Search tips
Search criteria 


Logo of transbThe Royal Society PublishingPhilosophical Transactions BAboutBrowse By SubjectAlertsFree Trial
Philos Trans R Soc Lond B Biol Sci. 2009 September 12; 364(1529): 2577–2587.
PMCID: PMC2865112

Structural and functional architecture of respiratory networks in the mammalian brainstem


Neural circuits controlling breathing in mammals are organized within serially arrayed and functionally interacting brainstem compartments extending from the pons to the lower medulla. The core circuit components that constitute the neural machinery for generating respiratory rhythm and shaping inspiratory and expiratory motor patterns are distributed among three adjacent structural compartments in the ventrolateral medulla: the Bötzinger complex (BötC), pre-Bötzinger complex (pre-BötC) and rostral ventral respiratory group (rVRG). The respiratory rhythm and inspiratory–expiratory patterns emerge from dynamic interactions between: (i) excitatory neuron populations in the pre-BötC and rVRG active during inspiration that form inspiratory motor output; (ii) inhibitory neuron populations in the pre-BötC that provide inspiratory inhibition within the network; and (iii) inhibitory populations in the BötC active during expiration that generate expiratory inhibition. Network interactions within these compartments along with intrinsic rhythmogenic properties of pre-BötC neurons form a hierarchy of multiple oscillatory mechanisms. The functional expression of these mechanisms is controlled by multiple drives from more rostral brainstem components, including the retrotrapezoid nucleus and pons, which regulate the dynamic behaviour of the core circuitry. The emerging view is that the brainstem respiratory network has rhythmogenic capabilities at multiple hierarchical levels, which allows flexible, state-dependent expression of different rhythmogenic mechanisms under different physiological and metabolic conditions and enables a wide repertoire of respiratory behaviours.

Keywords: breathing, brainstem, respiratory rhythm and pattern generation, pons, pre-Bötzinger complex, ventral respiratory column

1. Introduction

Vital circuits in the mammalian brainstem control respiratory movements and maintain homeostasis of the internal physiological state of the brain and body. Breathing is a primal homeostatic neural process, regulating levels of oxygen and carbon dioxide in blood and tissues, which are crucial for life. Rhythmic respiratory movements must occur continuously throughout life and originate from neural activity generated by specially organized circuits in the brainstem. The automatic and seemingly simple nature of the act of inspiration and expiration belies the complexity of the neural machinery involved. Indeed, although neuroscientists have been conducting intensive investigations of neural circuits involved in respiratory rhythm and pattern generation for more than a century, the underlying neural mechanisms are still not fully understood, and many key issues remain unresolved. Here, we provide a synopsis of new emerging views on the spatial and functional architecture of brainstem respiratory circuits.

Innate rhythmic movements such as breathing are thought to be produced by central pattern generator (CPG) networks. These are specialized neuronal circuits in the central nervous system that are intrinsically capable of generating rhythmic activity and motor output without rhythmic inputs from other central circuits and sensorimotor feedback signals (Marder & Calabrese 1996; Grillner et al. 2005). The patterned rhythmic activities generated by CPGs emerge from a combination of synaptic interactions among spatially distributed populations of neurons and intrinsic cellular properties of the neurons comprising the CPG. In addition, the CPG circuits are incorporated into larger neural systems and operate under the control of various central and peripheral sensory inputs that modify the CPG-generated motor pattern, adjusting it to the internal and/or external environment, current motor task and organismal needs. Such inputs regulate not only the frequency and amplitude of the output rhythmic motor activity, but can also dramatically reconfigure the CPG by transforming the operational rhythmogenic and neural pattern formation mechanisms. Thus, CPG circuits not only serve as generators of spatio-temporal patterns of neural activity, but also function as neural substrates for sensorimotor integration. Understanding the complex neural processes involved in the operation of different CPGs, and in particular the respiratory CPG, including the mechanisms underlying the circuit dynamic reconfiguration under different conditions represents a central and challenging problem in neuroscience.

Indeed, breathing is a dynamically mutable motor behaviour that not only performs a vital homeostatic function, but is also integrated with many other physiological functions controlled by the brainstem and higher CNS circuits, such as suckling, swallowing, sniffing, chewing, coughing, vomiting and vocalization. This requires a robust yet highly flexible network organization that can permit multiple state-dependent modes of operation. At the brainstem level, this functional flexibility appears to result from a hierarchically arranged architecture of neural circuits with structurally and functionally compartmentalized components, whose interactions can be dynamically re-organized to produce multiple rhythmic respiratory motor behaviours.

2. Spatially arrayed functional respiratory compartments: overview

Breathing movements are produced by a spatially distributed pontine–medullary respiratory network generating rhythmic patterns of alternating inspiratory and expiratory activities that drive and coordinate the activity of spinal and cranial motoneurons (Cohen 1979; Bianchi et al. 1995; Feldman & Smith 1995; Richter 1996). The motor pattern during normal breathing is considered to consist of three phases: inspiration, post-inspiration (post-I or P-I) and the later stage 2 (E-2) of expiration (Richter 1996). This pattern originates within interconnected bilateral columns of medullary neurons, the ventral respiratory columns (VRCs), and is controlled by inputs from other medullary structures including the retrotrapezoid nucleus (RTN), raphé nuclei and more rostrally pontine circuits. The VRC includes several rostro-caudally arranged compartments (figure 1) (see also Alheid & McCrimmon 2008): the Bötzinger complex (BötC), the pre-Bötzinger complex (pre-BötC) and the rostral (rVRG) and caudal (cVRG) parts of the ventral respiratory group (VRG). Respiratory neurons in these compartments are typically classified based on their firing patterns (e.g. decrementing and augmenting) and their phases of activity relative to the breathing cycle, such as: early-inspiratory (early-I) neurons, with a decrementing firing pattern during inspiration; ramp-inspiratory (ramp-I) neurons, with an augmenting firing pattern during inspiration; post-I neurons, with a decrementing firing pattern during expiration; augmenting expiratory (aug-E) neurons, with an augmenting firing pattern during expiration; and a heterogeneous population of pre-inspiratory (pre-I or also referred to as pre-I/I) neurons that start firing before the onset of inspiration and continue throughout the inspiratory phase (see Richter 1996, for detailed review).

Figure 1.

Spatially arrayed respiratory compartments extending in the brainstem from the pons to caudal areas of the medulla are illustrated in a schematic parasagittal brainstem section at the level of nucleus ambiguus (NA), VRC, facial nucleus (VII) and the pons, ...

The new concept we propose is that these specific populations of respiratory neurons are arrayed in the spatially distinct compartments described earlier and constitute microcircuits that interact within and between these compartments. Furthermore, the microcircuits within each compartment are hypothesized to represent functionally distinct network building blocks. A synopsis of functional properties of these compartments and circuit components is provided in the following.

(a) Pontine circuits

The pontine respiratory regions, called the pontine respiratory group (PRG), include neuronal populations in the Kölliker–Fuse (KF) nucleus and parabrachial (PB) complex (lateral, LPB, and medial, MPB, nuclei) in the rostral dorsolateral pons (figure 1) and several areas in the ventrolateral pons. Although the functional role of the pons in the generation and control of respiratory rhythm and pattern has not been fully established, the above-mentioned pontine regions have been shown to interact with multiple medullary compartments; these interactions provide strong modulation of medullary respiratory network activity and control respiratory phase transitions (e.g. Cohen 1979; St-John 1998; Okazaki et al. 2002; Alheid et al. 2004; Cohen & Shaw 2004; Ezure 2004; Song & Poon 2004; Dutschmann & Herbert 2006; Ezure & Tanaka 2006). Pontine connections to caudal medullary circuits appear to be critical for coordinating the activity of expiratory muscles and upper airway musculature during expiration and for expression and regulation of post-I activity (Dutschmann & Herbert 2006; reviewed in Mörschel & Dutschmann 2009). There is also an emerging consensus that pontine circuits may play a fundamental role in sensorimotor integration (Song & Poon 2004; Potts et al. 2005).

(b) Retrotrapezoid nucleus/parafacial respiratory group compartment

Neuronal clusters that constitute the RTN (Smith et al. 1989) are located below and extend through the rostro-caudal levels of the facial motor nucleus. RTN contains neurons whose activity is sensitive to the brain carbon dioxide level and also receives input from peripheral, oxygen-sensitive chemoreceptors and hence is thought to function as a major chemosensory region (Mulkey et al. 2004; Guyenet et al. 2005). RTN provides chemosensitive excitatory drives to most respiratory populations of the VRC, which modulates the activity of the medullary respiratory network and adapts the activity to the metabolic state of the system (Nattie 1999; Mulkey et al. 2004; Guyenet et al. 2005). Onimaru et al. (1988, 2006) and Onimaru & Homma (1997), studying this area in neonatal rodent in vitro brainstem–spinal cord preparations, found a group of ‘pre-I’ oscillatory neurons located in the so-called parafacial respiratory group (pFRG) that appears to spatially overlap the RTN. They have proposed that the pFRG provides a primary rhythmic excitatory drive that entrains the pre-BötC inspiratory oscillator (below). In contrast, other investigators have proposed that pFRG activity represents an expiratory oscillator that interacts with the inspiratory oscillator in the pre-BötC to generate coordinated patterns of inspiratory and expiratory activity (Feldman & Del Negro 2006; Janczewski & Feldman 2006, also see the following). Thus, an emerging view is that the RTN/pFRG compartment may contain functionally distinct populations of neurons involved in several interrelated basic functions, including coordinating inspiratory and expiratory activities in a metabolic state-dependent manner (see Champagnat et al. 2009 for additional information on structural/functional organization, including developmental origins).

(c) Bötzinger complex compartment

The BötC, with predominately expiratory neurons (post-I and aug-E), is considered to be a major source of expiratory activity in the system during normal breathing (Ezure 1990; Jiang & Lipski 1990; Tian et al. 1999; Ezure et al. 2003). This function of the BötC and its interactions with other VRC compartments have been extensively investigated (Fedorko & Merrill 1984; Long & Duffin 1986; Ezure & Manabe 1988; Ezure 1990; Jiang & Lipski 1990; Tian et al. 1999; Ezure et al. 2003; Shen et al. 2003), although an exclusive role of BötC neurons in the generation of expiration has been recently debated (see RTN/pFRG mentioned earlier). Recent work indicates that the BötC contains critical respiratory network elements generating two major forms of expiratory activity (Smith et al. 2007), which is consistent with a basic role of the BötC in expiratory pattern generation. This role arises from the presence of rhythmic expiratory inhibitory neurons making widely distributed interconnections with other compartments (see §4). The BötC neurons are critically involved in the control of the transition between inspiratory and expiratory activities in the network, which is fundamental for the rhythmic inspiratory–expiratory alternation essential for normal breathing.

(d) Pre-Bötzinger complex compartment

Adjacent and caudal to the BötC, the pre-BötC contains interneuron circuitry essential for generating inspiratory activity (Smith et al. 1991, 2000; Rekling & Feldman 1998; Koshiya & Smith 1999; Feldman & Del Negro 2006). The pre-BötC has been of intense interest because it is thought to function as a kernel structure that is a main source of rhythmic inspiratory excitatory drive to premotor circuits. Furthermore, pre-BötC circuits can express autorhythmic or pacemaker-like activity that generates a rudimentary pattern of inspiratory activity when the structure is experimentally isolated in vitro (Koshiya & Smith 1999; Johnson et al. 2001). This activity is proposed to be based on excitatory synaptic interactions within the pre-BötC and intrinsic cellular mechanisms involving persistent sodium current (INaP, Butera et al. 1999a,b; Smith et al. 2000; Rybak et al. 2003, 2004b; Purvis et al. 2007; Koizumi & Smith 2008). The intrinsic rhythmic activity of the network may also involve calcium-activated non-selective cationic current (ICAN), which in combination with excitatory synaptic interactions can also provide cellular- and network-level rhythmic bursting (Ramirez et al. 2004; Del Negro et al. 2005; Pace et al. 2007). Collectively, these (and other) neuronal currents and excitatory synaptic interactions provide mechanisms for regenerative initiation, maintenance and termination of inspiratory network activity in the isolated pre-BötC in vitro. Mechanisms underlying rhythmic inspiratory pattern generation in the pre-BötC under more physiological conditions (i.e. when the pre-BötC is embedded in the intact brainstem) are more complex, because there is a dynamic overlay of rhythmic inhibitory, tonic excitatory and other modulatory inputs that converge on the pre-BötC excitatory network (Smith et al. 2000, 2007; see §4). Furthermore, the cellular composition of the pre-BötC is heterogeneous, containing different electrophysiological phenotypes (Smith et al. 2000; Del Negro et al. 2001, 2002a,b; Stornetta et al. 2003; Ramirez et al. 2004; Koizumi & Smith 2008), including subpopulations of cranial pre-motoneurons (Koizumi et al. 2008) and populations of rhythmically active GABAergic (Kuwana et al. 2006) and glycinergic inhibitory neurons (Winter et al. 2009). These latter inhibitory neurons function critically in dynamically coordinating inspiratory and expiratory activity (discussed subsequently). Thus, pre-BötC circuits subserve two basic functions: (i) generation of rhythmic excitatory inspiratory drive, including the pre-I component of this drive (Schwarzacher et al. 1995; Smith et al. 2000, 2007; Feldman & Del Negro 2006) that is important for initiation of the inspiratory phase and (ii) coordination of inspiratory–expiratory pattern formation via inspiratory inhibition provided by the pre-BötC inhibitory neurons.

(e) Rostral ventral respiratory group compartment

This compartment contains the main bilateral clusters of bulbospinal inspiratory (ramp-I) excitatory neurons that project to spinal phrenic and intercostal inspiratory motoneurons and shape the inspiratory motor output pattern (Bianchi et al. 1995; Richter 1996). These neurons are driven by the pre-BötC excitatory neurons and are inhibited during expiration by the BötC inhibitory neurons; both of these inputs along with other modulatory drives shape and control the characteristic ramping pattern of inspiratory rVRG activity. Thus, the rVRG is also a compartment with multiple convergent input drives essential for inspiratory pattern formation but in contrast to the pre-BötC, this neuronal population as a whole does not have intrinsic rhythmogenic capability (Smith et al. 2007).

(f) Caudal ventral respiratory group compartment

Excitatory bulbospinal expiratory neurons that project to spinal thoracic and lumbar expiratory motoneurons are concentrated in the retroambiguus area (Ezure 1990). This compartment is presumed to be the expiratory counterpart to the inspiratory rVRG. Convergent inputs including those from BötC that are synaptically integrated in the cVRG, locally shape the patterns of excitatory and inhibitory expiratory bulbospinal drives.

3. A spatial and functional hierarchy of multiple rhythmogenic mechanisms

The spatial and functional organization of the brainstem respiratory network has been recently studied in an in situ arterially perfused rat brainstem–spinal cord preparation (Paton 1996; Pickering & Paton 2006) using sequential rostral-to-caudal microtransections through the brainstem while recording cranial and spinal motor outflow, as well as compartmental neuronal population activity to observe transformations of network behaviour (Rybak et al. 2007; Smith et al. 2007). This approach is similar to that applied in neonatal rat brainstem–spinal cord preparations in vitro that resulted in the original discovery of the pre-BötC (Smith et al. 1991), but now applied in a mature rat nervous system generating neural activity patterns more similar to those in vivo. This approach was designed to test the concept that there exists structural and functional compartmentalization of circuit elements as outlined earlier. Informed by the anatomy, it was hypothesized that there is a rostral-to-caudal stacking of network building blocks subserving distinct circuit functions, which are fully integrated in the intact system, but can be revealed as particular compartments are removed.

The major results obtained from these studies are that sequential reduction of the network progressively re-organizes network dynamics, such that new rhythmogenic mechanisms emerge. Specifically, starting from a transection at the pontine–medullary junction (figure 2), the normal three-phase respiratory pattern is transformed to a two-phase rhythmic pattern lacking the post-I phase. With more caudal transections made close to or at the rostral boundary of pre-BötC (figure 2), the respiratory activity transforms to ‘one-phase’ inspiratory oscillations originating within the pre-BötC, without critical involvement of phasic expiratory inhibition (Rybak et al. 2007; Smith et al. 2007). From these results, we concluded that: (i) generation of the normal three-phase rhythmic pattern requires the presence of the pons (i.e. excitatory drive from pontine neurons to the VRC); (ii) generation of the two-phase pattern is intrinsic to reciprocal inhibitory synaptic interactions between the BötC and the pre-BötC and may also involve the RTN to provide excitatory drive to generate stable behaviour; and (iii) the one-phase inspiratory oscillations are generated within the pre-BötC and rely on intrinsic cellular mechanisms operating in the context of the pre-BötC excitatory network. The latter is inferred since systemic application of riluzole, a blocker of INaP, abolished the one-phase oscillations in contrast to the three-phase and two-phase patterns that persisted in the presence of riluzole (Smith et al. 2007). Remarkably, the disturbances of the pre-BötC inspiratory oscillations by blocking INaP in the reduced mature neuraxis in situ were very similar to those obtained with the pre-BötC compartment isolated in vitro within neonatal rat medullary slices (Koizumi & Smith 2008).

Figure 2.

Transformations of respiratory rhythm and pattern following sequential brainstem transection in the in situ arterially perfused brainstem–spinal cord preparation from juvenile rat, revealing three rhythmic states of the respiratory network as ...

This approach of sequentially reducing the pontine–medullary network, coupled with the analysis of network activity patterns and systematic probing for pre-BötC autorhythmic mechanisms, has provided a new unified view that integrates several earlier models (Smith et al. 2000; Richter & Spyer 2001; Rybak et al. 2004a). The basic conclusion is that there exists a spatial and dynamical hierarchy of interacting pontine, BötC and pre-BötC circuits, each of which controls different aspects of respiratory rhythm generation and pattern formation, which can be revealed as the network is progressively reduced. The expression of each rhythmogenic mechanism is state-dependent and produces specific motor patterns likely to underpin distinct motor behaviours.

4. Core circuitry of the respiratory pattern generation network

A minimal network configuration proposed to represent core circuitry (figure 3) that incorporates the multiple rhythmic states described earlier and their transformations as observed experimentally includes: (i) a mutually inhibitory circuit with a ring-like architecture composed of post-I and aug-E neurons of the BötC compartment, and early-I neurons within the pre-BötC; and (ii) a pre-BötC kernel of excitatory pre-I/inspiratory neurons, with persistent sodium current (INaP)-dependent intrinsic bursting properties. The latter participate dynamically in the expiratory–inspiratory phase transition and generation of the inspiratory phase. Detailed descriptions and computational models of the dynamic operation of this core network and the circuit elements participating in each rhythmic state can be found in a series of related publications (Rybak et al. 2007; Smith et al. 2007; Rubin et al. 2009). Figure 3 shows examples of simulations from these computational models illustrating activity patterns of each population within the core BötC and pre-BötC compartments and the resulting activity pattern downstream in the rVRG, which shapes and provides the essential excitatory drive to spinal inspiratory motoneurons.

Figure 3.

Schematic of the core microcircuitry within BötC and pre-BötC respiratory compartments and activity patterns of respiratory neuron populations of the core circuitry. Shown are the main interacting neuronal populations in these compartments ...

The rhythmogenic mechanisms expressed by this core circuitry are postulated to depend on ‘excitatory drives’ that carry state-characterizing information provided by multiple sources distributed within the brainstem, particularly the pons, RTN, raphé nuclei and nucleus of the solitary tract (not shown in figure 3). These drives include those considered to be major chemoreceptor sites sensing CO2/pH and/or receiving input from peripheral chemoreceptors sensing CO2/pH and low O2 (i.e. RTN and raphé nuclei, Nattie 1999; Richerson 2004; Guyenet et al. 2005). Although currently not completely defined, these drives appear to have a spatial organization that maps specifically on the spatial organization of the network. These drives are conditionally represented in the core circuit schematic (figure 3) by three separate sources located in the pons, RTN and raphé, the latter of which provides excitatory serotonergic modulatory inputs to the pre-BötC and other network components (Richerson 2004; Ptak et al. 2009). These serotonergic inputs have been implicated in state-dependent regulation of respiratory network activity including that in response to hypercapnia (figure 4) and also in pathological respiratory behaviour including sudden infant death syndrome (Paterson et al. 2006). The computational models incorporating the core circuitry and these three major drives (figure 3) have been able to reproduce the three-, two- and one-phase rhythmic regimes and their transformations (Rybak et al. 2007; Smith et al. 2007; Rubin et al. 2009). Importantly, recent modelling has shown that the core circuitry is a robust dynamical structure in which each element of the circuitry represents a node for control of oscillatory frequency by the input drives (Rubin et al. 2009). For example, augmenting net excitatory drive to the pre-BötC excitatory kernel can increase inspiratory frequency by a factor of three, accompanied by a reduction in expiratory duration and a smaller decrease in inspiratory duration. In contrast, increasing drive to the BötC aug-E population can reduce frequency by a factor of three owing to the associated prolongation of expiratory phase duration (Rubin et al. 2009).

Figure 4.

Schematic of an extended model of the brainstem respiratory network indicating possible state-dependent neuronal interactions between the RTN/pFRG, the core BötC and pre-BötC circuitry and more caudal compartmentalized circuit components. ...

5. Other oscillatory mechanisms in the network

The studies of Onimaru & Homma (1997, 2003) and Janczewski & Feldman (2006; see also Feldman & Del Negro 2006) have lead to different proposals for fundamental respiratory rhythm generation mechanisms. It has been suggested that a neural oscillator in the pFRG interacts with the pre-BötC inspiratory oscillator to generate coordinated patterns of inspiratory and expiratory activity in neonatal rodents. While there is currently no clear consensus about the nature and functional role of this mechanism, there has been intense general interest in properties of the pFRG, including developmental origins (Champagnat et al. 2009). This has been bolstered by data of Guyenet et al. (2005), who have shown that the RTN, which as noted earlier spatially overlaps the pFRG, appears to provide a fundamental chemosensory drive to the VRC. Furthermore, developmental studies show that RTN/pFRG cells embryologically have a transcription factor (Phox2b) expression phenotype and that deletion of these cells in mutant mice causes a congenital central hypoventilation-like syndrome associated with irregular breathing, lack of responsiveness to hypercapnia and fatality at birth (Dubreuil et al. 2009).

Recent studies investigating the nature of RTN/pFRG activity and its potential role as an oscillatory mechanism within in situ perfused brainstem–spinal cord preparations from mature rats (Abdala et al. in press) suggest that a subpopulation of RTN/pFRG cells become rhythmically active during hypercapnia. This emergent activity is associated with a transformation of activity of motor outflows generating expiratory muscle activity, particularly abdominal nerve (AbN) motor outflow. During eupnoeic-like respiratory activity, only low-amplitude post-I discharge was observed in AbN expiratory discharge, but AbN late-expiratory (late-E) activity, immediately preceding inspiratory phrenic nerve bursts, occurred during hypercapnia (2–5% elevation of CO2). This AbN activity constitutes active/forced expiration. This activity did not occur after removing the pons by pontine–medullary transections or after chemical inactivation of the RTN region by site-specific application of a GABAA receptor agonist (isoguvacine) to locally silence neurons.

One plausible conclusion from these results is that there are distinct subpopulations of RTN/pFRG neurons operating under control of pontine circuits and the metabolic state of the system that provide the necessary excitation to activate an additional RTN/pFRG oscillatory mechanism. This mechanism operates only under conditions of high respiratory drive to generate a central late-E activity projected to AbN. We hypothesize that this oscillatory mechanism is not a necessary component of the respiratory CPG, but constitutes an adaptive mechanism activated under critical metabolic conditions to entrain both inspiration and active expiration. This view is congruent with the ideas that the pFRG provides both tonic (Guyenet et al. 2005) and rhythmic (Onimaru et al. 2003) pre-I excitatory drives to the BötC, pre-BötC and other VRC circuit elements, and that this region is involved in the generation of active expiration in vivo (Feldman & Del Negro 2006). This late-E/pre-I activity that emerges under specific metabolic conditions is provided by rhythmogenic mechanisms that are distinct from, and additional to, the mechanisms operating within the core VRC circuitry (figure 3). Figure 4 shows a schematic of our extended model of the core circuitry with state-dependent rhythmic interactions between the core respiratory network and the RTN/pFRG that incorporates this additional rhythmogenic mechanism emerging during hypercapnic conditions. As depicted in this figure, it is possible that the RTN/pFRG not only subserves a central chemosensory function but also contains cells integrating a variety of state-dependent signals, including those from the pons to coordinate the activity of abdominal and upper airway musculature, particularly at high levels of respiratory drive.

6. General insights into the functional organization of brainstem respiratory circuits

A number of models describing the functional organization of brainstem respiratory networks have been proposed over the past two decades. These include three-phase models involving predominantly inhibitory network interactions (Richter 1996; Rybak et al. 1997; Richter & Spyer 2001), hybrid pacemaker-network models (Smith et al. 2000; Rybak et al. 2004a, 2009; Rubin et al. 2009) in which the pre-BötC with autorhythmic properties is dynamically controlled by inhibitory networks, a dual inspiratory pacemaker neuron model (Ramirez et al. 2004), an alternating inspiratory–expiratory two-oscillator model (Feldman & Del Negro 2006; Janczewski & Feldman 2006) and a coupled dual excitatory oscillator model (Onimaru & Homma 2003; Onimaru et al. 2006). However, the actual spatial and dynamical organization of the brainstem respiratory network has been difficult to define unequivocally. The core circuitry described earlier and its interactions with other brainstem elements suggest that the pontine–medullary respiratory network has a specific spatial and functional organization extending from the rostral pons to the caudal VRC. Although some respiratory neuron types (e.g. post-I, aug-E and early-I) are not strictly localized within particular medullary compartments, but rather are distributed throughout the VRC (Segers et al. 1987; Ezure 1990; Bianchi et al. 1995), each compartment nevertheless contains dominant populations that may define each compartment's specific functional role. A basic principle is that each compartment operates under the control of more rostral compartments (Smith et al. 2007), constituting a rostro-caudal functional hierarchy of interacting circuit building blocks. Specifically, at the caudal end of the VRC, the cVRG bulbospinal expiratory neurons are controlled by the BötC and inputs from more rostral compartments (e.g. arising within the RTN). Similarly, inspiratory activity of rVRG bulbospinal premotor inspiratory neurons is formed by an excitatory inspiratory synaptic drive from the pre-BötC excitatory neurons and phasic inhibition from BötC expiratory neurons. In turn, the pre-BötC is controlled by the rostrally placed BötC that inhibits the pre-BötC during expiration, whereas the RTN and pontine nuclei provide excitatory drives to BötC, pre-BötC and rVRG. The pontine activation of expiratory BötC populations (especially post-I neurons) provides a widely distributed inhibition within the network during expiration, which appears critical for rhythm and pattern generation under normal conditions. At the level of RTN, at least some neurons (late-E) appear to operate under state-dependent control by the rostral pons.

7. Functional significance of multiple rhythmogenic mechanisms

The functional diversity of the respiratory CPG necessitates that respiratory circuits have a flexible organization, permitting multiple state-dependent modes of operation. The experimental results reviewed earlier indicate that respiratory circuits incorporate at least four oscillatory mechanisms (others may exist) and that such rhythmic states can occur under different physiological and pathophysiological conditions. Changes in metabolic conditions (levels of carbon dioxide/pH, or oxygen) that alter the balance of excitatory/inhibitory drives and excitability of pontine, RTN, BötC and pre-BötC neuronal populations can change network interactions, producing transformations from the normal three-phase rhythmic state to the other rhythmogenic states inherent in the system. For example, severe hypoxia transforms the respiratory system to generate a one-phase, INaP-dependent inspiratory gasping rhythm (Paton et al. 2006) that represents a functional unembedding of INaP-dependent oscillatory mechanisms intrinsic to the pre-BötC. Furthermore, the INaP-dependent mechanism operating in reduced states such as in vitro may not be exclusive, and the pre-BötC may have several intrinsic mechanisms of inspiratory rhythm generation as noted earlier. Thus the pre-BötC, as one of the key compartments, when functionally embedded in the spatially distributed brainstem network, can operate in multiple modes of rhythm generation, depending on interactions with other compartments. Other studies have shown that hypocapnia can convert the network three-phase to a two-phase rhythmic pattern (A. P. L. Abdala, J. C. Smith, I. A. Rybak & J. F. R. Paton 2007, unpublished data; Sun et al. 2001). Emergence of other oscillatory mechanisms such as those underlying active expiration at high levels of respiratory drive (e.g. hypercapnia) was discussed earlier.

8. Summary and perspectives

An emerging view is that the architecture of respiratory CPG circuits incorporates multiple rhythmogenic mechanisms that are inherent in the spatial and functional organization of the brainstem respiratory network. This network has rhythmogenic capabilities at multiple levels of cellular and network organization with oscillatory mechanisms ranging from inhibitory network-based synaptic interactions, to neuronal conductance-based endogenous rhythmogenic mechanisms. The emergence of each oscillatory mechanism depends on the metabolic and functional state of the system and is controlled by multiple sources of drives located within the medulla and pons, some of which are sensitive to the levels CO2, O2 and pH that provide for homeostatic regulation, while others arise from multiple descending systems involved in sensorimotor integration. The architecture of the core circuitry with an embedded excitatory kernel provides multiple nodes for external control of respiratory oscillation frequency and phase durations as well as for dramatically changing the mode of respiratory CPG operation.

Further experimental investigations are required to more precisely define the neuronal connection patterns and dynamic interactions of the core circuit elements and input drives to test and refine the concept of the hierarchical organization of the respiratory CPG. It remains to be established how the brain exploits this architecture and alters network interactions and cellular properties to transform the respiratory rhythm and pattern to behaviourally adapt breathing to different physiological and metabolic states.


The work was supported by the Intramural Research Program of the NIH, NINDS and NINDS, grant R01 NS057815. J.F.R.P. is the recipient of a Royal Society Wolfson Research Merit Award.


One contribution of 17 to a Discussion Meeting Issue ‘Brainstem neural networks vital for life’.


  • Abdala A. P. L., Rybak I. A., Smith J. C., Paton J. F. R. In press Abdominal expiratory activity in the rat brainstem–spinal cord in situ: patterns, origins, and implications for respiratory rhythm generation. J. Physiol [PubMed]
  • Alheid G. F., McCrimmon D. R. 2008. The chemical neuroanatomy of breathing. Respir. Physiol. Neurobiol. 164, 3–11 (doi:10.1016/j.resp.2008.07.014) [PMC free article] [PubMed]
  • Alheid G. F., Milsom W. K., McCrimmon D. R. 2004. Pontine influences on breathing: an overview. Respir. Physiol. Neurobiol. 143, 105–114 (doi:10.1016/j.resp.2004.06.016) [PubMed]
  • Bianchi A. L., Denavit-Saubie M., Champagnat J. 1995. Central control of breathing in mammals: neuronal circuitry, membrane properties, and neurotransmitters. Physiol. Rev. 75, 1–45 [PubMed]
  • Butera R. J., Rinzel J. R., Smith J. C. 1999a. Models of respiratory rhythm generation in the pre-Bötzinger complex. I. Bursting pacemaker neurons. J. Neurophysiol. 82, 382–397 [PubMed]
  • Butera R. J., Rinzel J. R., Smith J. C. 1999b. Models of respiratory rhythm generation in the pre-Bötzinger complex. II. Populations of coupled pacemaker neurons. J. Neurophysiol. 82, 398–415 [PubMed]
  • Champagnat J., Morin-Surun M. P., Fortin G., Thoby-Brisson M. 2009. Developmental basis of the rostro-caudal organization of the brainstem respiratory rhythm generator. Phil. Trans. R. Soc. B 364, 2469–2476 (doi:10.1098/rstb.2009.0090) [PMC free article] [PubMed]
  • Cohen M. I. 1979. Neurogenesis of respiratory rhythm in the mammal. Physiol. Rev. 59, 1105–1173 [PubMed]
  • Cohen M. I., Shaw F. 2004. Role in the inspiratory off-switch of vagal inputs to rostral pontine inspiratory-modulated neurons. Respir. Physiol. Neurobiol. 143, 127–140 (doi:10.1016/j.resp.2004.07.017) [PubMed]
  • Del Negro C. A., Johnson S. M., Butera R. J., Smith J. C. 2001. Models of respiratory rhythm in the pre-Bötzinger complex. III. Experimental tests of model predictions. J. Neurophysiol. 86, 59–74 [PubMed]
  • Del Negro C. A., Koshiya N., Butera R. J., Smith J. C. 2002a. Persistent sodium current, membrane properties and bursting behavior of pre-Bötzinger complex inspiratory neurons in vitro. J. Neurophysiol. 88, 2242–2250 (doi:10.1152/jn.00081.2002) [PubMed]
  • Del Negro C. A., Morgado-Valle C., Feldman J. L. 2002b. Respiratory rhythm: an emergent network property? Neuron 34, 821–830 (doi:10.1016/S0896-6273(02)00712-2) [PubMed]
  • Del Negro C. A., Morgado-Valle C., Hayes J. A., Mackay D. D., Pace R. W., Crowder E. A., Feldman J. L. 2005. Sodium and calcium current-mediated pacemaker neurons and respiratory rhythm generation. J. Neurosci. 25, 446–453 (doi:10.1523/JNEUROSCI.2237-04.2005) [PubMed]
  • Dubreuil V., Burhanin J., Goridis C., Brunet J.-F. 2009. Breathing with Phox2b. Phil. Trans. R. Soc. B 364, 2477–2483 (doi:10.1098/rstb.2009.0085) [PMC free article] [PubMed]
  • Dutschmann M., Herbert H. 2006. The Kölliker–Fuse nucleus gates the post-inspiratory phase of the respiratory cycle to control inspiratory off-switch and upper airway resistance in rat. Eur. J. Neurosci. 24, 1071–1084 (doi:10.1111/j.1460-9568.2006.04981.x) [PubMed]
  • Ezure K. 1990. Synaptic connections between medullary respiratory neurons and consideration on the genesis of respiratory rhythm. Prog. Neurobiol. 35, 429–450 (doi:10.1016/0301-0082(90)90030-K) [PubMed]
  • Ezure K. 2004. Respiration-related afferents to parabrachial pontine regions. Respir. Physiol. Neurobiol. 143, 167–175 (doi:10.1016/j.resp.2004.03.017) [PubMed]
  • Ezure K., Manabe M. 1988. Decrementing expiratory neurons of the Bötzinger complex. Exp. Brain Res. 72, 159–166 (doi:10.1007/BF00248511) [PubMed]
  • Ezure K., Tanaka I. 2006. Distribution and medullary projection of respiratory neurons in the dorsolateral pons of the rat. Neuroscience 141, 1011–1023 (doi:10.1016/j.neuroscience.2006.04.020) [PubMed]
  • Ezure K., Tanaka I., Kondo M. 2003. Glycine is used as a transmitter by decrementing expiratory neurons of the ventrolateral medulla in the rat. J. Neurosci. 23, 8941–8948 [PubMed]
  • Fedorko L., Merrill E. G. 1984. Axonal projections from the rostral expiratory neurones of the Bötzinger complex to medulla and spinal cord in the cat. J. Physiol. 350, 487–496 [PubMed]
  • Feldman J. L., Del Negro C. A. 2006. Looking for inspiration: new perspectives on respiratory rhythm. Nat. Rev. Neurosci. 7, 232–241 (doi:10.1038/nrn1871) [PMC free article] [PubMed]
  • Feldman J. L., Smith J. C. 1995. Neural control of respiratory pattern in mammals: an overview. In Regulation of breathing (eds Dempsey J. A., Pack A. I., editors. ), pp. 39–69 New York, NY: Decker
  • Grillner S., Markram H., DeSchutter E., Silberberg G., LeBeau F. E. N. 2005. Microcircuits in action: from CPGs to neocortex. Trends Neurosci. 28, 525–533 (doi:10.1016/j.tins.2005.08.003) [PubMed]
  • Guyenet P. G., Mulkey D. K., Stornetta R. L., Bayliss D. A. 2005. Regulation of ventral surface chemoreceptors by the central respiratory pattern generator. J. Neurosci. 25, 8938–8947 (doi:10.1523/JNEUROSCI.2415-05.2005) [PubMed]
  • Janczewski W. A., Feldman J. L. 2006. Distinct rhythm generators for inspiration and expiration in the juvenile rat. J. Physiol. 570, 407–420 [PubMed]
  • Jiang C., Lipski J. 1990. Extensive monosynaptic inhibition of ventral respiratory group neurons by augmenting neurons in the Bötzinger complex in the cat. Exp. Brain Res. 81, 639–648 (doi:10.1007/BF02423514) [PubMed]
  • Johnson S. M., Koshiya N., Smith J. C. 2001. Isolation of the kernel for respiratory rhythm generation in a novel preparation: the pre-Bötzinger complex ‘island’. J. Neurophysiol. 85, 1772–1776 [PubMed]
  • Koizumi H., Smith J. C. 2008. Persistent Na+ K+-dominated leak currents contribute to respiratory rhythm generation in the pre-Bötzinger complex in vitro. J. Neurosci. 28, 1773–1785 (doi:10.1523/JNEUROSCI.3916-07.2008) [PubMed]
  • Koizumi H., Wilson C. G., Wong S., Yamanishi T., Koshiya N., Smith J. C. 2008. Functional imaging, spatial reconstruction, and biophysical analysis of a respiratory motor circuit isolated in vitro. J. Neurosci. 28, 2353–2365 (doi:10.1523/JNEUROSCI.3553-07.2008) [PubMed]
  • Koshiya N., Smith J. C. 1999. Neuronal pacemaker for breathing visualized in vitro. Nature 400, 360–363 (doi:10.1038/22540) [PubMed]
  • Kuwana S., Tsunekawa N., Yanagawa Y., Okada Y., Kuribayashi J., Obata K. 2006. Electrophysiological and morphological characteristics of GABAergic respiratory neurons in the mouse pre-Bötzinger complex. Eur. J. Neurosci. 23, 667–674 (doi:10.1111/j.1460-9568.2006.04591.x) [PubMed]
  • Long S. E., Duffin J. 1986. The neuronal determinants of respiratory rhythm. Prog. Neurobiol. 27, 101–182 (doi:10.1016/0301-0082(86)90007-9) [PubMed]
  • Nattie E. 1999. CO2, brainstem chemoreceptors and breathing. Prog. Neurobiol. 59, 299–331 (doi:10.1016/S0301-0082(99)00008-8) [PubMed]
  • Marder E., Calabrese R. L. 1996. Principles of rhythmic motor pattern generation. Physiol. Rev. 76, 687–717 [PubMed]
  • Mörschel M., Dutschmann M. 2009. Pontine respiratory activity involved in inspiratory/expiratory phase transition. Phil. Trans. R. Soc. B 364, 2517–2526 (doi:10.1098/rstb.2009.0074) [PMC free article] [PubMed]
  • Mulkey D. K., Stornetta R. L., Weston M. C., Simmons J. R., Parker A., Bayliss D. A., Guyenet P. G. 2004. Respiratory control by ventral surface chemoreceptor neurons in rats. Nat. Neurosci. 7, 1360–1369 (doi:10.1038/nn1357) [PubMed]
  • Okazaki M., Takeda R., Yamazaki H., Haji A. 2002. Synaptic mechanisms of inspiratory off-switching evoked by pontine pneumotaxic stimulation in cats. Neurosci. Res. 44, 101–110 (doi:10.1016/S0168-0102(02)00091-3) [PubMed]
  • Onimaru H., Homma I. 1997. Neuronal mechanisms of respiratory rhythm generation: an approach using in vitro preparation. Jpn J. Physiol. 47, 385–403 (doi:10.2170/jjphysiol.47.385) [PubMed]
  • Onimaru H., Homma I. 2003. A novel functional neuron group for respiratory rhythm generation in the ventral medulla. J. Neurosci. 23, 1478–1486 [PubMed]
  • Onimaru H., Arata A., Homma I. 1988. Primary respiratory rhythm generator in the medulla of brainstem–spinal cord preparation from newborn rat. Brain Res. 445, 314–324 (doi:10.1016/0006-8993(88)91194-8) [PubMed]
  • Onimaru H., Kumagawa Y., Homma I. 2006. Respiration-related rhythmic activity in the rostral medulla of newborn rats. J. Neurophysiol. 96, 55–61 (doi:10.1152/jn.01175.2005) [PubMed]
  • Pace R. W., Mackay D. D., Feldman J. L., Del Negro C. A. 2007. Inspiratory bursts in the preBötzinger complex depend on calcium-activated nonspecific cationic current linked to glutamate receptors. J. Physiol. 582, 113–125 (doi:10.1113/jphysiol.2007.133660) [PubMed]
  • Paterson D. S., Trachtenberg F. L., Thompson E. G., Belliveau R. A., Beggs A. H., Darnall R., Chadwick A. E., Krous H. F., Kinney H. C. 2006. Multiple serotonergic brainstem abnormalities in sudden infant death syndrome. JAMA 296, 2124–2132 (doi:10.1001/jama.296.17.2124) [PubMed]
  • Ptak K., Yamanishi T., Aungst J., Milescu L., Zhang R., Richerson G. B., Smith J. C. 2009. Raphé neurons stimulate respiratory circuit activity by multiple mechanisms via endogenously released serotonin and substance P. J. Neurosci. 29, 3720–3737 (doi:10.1523/JNEUROSCI.5271-08.2009) [PMC free article] [PubMed]
  • Paton J. F. R. 1996. A working heart–brainstem preparation of the mouse. J. Neurosci. Methods 65, 63–68 (doi:10.1016/0165-0270(95)00147-6) [PubMed]
  • Paton J. F. R., Abdala A. P. L., Koizumi H., Smith J. C., St-John W. M. 2006. Respiratory rhythm generation during gasping depends on persistent sodium current. Nat. Neurosci. 9, 311–313 (doi:10.1038/nn1650) [PubMed]
  • Pickering A. E., Paton J. F. R. 2006. A decerebrate, artificially-perfused in situ preparation of rat: utility for the study of autonomic and nociceptive processing. J. Neurosci. Methods 155, 260–271 (doi:10.1016/j.jneumeth.2006.01.011) [PubMed]
  • Potts J. T., Rybak I. A., Paton J. F. R. 2005. Respiratory rhythm entrainment by somatic afferent stimulation. J. Neurosci. 25, 1965–1978 (doi:10.1523/JNEUROSCI.3881-04.2005) [PubMed]
  • Purvis L., Smith J. C., Koizumi H., Butera R. J. 2007. Intrinsic bursters increase the robustness of rhythm generation in an excitatory network. J. Neurophysiol. 97, 1515–1526 (doi:10.1152/jn.00908.2006) [PubMed]
  • Ramirez J. M., Tryba A. K., Pena F. 2004. Pacemaker neurons and neuronal networks: an integrative view. Curr. Opin. Neurobiol. 14, 665–674 (doi:10.1016/j.conb.2004.10.011) [PubMed]
  • Rekling J. C., Feldman J. L. 1998. Pre-Bötzinger complex and pacemaker neurons: hypothesized site and kernel for respiratory rhythm generation. Ann. Rev. Physiol. 60, 385–405 (doi:10.1146/annurev.physiol.60.1.385) [PubMed]
  • Richerson G. B. 2004. Serotonergic neurons as carbon dioxide sensors that maintain pH homeostasis. Nat. Rev. Neurosci. 5, 449–461 (doi:10.1038/nrn1409) [PubMed]
  • Richter D. W. 1996. Neural regulation of respiration: rhythmogenesis and afferent control. In Comprehensive human physiology, vol. II (eds Gregor R., Windhorst U., editors. ), pp. 2079–2095 Berlin, Germany: Springer-Verlag
  • Richter D. W., Spyer K. M. 2001. Studying rhythmogenesis of breathing: comparison of in vitro and in vivo models. Trends Neurosci. 24, 464–472 (doi:10.1016/S0166-2236(00)01867-1) [PubMed]
  • Rubin J. E., Shevtsova N. A., Ermentrout G. B., Smith J. C., Rybak I. A. 2009. Multiple rhythmic states in a model of the respiratory CPG. J. Neurophysiol. 101, 2146–2165 (doi:10.1152/jn.90958.2008) [PubMed]
  • Rybak I. A., Paton J. F. R., Schwaber J. S. 1997. Modeling neural mechanisms for genesis of respiratory rhythm and pattern. II. Network models of the central respiratory pattern generator. J. Neurophysiol. 77, 2007–2026 [PubMed]
  • Rybak I. A., Shevtsova N. A., St-John W. M., Paton J. F. R., Pierrefiche O. 2003. Endogenous rhythm generation in the pre-Bötzinger complex and ionic currents: modelling and in vitro studies. Eur. J. Neurosci. 18, 239–257 (doi:10.1046/j.1460-9568.2003.02739.x) [PubMed]
  • Rybak I. A., Shevtsova N. A., Paton J. F. R., Dick T. E., St-John W. M., Mörschel M., Dutschmann M. 2004a. Modeling the ponto-medullary respiratory network. Respir. Physiol. Neurobiol. 143, 307–319 (doi:10.1016/j.resp.2004.03.020) [PubMed]
  • Rybak I. A., Shevtsova N. A., Ptak K., McCrimmon D. R. 2004b. Intrinsic bursting activity in the pre-Bötzinger complex: role of persistent sodium and potassium currents. Biol. Cybern. 90, 59–74 (doi:10.1007/s00422-003-0447-1) [PubMed]
  • Rybak I. A., Abdula A. P. L., Markin S. N., Paton J. F. R., Smith J. C. 2007. Spatial organization and state-dependent mechanisms for respiratory rhythm and pattern generation. Prog. Brain Res. 165, 201–220 (doi:10.1016/S0079-6123(06)65013-9) [PMC free article] [PubMed]
  • Rybak I. A., et al. 2009. Reconfiguration of the pontomedullary respiratory network: a computational modeling study with coordinated in vivo experiments. J. Neurophysiol. 100, 1770–1799 (doi:10.1152/jn.90416.2008) [PubMed]
  • Schwarzacher S., Smith J. C., Richter D. W. 1995. The pre-Bötzinger complex in the cat. J. Neurophysiol. 73, 1452–1461 [PubMed]
  • Segers L. S., Shannon R., Saporta S., Lindsey B. G. 1987. Functional associations among simultaneously monitored lateral medullary respiratory neurons in the cat. I. Evidence for excitatory and inhibitory connections of inspiratory neurons. J. Neurophysiol. 57, 1078–1100 [PubMed]
  • Shen L. L., Li Y. M., Duffin J. 2003. Inhibitory connections among rostral medullary expiratory neurones detected with cross-correlation in the decerebrate rat. Pflügers Arch 446, 365–372 (doi:10.1007/S00424-003-1024-0) [PubMed]
  • Smith J. C., Morrison D. E., Ellenberger H. H., Otto M. R., Feldman J. L. 1989. Brainstem projections to the major respiratory neuron populations in the medulla of the cat. J. Comp. Neurol. 281, 69–96 (doi:10.1002/cne.902810107) [PubMed]
  • Smith J. C., Ellenberger H., Ballanyi K., Richter D. W., Feldman J. L. 1991. Pre-Bötzinger complex: a brain stem region that may generate respiratory rhythm in mammals. Science 254, 726–729 (doi:10.1126/science.1683005) [PMC free article] [PubMed]
  • Smith J. C., Butera R. J., Koshiya N., Del Negro C., Wilson C. G., Johnson S. M. 2000. Respiratory rhythm generation in neonatal and adult mammals: the hybrid pacemaker-network model. Respir. Physiol. Neurobiol. 122, 131–147 [PubMed]
  • Smith J. C., Abdala A. P. L., Koizumi H., Rybak I. A., Paton J. F. R. 2007. Spatial and functional architecture of the mammalian brainstem respiratory network: a hierarchy of three oscillatory mechanisms. J. Neurophysiol. 98, 3370–3387 (doi:10.1152/jn.00985.2007) [PMC free article] [PubMed]
  • Song G., Poon S. 2004. Functional and structural models of pontine modulation of mechanoreceptor and chemoreceptor reflexes. Respir. Physiol. Neurobiol. 143, 281–292 (doi:10.1016/j.resp.2004.05.009) [PubMed]
  • St-John W. M. 1998. Neurogenesis of patterns of automatic ventilatory activity. Prog. Neurobiol. 56, 97–117 (doi:10.1016/S0301-0082(98)00031-8) [PubMed]
  • Stornetta R. L., Rosin D. L., Wang H., Sevigny C. P., Weston M. C., Guyenet P. G. 2003. A group of glutamatergic interneurons expressing high levels of both neurokinin-1 receptors and somatostatin identifies the region of the pre-Bötzinger complex. J. Comp. Neurol. 455, 499–512 (doi:10.1002/cne.10504) [PubMed]
  • Sun Q.-J., Goodchild A. K., Pilowsky P. M. 2001. Firing patterns of pre-Bötzinger and Bötzinger neurons during hypocapnia in the adult rat. Brain Res. 903, 198–206 (doi:10.1016/S0006-8993(01)02447-7) [PubMed]
  • Tian G. F., Peever J. H., Duffin J. 1999. Bötzinger-complex, bulbospinal expiratory neurones monosynaptically inhibit ventral-group respiratory neurones in the decerebrate rat. Exp. Brain Res. 124, 173–180 (doi:10.1007/s002210050612) [PubMed]
  • Winter S. M., Fresemann J., Schnell C., Oku Y., Hirrlinger J., Hülsmann S. 2009. Glycinergic interneurons are functionally integrated into the inspiratory network of mouse medullary slices. Pflugers Arch. 458, 459–469 (doi:10.1007/s00424-009-0647-1) [PMC free article] [PubMed]

Articles from Philosophical Transactions of the Royal Society B: Biological Sciences are provided here courtesy of The Royal Society