, moving and gas compression artefacts, the analysis of changes in relative tidal volume (VT) and VE was limited to brief AS epochs that displayed ABD recruitment in order to determine differences in VT and VE with ABD onset.

INT and ABD_EMG signals were rectified and integrated (time constant decay 0.08 seconds) and used, together with respiratory flow data, to determine inspiratory and expiratory phases and relative contribution of ABD activity to respiration. ABD_EMG activity was classified as being either tonic (nonrespiratory modulated), weakly expiratory modulated, or robustly recruited (peak amplitude >50% of preceding baseline activity) based on at least three consecutive respiratory cycles.

Analysis of respiratory parameters within each age group was initially performed in QS and AS irrespective of the observed pattern of ABD_EMG activity. Further analysis was performed in order to compare respiratory parameters within sleep states in which robust ABD_EMG was either present (ABD^W) or absent (ABD^W/O). AS epochs that displayed recruitment of ABD_EMG activity were further analyzed to compare breathing characteristics (bpm, CV_RP, VT, and VE) in six respiratory cycles preceding ABD_EMG onset (ABD⁻) with the breathing pattern observed with ABD_EMG onset (ABD⁺). Pressure volume artefacts in the plethysmograph signal and large swings in INT and ABD_EMG activity occurring together with myoclonic twitches were eliminated from our ABD^+/− analysis.

In order to further characterize the relationship between the occurrence of twitches and INT and ABD_EMG activity in AS, we performed twitch-triggered analysis of ʃINT and ʃABD_EMG amplitude in the 2-second interval preceding and following the occurrence of twitches. Twitches at AS onset and occurring within a 4-second intertwitch interval were eliminated from the analysis to avoid overlap.

Average values for each age group were calculated, normalized, and compared (analysis of variance [ANOVA], Tukey test, and post hoc t-test) to test for significance between means of respiratory variables within age groups and across age groups. Data are presented as mean ± standard error of the mean. p values less than .05 were considered significant.

RESULTS

All rats used in this study were the result of spontaneous deliveries and were tested to determine sleep states, ventilatory parameters, and the contribution of ABD muscle recruitment to ventilation during normal “healthy” postnatal development. P0 rats were tested within 10 hours of delivery to investigate ABD recruitment in presence of an immature and often irregular breathing pattern.³⁴ In contrast, P1 rats displayed a more regular breathing pattern and therefore were analyzed separately. Other groups were tested within a 48-hour interval: P3-4; P7-8, and P14-15, for simplicity, these groups are indicated as P4, P8, and P15.

Sleep Characteristics in the Postnatal Period

The percentage of time spent in W, QS, and AS for the five age groups examined is displayed in Table 1. Similar to human infants,² and in agreement with previous work done in rodents,^33,35 neonatal rats spent most of their time asleep. Developmentally, the amount of time spent sleeping increased significantly from P0 to P1 (ANOVA, Tukey test, p < .05), did not change between P1 and P8 and then decreased at P15 relative to the P1 and P4 rats (ANOVA, Tukey test, p < .05).

By using nuchal_EMG traces combined with video monitoring of overt behavior, we classified epochs of QS and AS according to sleep-scoring criteria developed by Blumberg et al. for infant rats.²⁹ We extended this sleep-scoring approach for the first time to P0 rodents.

The amount of time spent in QS increased in P15 rats compared to P0 rats (ANOVA, Tukey test, p < .01). In contrast, the amount of time spent in AS decreased significantly during the second week of postnatal development (ANOVA, Tukey test, p < .01; Table 1).

Table 1 also indicates the total number of W, QS, and AS events/hour and their average duration in seconds. W epochs were most frequent in P8 compared to P0-1 and P15 rats (ANOVA, Tukey test, p < .01; Table 1). Their duration was highest at P0 and P15 when rats spent a large part of their time awake. QS epochs increased in number from P0–4 to P8 and then fall precipitously at P15 relative to P4–P8 (ANOVA, Tukey test, p < .05). The duration of QS events increased in P15 compared to P0 and P8 rats (ANOVA, Tukey test, p < .05). Frequency of AS events increased in P0 compared to P8 (ANOVA, Tukey test, p < .05), and decreased at P15 compared to the first postnatal week (ANOVA, Tukey test, p < .01). Mean duration of AS events decreased in P8–P15 relative to P0 (ANOVA, Tukey test, p < .05).

Respiratory Pattern in the Postnatal Period

Respiratory frequency, when calculated across the entire recording session of instrumented rats without consideration of sleep state, averaged 130 ± 11 bpm at P0 (n = 6), 132 ± 12 bpm at P1 (n = 6), 116 ± 5 bpm at P4 (n = 10), 126 ± 8 bpm at P8 (n = 10), and 142 ± 10 bpm in P15 rats (n = 9), with no significant difference in breathing rate across age groups (ANOVA, Tukey test, p > .05). Frequency values tended to be numerically less compared to rats that did not undergo anesthesia and surgery (132 ± 25, 196 ± 1, 195 ± 13, 153 ± 12 bpm at P1, P4, P8, and P15, respectively), but there was no significant difference in comparison to values in noninstrumented rats and across age groups (ANOVA, Tukey test, p > .05).

Respiratory frequency was then specifically examined in AS and QS and are plotted in Figure 1A. Respiratory rate tended to be numerically greater in AS compared to QS, with differences being significant at P4 (p = 3.15 × 10⁻²), P8 (p = 2.615 × 10⁻²), and P15 (p = 5.57 × 10⁻⁴).

Respiratory variability (calculated as CV_RP) across the entire recording session was 0.64 ± 0.06 at P0, 0.31 ± 0.03 at P1, 0.36 ± 0.04 at P4, 0.30 ± 0.03 at P8, and 0.35 ± 0.04 at P15, where variability was significantly higher at P0 relative to all other age groups (ANOVA, Tukey test, p < .01). Respiratory variability was not different in noninstrumented rats (0.39 ± 0.08, 0.19 ± 0.02, 0.28 ± 0.06, 0.43 ± 0.07 at P1, P4, P8, P15, respectively) compared to instrumented rats (ANOVA, Tukey test, p > .05).

Analysis of respiratory variability across sleep states showed that CV_RP was significantly greater in AS compared to QS at all ages (Figure 1B), and the ratio between the CV_RP of AS/QS was greater in P15 rats compared to P0–P8 rats (ANOVA, Tukey test, p < .01). Further, when CV_RP in both AS and QS sleep were compared to the CV_RP obtained through the recording session (no sleep states considered), these values were significantly lower at all ages with the exception of AS in P15 rats.

The occurrence of central apneas, defined as an interruption in respiratory flow lasting longer than the average duration of two missed breaths,³⁶ was greatest at P0 (Table 2). Apneas were more frequent during AS compared to QS; 75 ± 23%, 63 ± 16%, 86 ± 42%, 44 ± 26%, and 60 ± 29% of apneas occurred in AS at P0, P1, P4, P8, and P15, respectively. Post-sigh apneas were rare compared to central apneas in all age groups (Table 2).

Recruitment of ABD_EMG Activity in the Postnatal Period

Analysis of ABD_EMG signals through development and across sleep/wake cycles revealed that ABD_EMG activity was either tonic (i.e., not expiratory modulated), weakly expiratory modulated (i.e., characterized by phasic activity during expiration and silent during inspiration), or robustly recruited during expiration (i.e., active expiration). Multiple patterns of ABD_EMG activity could be observed in any single QS or AS epoch across the different age groups (Figures 2–4), although bouts of high amplitude, expiratory-modulated ABD_EMG recruitment occurred most frequently in AS compared to QS throughout development (Table 3).

Figures 2 and 3 display nuchal_EMG and ABD_EMG activities (raw and integrated signals) and airflow traces recorded across wake/sleep alternations in P0, P1, P4 (Figure 2), and P8 rats (Figure 3A). Analysis of EMG activity and behavioral data indicates that rats transitioned from W, characterized by large body movements and high nuchal tone, to QS, characterized by low nuchal tone and behavioral quiescence, to AS, characterized by intermittent muscle twitches.

Myoclonic twitches could be followed by lasting bouts of high amplitude ABD_EMG activity in AS. These high-amplitude events were preceded by twitches within a 2 second interval in 51.4 ± 13.3% of occurrences at P0, 57.7 ± 11.3% at P1, 62.0 ± 22.5% at P4, 73.9 ± 10.9% at P7 and 49.1 ± 12.3% at P15. The remainder of these events, although occurring in AS, were not preceded by myoclonic twitches. As shown in Figure 3B for P8 rats, we observed a very brief (~500ms on average) and small amplitude increase in both INT_EMG and ABD_EMG associated with twitch occurrence. These small increases did not affect our analysis of changes in INT_EMG and ABD_EMG activation since the immediate time window surrounding twitch events were eliminated from our analysis. Further, twitch-triggered analysis of changes in ABD_EMG (P0–P15) and INT_EMG (P8-P15) peak amplitude activity preceding and following a twitch event (Figure 3C) indicated that, with the exception of P0 rats (+13 ± 0.3% increase in ABD_EMG activity following a twitch compared to pre-twitch activity; p = .01; n = 6) respiratory muscle activity in both INT and ABD muscles was not significantly influenced by occurrence of twitches (p >.05).

Figure 4 displays a sleep/wake cycle from a P15 rat in which sleep pattern was evaluated not only by nuchal_EMG and behavior, but also by cortical EEG activity. QS was characterized by low neck_EMG activity, behavioral quiescence and high power low frequency cortical EEG activity. Similar to previous reports,³³ we often observed transitions from QS to AS and back to QS within the same sleep cycle in P15 rats. The occurrence of twitches throughout AS events and high amplitude ABD bursts following a twitch were less frequent in P15 compared to the younger age groups.

Recruitment of ABD_EMG activity in QS is not associated with significant changes in respiratory variability

Recruitment of high amplitude expiratory ABD_EMG activity during QS epochs was most common in P0 to P4 rats and declined thereafter (P8–15). Bouts of high amplitude ABD_EMG activity occurred in 11.4 ± 3.7% of QS epochs in 5 out of 6 P0 rats, 6.3 ± 1.3% of QS epochs in 4 out of 6 P1 rats and 5.5 ± 3.5% of QS epochs in 6 out of 10 P4 rats (Table 3). High amplitude ABD_EMG recruitment was observed in only 3.0 ± 1.2% of QS epochs in 4 out of 10 P8 rats and in only a few QS epochs in 1 out of 9 P15 rats.

We further characterized breathing pattern by separating QS epochs that did (ABD^W) or did not (ABD^W/O) display high amplitude ABD_EMG recruitment. Our results indicate that respiratory frequency in QS was not affected by the presence of ABD recruitment in the first postnatal week (Figure 5A).

Figure 5B illustrates the distribution of CV_RP in QS epochs that displayed ABD recruitment (ABD^W) or not (ABD^W/O). The CV_RP in ABD^W increased compared to ABD^W/O in QS only at P0 (0.37 ± 0.06 and 0.23 ± 0.04; p = 2.25 × 10⁻³, n = 5), whereas CV_RP did not change at P1 (0.16 ± 0.04 vs 0.12 ± 0.01; p = .20, n = 4), P4 (0.21 ± 0.05 vs 0.16 ± 0.02; p = .09, n = 6) and P8 (0.27 ± 0.06 vs 0.17 ± 0.03; p = .09, n = 4). Only one P15 rat recruited ABD_EMG during QS. Thus, these data were not analyzed further.

Occurrence of apneas was also investigated in ABD^W and ABD^W/O QS epochs across development. 82.3%, 40%, 71.4%, 100% and 89.5% of apneas occurred in ABD^W/O in P0, P1, P4, P8 and P15 rats, respectively. Post sigh apneas in QS were present only in P4-P15 rats, with 100% of them being in ABD^W/O at P4 and 66.7% of them being in ABD^W/O at P8 and P15.

Recruitment of ABD_EMG activity in AS is associated with increased breathing variability

AS epochs were also analyzed according to the presence (ABD^W) or the absence (ABD^W/O) of high amplitude ABD_EMG activity. The proportion of AS epochs that displayed bouts of high amplitude ABD_EMG activity increased within P0-P8 rats and was present in all rats examined at each age group (Table 3). 40.0 ± 4.2% of the AS epochs displayed ABD_EMG recruitment at P0, 48.4 ± 8.2% at P1, 40.8 ± 9.2% at P4, and 39.3 ± 5.8% at P8. Contrary to P4-P8, the percentage of ABD^W events in AS decreased to 22.7 ± 3.8% of AS epochs at P15 (ANOVA, Tukey test p <.01). This value is comparable to what we observed previously in adult rats.^32,37

While we did not observe differences in respiratory rate with occurrence of bouts of high amplitude ABD_EMG recruitment at P0-P1, and P8-P15 (Figure 6A), respiratory rate significantly decreased with ABD_EMG recruitment only in P4 rats (109 ± 7.8 bpm in ABD^W/O vs 104 ± 7.7 bpm in ABD^W; p = 2.2 × 10⁻²).

When CV_RP was compared between ABD^W/O and ABD^W events in AS epochs, we observed an increase in respiratory variability in ABD^W events compared to ABD^W/O events in all age groups, with the exception of P0 (Figure 6B). CV_RP increased from 0.17 ± 0.01 to 0.23 ± 0.01 with ABD recruitment at P1 (p = 3.6 × 10⁻⁴), from 0.19 ± 0.01 to 0.24 ± 0.02 at P4 (p = 1.8 × 10⁻³), from 0.21 ± 0.03 vs 0.25 ± 0.03 at P8 (p = 2.0 × 10⁻²) and from 0.33 ± 0.04 to 0.41 ± 0.04 at P15 (p = 2.3 × 10⁻³).

Apneas occurring in AS were more frequent in ABD^W compared to ABD^W/O events in the first postnatal week (70.6%, 100%, 82.4% and 60.0% of apneas were in ABD^W at P0, P1,P4,P8), whereas at P15 apneas occurred more frequently in ABD^W/O events (36.4%). Post sigh apneas occurred in P0-P4 and P15 rats. At P0 40.0% of post-sigh apneas occurred in ABD^W/O, at P1 100% of post-sigh apneas occurred in ABD^W/O and at P4 and P15 66.7% of post-sigh apneas occurred in ABD^W/O.

The occurrence of ABD recruitment was investigated in relation with the timing of apnea episodes (within a 10s period). In P0 rats, 44.7% of apneas occurred following ABD_EMG recruitment and 28.9% of apneas were both preceded and followed by respiratory cycles that displayed ABD_EMG activity within 10 seconds from the apneas, whereas only 2.6% of apneic events were followed by ABD_EMG recruitment. In P1 rats, 42.9% of apneas were preceded by ABD_EMG recruitment, whereas 28.6% of apneas were both preceded and followed by forced expiratory activity. In P4 rats, 22.2% of apneas were followed by expiratory ABD recruitment, 37.0% of apneas were preceded by expiratory ABD recruitment and 14.8% of apneas were both preceded and followed by forced expiratory activity. In P8 rats, 100% of apneas were both preceded and followed by forced expiratory activity, whereas in P15 rats 50% of apneas were preceded by ABD recruitment and 16.7% of apneas were followed by forced expiratory activity.

The onset of ABD muscle recruitment in AS is associated with a decrease in respiratory variability and an increase in VE

We further analyzed breathing characteristics within ABD^W events occurring in AS. Six respiratory cycles preceding ABD recruitment (ABD^-) and following the onset of ABD recruitment (ABD⁺) were analyzed to determine changes in respiratory rate, respiratory variability, relative VT and VE with ABD recruitment across all age groups.

With the onset of high amplitude ABD_EMG recruitment in ABD^W AS (ABD⁺), there was no change in respiratory frequency compared to the preceding respiratory cycles that did not display ABD_EMG recruitment (ABD⁻; Figure 7A), with the exception of a modest but significant increase in respiratory rate measured at P1 (+11.1 ± 0.3%, p = .02). As shown in Figure 7B, CV_RP significantly decreased in ABD⁺ with the onset of ABD_EMG recruitment compared to preceding ABD^- events at P0 (0.29 ± 0.04 vs 0.18 ± 0.03; p = 1.8 × 10⁻³, n = 6), at P1 (0.13 ± 0.02 vs 0.07 ± 0.01; p = .02, n = 6), at P4 (0.14 ± 0.02 vs 0.06 ± 0.01; p = 2.4 × 10⁻³, n = 10), at P8 (0.13 ± 0.02 vs 0.10 ± 0.01; p = .01, n = 10), and at P15 (0.19 ± 0.05 vs 0.08 ± 0.01; p = .03, n = 9). These data suggest that variability in respiratory rhythm decreases with the onset of high amplitude ABD_EMG recruitment in AS.

Although VT tended to increase in all age groups with recruitment of ABD activity, the increase was only significant at P0 and P8 (+100 ± 9%; p = 6.6 × 10⁻³, Figure 8A) and P8 (+9.7 ± 18.1%; p = .03). Similarly, in older rats (P4-P15) that were also instrumented with INT_EMG electrodes, the peak amplitude of ʃINT_EMG activity indicated a tendency to increase with onset of ABD recruitment, but the changes were not significant (+8.1 ± 24.6% at P4, n = 4; +7.7 ± 33.1% at P8, n = 4; +14.5 ± 16.2%, P15, n = 3).

However, due to the combined changes in rhythm and amplitude that were associated with active expiration, VE increased significantly in all ages except P4 and P15 (P0 = +146 ± 37%, p = 8.2 × 10⁻³; P1 = +36 ± 16%, p = .03; +P8 = 8.9 ± 2.6%, p = 6.9 × 10⁻³; Figure 8B).

DISCUSSION

In this study we investigated respiratory changes occurring in the postnatal period of rodents across QS and AS states using electromyogram and behavioral criteria previously developed and validated by others.^29,35,38 We confirmed that changes in sleep pattern occurs with development and we reported differences in breathing characteristics associated with AS and QS through the first two weeks of postnatal life in rats. Further, for the first time, we analyzed the recruitment of expiratory muscle activity across sleep/wake cycles in postnatal rats. We reported the occurrence of bouts of high amplitude expiratory ABD muscle activity during AS and QS and we demonstrated that the onset of recruitment of expiratory muscle activity was associated with an increase in ventilation and an improvement in respiratory stability in AS. These results indicate that active expiration contributes to ventilation in the first post-natal weeks of rats and its occurrence is associated with increased respiratory stability and improvement of ventilation. The ability of increasing active ventilation during unstable breathing may have important clinical implication for sleep-related breathing disorders in the perinatal period.

Sleep changes across development

Our findings confirmed that sleep pattern changes through development, with a large amount of time spent in AS in the first post-natal week followed by a dramatic AS reduction in favor of QS and W times during the second postnatal week.

Because of absence of cortical delta activity in rats younger than P11, we used criteria developed by Blumberg and colleagues²⁹ to determine sleep states. These criteria established that low nuchal muscle tone and behavioral quiescence in <P11 rats are indicative of QS in absence of a fully developed cortical delta activity. The occurrence of myoclonic twitches on top of the low nuchal muscle tone was indicative of AS. These standards have been extensively validated by further analyses of cortical and hippocampal activity³¹ in addition to studies that have further investigated mechanisms associated with development of extraocular muscle activity and sleep muscle atonia in the first two post-natal weeks.^31,38

Based on these criteria, we observed an increased amount of time spent in W both in P0 (few hours after birth) and in P15 rats. The high prevalence of awake time in P0 rats is in agreement with what was previously observed in humans, where infants spent 30 to 50% of time awake in the first hours post-partum.³⁹ In accordance with previous studies, an increase in awake time and a decrease in AS time occurring in the second post-natal week (compared to the first postnatal week) is evidence of progressive brain development and organization of brain dynamics that regulate brain sleep state alternations and cortical activity.^33,35,40,41

Respiratory changes across development

In this study, we systematically analyzed respiratory parameters and recruitment of expiratory activity in AS and QS by using nuchal_EMG and behavioral activity to rigorously assess wake/sleep cycles in the first two post-natal weeks. This approach has given us the opportunity to determine in each sleep state (AS and QS), respiratory frequency and variability, apnea occurrence, and expiratory muscle recruitment across development.

Previous studies have investigated breathing and its variability across development and its response to different ventilatory challenges regardless of brain or behavioural states.^15–25 Compared to previous studies, we observed a lower respiratory rate in both AS and QS at each age considered.^17,23 We attribute the higher respiratory frequency¹⁷ to the measurement of breathing not only during sleep but also during W and active behaviors. In order to verify this observation we analyzed our data regardless of the sleep states and reported higher values that are comparable to previous studies.^17,22 Because respiratory disorders in the perinatal period occur mostly during sleep, in particular AS,² we believe that the investigation of breathing in specific sleep states may give important contribution to understanding respiratory control in both physiological and pathological conditions.

Analysis of breathing characteristics in developing rats indicates that respiratory rate in baseline conditions was consistent in AS and QS during the first post-natal week and changes in respiratory rate occurred only in P8 and P15 rats, with a significant increase in respiratory rate occurring in AS compared to QS. Nonetheless, in each age group respiratory variability, as indicated by the CV_RP, was consistently higher in AS compared to QS.

Two age groups displayed the highest values in coefficient of variability of the respiratory period in AS P0 and P15. In P0, we observed high respiratory variability in both AS and QS. In the immediate time following birth, respiration is usually more variable in mammals,^34,42,43 with frequent apneas and respiratory irregularities. Post-natal respiratory variability has been attributed to rapidly occurring neonatal adjustments in respiratory mechanics, blood oxygenation, metabolic rate and airway receptors input function associated with air breathing, in addition to the necessary and progressive reabsorption of lung fluids. These adjustments usually occur within the first hours after birth and by P1 respiratory rhythms in mammals become more stable. In support of this, we observed the highest frequency of apneas immediately following birth (P0) and a progressive reduction in respiratory variability and respiratory disturbances in ≥ P1 rats.

Sleep states in P15 rats were analyzed based on their EEG signal and behavioral criteria. Although EEG activity is not as fully developed as it is in adult rats, clear cortical delta activity helped us differentiate the occurrence of QS from AS.³³ The age around P12-15 has been proposed to be a critical respiratory period in rats because of changes in the expression of neurotransmitters and receptors within the respiratory network and variations in ventilatory and metabolic responses to hypoxia.^15–17,28 Here, we show that breathing variability is highest during AS sleep at P15. Although we did not observe the peak of respiratory frequency (~300bpm) that Liu and Wong Riley reported in their study at P15,¹⁷ in either instrumented or non-instrumented rats, we demonstrated that respiratory variability was highest in P15 rats compared to other age groups ≥ P1 and the highest variability was observed in AS compared to QS.

ABD recruitment across development

An objective of our study was to determine if expiratory ABD muscle activity was recruited across sleep/wake cycles in the first two postnatal weeks of rats. Although resistant to hypoxic challenges, at this time of development, critical changes in ventilation, chemosensitivity, lung development and lung compliance occur^28,42,44 and are source of respiratory variability and increased frequency of respiratory disturbances.

Because robust recruitment of expiratory activity was shown to be associated with increased VE and more stable breathing in REM sleep of adult rats,³² an in depth analysis on the occurrence of this activity across development was therefore performed. Similar to our previous results in adult rats,³² we observed clear recruitment of expiratory ABD_EMG activity in all age groups in multiple sleep/wake cycles. Expiratory ABD activity displayed bursts of high amplitude expiratory modulated activity that was associated with expiratory flow. Occurrence of ABD activity was often preceded by myoclonic twitches, although twitches did not influence activity of either inspiratory or expiratory muscles following twitches. Most significantly, recruitment of high amplitude ABD_EMG activity was associated with a reduction in respiratory variability and an increase in ventilation in AS, when the majority of respiratory disturbances occur in infancy. From the results of this study we conclude that in postnatal rats occurrence of high amplitude ABD_EMG activity is consistently present in sleeping rats and it is associated with increased VE and reduction in respiratory frequency variability in AS. It will now be vital to determine the contribution of expiratory ABD_EMG activity in pathological conditions given the association of active expiration with increased ventilation and respiratory stability in AS and REM sleep.³²

The source of the state dependent excitatory drive to expiratory ABD muscle is currently unknown. The retrotrapezoid nucleus/parafacial respiratory group (RTN/pFRG) is the area of the brain that has been proposed to be key for the generation of active recruitment of ABD muscles (forced expiration).^45–49 In vitro and in vivo studies across perinatal development of the expiratory oscillator (epF in embryonic stage and pFRG in juvenile and adult rats)^46,50,51 suggest that the expiratory oscillator is crucial for the onset of respiratory rhythms in rodents embryos⁵⁰ and may still be rhythmically active in the first post-natal period to pace and drive respiration.^46,51 In vivo studies in adult rodents further suggest that with development, pFRG becomes a conditional oscillator that is only active in presence of an increased respiratory drive.^45–47,52

These current results demonstrate that ABD muscle recruitment in the postnatal period is associated with potentiation of VE and with a reduction in respiratory variability. Even though the cellular and network mechanisms underlying the activation of expiratory activity during AS in the postnatal period are still unknown, our data support the hypothesis that recruitment of forced expiration contributes to breathing in the postnatal period of behaving rats.

SUPPLEMENTARY MATERIAL

Supplementary material is available at SLEEP online.

FUNDING

Research funded by NSERC (SP), WCHRI recruitment grant (SP) and CIHR Catalyst Grant on preterm birth (SP).

DISCLOSURE STATEMENT

None declared.

ACKNOWLEDGMENTS

We thank Drs. Funk and Dickson for invaluable comments on the manuscript and assistance with the twitch-triggered analysis of respiratory muscle activity.

REFERENCES

Author notes