- Donald EdwardsDonald EdwardsGeorgia State University
Crayfish are decapod crustaceans that use different forms of escape to flee from different types of predatory attacks. Lateral and Medial Giant escapes are released by giant interneurons of the same name in response to sudden, sharp attacks from the rear and front of the animal, respectively. A Lateral Giant (LG) escape uses a fast rostral abdominal flexion to pitch the animal up and forward at very short latency. It is succeeded by guided swimming movements powered by a series of rapid abdominal flexions and extensions. A Medial Giant (MG) escape uses a fast, full abdominal flexion to thrust the animal directly backward, and is also followed by swimming that moves the animal rapidly away from the attacker. More slowly developing attacks evoke Non-Giant (NG) escapes, which have a longer latency, are varied in the form of abdominal flexion, and are directed initially away from the attacker. They, too, are followed by swimming away from the attacker. The neural circuitry for LG escape has been extensively studied and has provided insights into the neural control of behavior, synaptic integration, coincidence detection, electrical synapses, behavioral and synaptic plasticity, neuroeconomical decision-making, and the modulatory effects of monoamines and of changes in the animal’s social status.
- Invertebrate Neuroscience
Overview: Animal Escape Behaviors
Escape is as old and as universal as animal predation, being one of the major ways prey species avoid predators (Sillar, Picton, & Heitler, 2016). It is part of the predator/prey evolutionary arms race, so that in higher forms it depends on short-latency detection of a predatory attack and on very fast mechanisms of propulsion away from the predator. In many insects, including flies, locusts, and crickets, escape depends on a powerful jump into the air that precedes flight; other insects, such as cockroaches, use a short-latency turn away from the predator and rapid locomotor acceleration. Teleost fish respond to an attack with a rapid “C”-shaped asymmetrical body contraction and rapid swimming. Mammals produce a short latency startle response that can lead to escape or freezing.
Escape in Crayfish
Crayfish are freshwater decapod crustaceans that have three distinct escape responses (Figure 1): one evoked by an attack to the rear that launches the animal up and forward into a somersault within 30 ms (Figure 1A; Video 1), one evoked by an attack to the front that thrusts the animal directly backward with a similar latency (Figure 1B; Video 2), and one at variable latency (~25 ms if the attack is expected; ~100–200 ms if the attack is unexpected) that directs the animal directly away from the direction of attack (Figure 1C) (Herberholz, Sen, & Edwards, 2004; Wine & Krasne, 1972). The first two escapes are powered by a rapid flexion of the abdomen, a “tailflip,” that is followed by re-extension of the abdomen and a rapid series of short abdominal flexions and extensions called swimming that moves the crayfish away from the direction of attack. The longer latency escape begins with an abdominal extension that transitions directly to swimming. Swimming is guided by spreading and retracting the uropods of the tailfan, which then generate thrust in the desired direction during flexion.
Each of these escapes is mediated by a well-studied neural circuit that has provided insights over the last 70 years into both the specific neurobiological mechanisms of escape and the mechanisms of several widely shared neural processes (Edwards, 2009; Edwards, Heitler, & Krasne, 1999). These include synaptic integration, decision-making, habituation, descending control, electrical synapses, neuromodulation, neural mechanisms of social status, and neuroeconomics. Here I will describe the escape circuitry, its mechanisms, and the associated neural processes. In order to show how the crayfish structure and circuit mechanisms work as a system to produce escape behavior, I illustrate them with simulations from a neuromechanical model of the crayfish and its escape circuits developed in AnimatLab, a neuromechanical simulator (Cofer et al., 2010).
Crayfish escapes are whole body movements that engage the abdomen and all the appendages either actively or passively. Crayfish resemble small clawed lobsters, with pairs of antennae, antennules, eyestalks, and several jaw parts that emerge from the cephalothorax, a pair of large claws that extend forward, and four pairs of walking legs that emerge from the underside and extend laterally (Figure 2A; Holdich, 2002; Huxley, 1880). The cephalothorax is covered by a hard cuticular carapace that provides armor-like protection for the crayfish and for most decapod crustaceans. The abdomen contains five serially arranged ring-like segments and ends in a tailfan that consists of a medial flap, the telson, and a pair of jointed flaps on the left and right sides, the uropods. The uropods can extend laterally to create a large fan-like structure to catch the water and thrust the animal backward during tail flexion, and they can fold medially beneath the telson to present a minimum profile and reduce drag through the water during tail extension.
Abdominal Flexion and Extension
Escape and swimming are powered by two sets of large muscles that fill most of the abdomen and extend into the thorax (Figure 2B). The Fast Flexor (FF) muscles occupy the ventral two-thirds of the abdomen and are responsible for the rapid fast abdominal flexion (“fast flexion”) phase of the escape response. The Fast Extensor (FE) muscles fill the dorsal one-third of the abdomen and are responsible for the rapid re-extension of the abdomen that immediately follows fast flexion.
The serially arranged abdominal segments are connected to their anterior and posterior neighbors through a pair of laterally placed cuticular contacts, or condyles, that together act as a hinge. The hinge allows dorsal/ventral rotation that is limited dorsally and is much greater ventrally (Figure 2C). With such a hinge between each pair of abdominal segments, the abdomen can produce a modest hyperextension and a strong flexion, much like the human elbow or knee. The FF muscles produce abdominal flexion through a complex pulley mechanism. The series of longitudinal muscles originate in each segment, beginning in the telson and project anteriorly for three segments, ending deep into the caudal portion of the thorax (Figure 2B). In each segment, they loop over the central and transverse muscles to form a segmental muscle pulley that amplifies the power of fast abdominal flexion (Figure 2C; Rayner & Wiersma, 1967). The transverse muscles lift the longitudinal muscles and increase their mechanical advantage for flexion. The series of dorsally running extensor muscles are simpler, projecting in parallel between segments from the telson into the thorax (Figure 2B).
The Crayfish Nervous System: Brain
The central nervous system extends from the head through the thorax to the caudal end of the abdomen in a chain of segmental ganglia connected by paired bundles of axons called connectives (Figure 3A). Each ganglion serves as a local processing center for segmental sensory information and motor outputs, and for descending and ascending signals along the connectives. The crayfish brain (Figure 3B), like that of vertebrates and other invertebrate arthropods, is in the most anterior portion of the body (i.e., the head), where it receives afferent information at short latency from the eyes (vision), antennae (touch), antennules (smell and touch), and mouthparts (touch and taste), that enable the animal to determine what lies ahead. The brain also receives sensory information from more caudal parts of the body and from proprioceptors and internal organs through the axon bundles that compose the paired connectives that form the ventral nerve cord. It and the subesophageal ganglion (SEG) produce descending command signals that initiate, inhibit, and coordinate movement.
Thoracic and Abdominal Ganglia
The five thoracic ganglia (Figure 3A, TG) each control a pair of limbs, and the six abdominal ganglia (Figure 3A, AG) control slow flexions and extensions of each segment through the superficial flexor and extensor muscles, and fast abdominal flexion and extension through the FF and FE muscles. Each of the four middle abdominal ganglia also controls a segmental pair of small limbs, the swimmerets. Control is exerted through bilateral sets of nerves that connect sense organs and muscles to the ganglion. In each of the first five abdominal ganglia (Figure 3, A1–A5), three nerves project peripherally on each side (Figure 3C, N1–N3). The first nerve contains segmental afferents from touch receptors on the abdominal surface and from the swimmerets, and motor neurons to the swimmeret muscles. The second nerve contains other touch receptor afferents, muscle receptor organ (MRO) afferents, and motor neurons to both the slow and fast extensor muscles. The third nerve, which exits from the nerve cord caudal to the ganglion, contains only motor neurons to the slow and fast flexor muscles. A6 is the terminal ganglion (Figure 3D), and like the brain, is a fused ganglion that contains elements from the sixth and seventh abdominal segments. There are six paired nerves and one unpaired medial nerve that project from A6 to the periphery. The six paired nerves contain the axons of primary afferents from hair cells and proprioceptors, and axons of motor neurons to the muscles of the telson and uropods.
Giant Axons in Escape
Two of the three forms of crayfish escape are triggered by interneurons with “giant” axons: the Lateral Giant and the Medial Giant. Giant axons help mediate escape responses among a variety of invertebrates, including squid, earthworms, lobsters, crayfish, and flies, and the fish and amphibians among the vertebrates. In each case, these axons are significantly larger in diameter than other axons in the nervous system, and thereby provide a rapidly conducting pathway for an escape signal to travel from the decision point for escape to the motor neurons that effect the escape.
Lateral Giant (LG) Escape
A predator’s attack on the extended abdomen of a quietly resting crayfish will often trigger a fast flexion of the abdomen that thrusts the animal upward and forward, away from the attacker (Figure 1A; Video 1). This “tailflip” is evoked by a single spike in the Lateral Giant (LG) “command” neuron (Krasne, 1969; Wiersma, 1947).
The LG is actually a bilateral chain of tightly coupled interneurons, one per hemisegment, linked together by electrical synapses to form a ladder network similar to that found in earthworms (Figure 3; Günther, 1972). Each LG has a cell body and dendritic arbor in one abdominal hemisegment, and an axon that projects anteriorly to the next segment. This pattern extends into the thorax, and presumably all the way to the brain, where the LG axon ends. However, separate LG neurons in the anterior ganglia have not yet been identified. The LG receives synaptic inputs locally from ipsilateral sensory afferents and ipsi- and contralateral mechanosensory interneurons, and from the contralateral LG and the caudal and rostral segmental neighboring LGs (Figure 4). These LG-LG connections are mediated by high-conductance, non-rectifying electrical synapses. The synapses between contralateral LGs occur on their dendritic branches, whereas those that link longitudinal neighbors occur between the end of the anteriorly projecting axon of the more caudal LG and the initial axon segment of the more rostral LG (arrowheads in Figure 4A). One interneuron can be excited in any abdominal and posterior thoracic segment, and the resulting action potential will spread bilaterally through the network along the length of the abdomen, through the thorax to the brain (Figure 3A).
Model of Crayfish Escape
To illustrate the neural and biomechanical mechanisms of crayfish escape, a neuromechanical model of the crayfish and its escape circuits was developed using the AnimatLab version 1 simulator. The model is described briefly in Figure 5, and its simulated behavioral and neural responses to stimuli that evoke the three types of escape are described in Figure 6.
Converging Inputs to LG
Although descending inputs from the head can help excite LG (Liu & Herberholz, 2010), an LG escape is triggered when an attack primarily directed at the abdomen excites abdominal hair cell afferents and stretch receptors that span the joints of the tailfan uropods (Figure 6A; Video 3; Newland, Aonuma, & Nagayama, 1997). In each abdominal hemisegment, the afferents converge on the distal tips of individual LG dendrites through weakly rectifying electrical synapses, and on a set of mechanosensory interneurons through depression-prone nicotinic cholinergic synapses (Figures 4 and 5; Miller, Vu, & Krasne, 1992; Zucker, 1972a). Some few afferents excite LG directly through chemical synapses (Araki & Nagayama, 2003; Newland et al., 1997). The interneurons excite LG through rectifying electrical synapses along the major branches of the LG dendritic tree (Antonsen & Edwards, 2003). The earlier responding interneurons (e.g., Int A in A6) synapse with LG more distally, and the later-responding interneurons (e.g., Int C) synapse more proximally (Figures 4C, D; Figure 5B). This spatial distribution allows the later inputs to reinforce the incoming depolarizing synaptic potential wave evoked by the earlier, distal inputs. The result is a diphasic excitatory postsynaptic potential (EPSP): an early α component evoked by monosynaptic input from afferents provides a shoulder for a slightly later β component evoked by disynaptic inputs from interneurons (Figure 6A: LG; Krasne, 1969; Zucker, 1972a).
Inhibition helps shape LG’s response and prevents undesirable consequences of both the afferents’ and LGs’ responses (Figure 5B2). The afferent input that excites LG and the mechanosenory interneurons (MSIs) also excites postexcitatory inhibitory (PEI) interneurons that produce feed-forward inhibition of LG (Vu, Berkowitz, & Krasne, 1997). Postexcitatory inhibition limits the period after the initial afferent volley when LG can be excited (Figure 6A), and so helps select for phasic stimuli.
LG usually fires only a single spike in response to its afferent input, but may fire two or more (Figure 6A: LG; Wine & Krasne, 1972). This spike spreads through the ladder-like LG network and excites a recurrent depolarizing inhibition of LG that prevents repetitive firing in the cell (Figure 5B2; Figure 6A; Roberts, 1968). LG also excites inhibitory interneurons that inhibit the primary afferents excited by the attack and those that would be excited by the escape movement. The presynaptic inhibition of the afferents prevents recurrent excitation of the escape circuit by the tailflip movement and it protects afferent synapses to MSIs from synaptic depression (Bryan & Krasne, 1977a, 1977b; Kennedy, Calabrese, & Wine, 1974).
The LG spike excites the motor giant motor neuron (MoG) on each side of the two most caudal thoracic ganglia and the three most rostral abdominal ganglia (Figures 3A; 5C; Heitler & Fraser, 1993; Mittenthal & Wine, 1973; Wine, 1984). These 10 MoGs excite all of the FF muscle in the thorax and first three abdominal segments to contract symmetrically, synchronously, and strongly. The MoGs in more caudal segments receive only subthreshold excitation from LG, and in the terminal ganglion the MoGs are actively inhibited (Dumont & Wine, 1987; Kirk, Dumont, & Wine, 1986). The FF muscle contraction is reinforced by slightly later excitation of the entire set of non-giant FF motor neurons in the same segments (Fraser & Heitler, 1989). The FF motor neurons are weakly excited in each hemisegment by LG (Fraser & Heitler, 1991) and more strongly by the Segmental Giant (SG) interneuron, which acts as a sort of local power amplifier of the LG spike as it passes through the ganglion (Figures 4A, B; 5C; 6A: Mns; Kramer, Krasne, & Wine, 1981). Although SGs in each abdominal ganglion are excited by LG, only those in the first three abdominal segments evoke spikes in segmental FF motor neurons (Miller, Hagiwara, & Wine, 1985; Roberts, Krasne, Hagiwara, Wine, & Kramer, 1982). The FF motor neurons then each excite a subset of the segmental FF muscle.
Fast Abdominal Flexion
Excitation of the MoG and non-giant FF motor neurons (FF Mns) evokes a powerful, rapid abdominal flexion around the three most rostral joints of the abdomen (Figures 1A; 6A: Behavior; Video 1; Video 3). The caudal end of the abdomen remains extended, so that the animal makes a “jackknife” movement around the thoracic-abdominal joint. The uropods spread laterally during fast flexion to maximize the force generated against the substrate and water column. The fast flexion movement begins within 25 ms of the first hostile contact, and causes the extended abdomen to thrust downward against the substrate and thrust the animal upward and forward.
Coincident with excitation of the FF Mns, LG excites two inhibitory neurons to prevent reflexive opposition to fast flexion. LG excites the Extensor Inhibitor motor neuron (EI; Figure 5C) and the Accessory interneurons that inhibit the receptors of the muscle receptor organs (MROs) (Reichert, Wine, & Hagiwara, 1981; Wine, 1977a). EI is an inhibitory motor neuron that causes the fast extensor (FE) muscle to relax during fast flexion. The MROs span the segmental joints in parallel with the abdominal fast extensor (FE) muscle, and their stretch receptors respond vigorously to abdominal flexion. That response, which excites the FE motor neurons and muscle in an abdominal resistance reflex, is delayed during the tailflip escape response by LG excitation of the Accessory interneurons. Finally, LG excites the flexor inhibitor (FI) motor neuron, which inhibits the FF muscle (Wine & Mistick, 1977). In the terminal ganglion, this occurs early and blocks flexion of the telson (Dumont & Wine, 1987). In anterior segments, FI excitation occurs after fast abdominal flexion, and promotes FF muscle relaxation during abdominal extension.
The legs are swept forward during an LG-mediated escape in a manner that streamlines the animal’s movement through the water (Cooke, 1985; Cooke & Macmillan, 1985). The movement may be passive; the common inhibitor motor neuron, which inhibits all peripheral leg muscles, is excited at short latency, whereas the major leg promoter motor neurons are not (Heitler & Fraser, 1989).
Fast Abdominal Extension
Fast abdominal flexion is immediately followed by fast extension (Figure 6A: Behavior), which occurs in response to strong excitation of the FE motor neurons that innervate the FE muscle (Wine, 1977b). LG does not excite the FE motor neurons; rather they are excited in a chained reflex by the MRO stretch receptors that are stimulated by the fast flexion itself (Figures 5C; 6A: MROs, Mns; Reichert et al., 1981; Reichert & Wine, 1983). The MROs are stretched during the fast abdominal flexion, but their stretch receptors are kept from firing by the Accessory interneurons, which are excited by LG. The inhibition ends once the fast flexion is completed, which allows the stretch receptors to respond to the flexed abdomen. Their strong response excites the FE motor neurons to re-extend the abdomen.
Excitation of Swimming
The initial LG-induced tailflip launches the animal off the substrate into the water column and away from the point of attack. The abdominal re-extension begins a rapid series of extension/flexion abdominal movements called swimming that moves the animal rapidly out of the area (Figure 6A: Mns, Behavior). The rhythmic abdominal movements are governed by a central pattern generator (CPG) that is excited by the afferent response to the initial attack and not by LG; its response is delayed until after the LG-initiated fast flexion is complete (Krasne & Wine, 1984).
Medial Giant (MG) Escape
MG Structure and Function
The medial giant command neurons are a bilateral pair of interneurons that have their dendrites and cell bodies in the brain and project their unbranched giant axons to the terminal abdominal ganglion (Figure 3; Glantz & Viancour, 1983). The MGs are excited by attacks directed to the front of the animal and evoke a tailflip escape response that thrusts the animal directly backward (Figure 1B; Herberholz, Sen, & Edwards, 2004; Wine & Krasne, 1982). Like the LG escape, MG escape is immediately followed by abdominal re-extension and swimming (Video 2; Video 4).
The MGs receive visual and tactile sensory inputs, the latter coming from the antennae and from other mechanosensory structures on the cephalothorax and legs. The inputs are cholinergic, and converge on the cell’s dendrites to excite MG at a single spike zone near the initial segment. The MGs are electrically coupled in the brain, allowing synaptic potentials and action potentials to spread contralaterally between the MGs (Glantz & Viancour, 1983).
MG and LG excite different but overlapping sets of segmental MoG and SG neurons in the thoracic and abdominal segments, and these differences account for the different patterns of abdominal flexion that MG and LG spikes evoke (Heitler & Fraser, 1993). Both MG and LG excite SGs in the thoracic and first three abdominal segments, and both excite MoGs in those abdominal segments. But only LG excites thoracic MoGs, and this difference may account for the sharper abdominal flexion that LG produces around the first three abdominal joints. The stronger flexion of the caudal abdomen evoked by MG occurs because it, and not the LG, excites the MoGs and SGs in those segments. This caudal pattern excites all of the abdominal FF muscles (Figure 5C) and produces the strong flexion around each abdominal joint that thrusts the animal backward (Figure 6B; Video 4). Like LG escape, MG fast abdominal flexion is followed by abdominal re-extension and swimming.
Unlike an LG spike, an MG spike actively promotes the legs as part of the escape response. MG drives the fast leg promotor motor neuron (PMM) one-for-one via a rectifying electrical synapse, and the PMM excites leg promotor muscles through a strong, but antifacilitating synapse (Heitler & Fraser, 1989, 1993).
Non-Giant (NG) Escape and Swimming
A “non-giant” (NG) escape can be excited by a phasic stimulus that fails to excite an LG or MG escape (Figure 1C), or it can be “voluntary,” with no apparent stimulus (Reichert & Wine, 1983). Unlike LG and MG escapes, the first movement of an NG escape is an abdominal extension (Reichert et al., 1981). This is followed by the rhythmic abdominal flexion/extension swimming movements that are driven by a central pattern generator (CPG). These movements appear to be driven by the same system that drives swimming that follows LG or MG escapes (Figure 6C; Video 5). The uropods are extended during abdominal flexion to catch as much water as possible and retracted during extension to minimize drag. Crayfish steer during swimming by varying the degree of uropod extension/retraction asymmetrically, and so change the direction of the net force vector. Swimming is visually guided, but the means by which visual stimuli affect the uropod motor output are not known.
Unlike the rigid forms of LG and MG escapes, NG escapes take on whatever form is needed to thrust the animal directly away from the direction of the attack. For attacks from one side, this would require an asymmetric extension of the uropods and twisting the abdomen to thrust the animal away from the side of the attack (Edwards & Mulloney, 1987). Latencies of giant-evoked tailflips are significantly shorter than non-giant tailflips, but only if the triggering attack was unexpected (Herberholz, Sen, & Edwards, 2004). In juvenile crayfish, the latencies of tailflips triggered by LG or MG spikes were identical (~11 ms) to all attacks, whereas the latency of a non-giant escape was ~16 ms when the attack was expected (a visible approaching predator) and ~57 ms when it was unexpected (from a handheld probe).
It is apparent that when a crayfish can predict the direction of a coming attack, it can plan its NG escape and greatly shorten its response latency. The escape plan must include the initial pattern of motor neuron excitation needed to escape directly away from the predator, together with excitation of the swimming CPG. How such a plan is produced is not known, but its production must require much of the ~40 ms difference between the latencies of NG escapes to unexpected and expected attacks. Once the crayfish has an appropriate plan, it can execute it in response to the predator’s attack with the short latency.
Adaptive Features of the Escape Circuits
A particularly striking aspect of the LG escape circuit is the ubiquity and variety of electrical synapses. Electrical synapses occur through gap junctions between two cells and allow small molecules, including ions and water, to pass between them. A difference in the membrane potential of two cells linked by gap junctions will cause a potential difference across the junction itself, and act to drive electrical current in the form of ions through the junction. Action potentials occurring in one neuron can drive current into a coupled cell to produce an electrical synaptic potential that has zero delay and no reversal potential. Because the spread of current through an electrical synapse is passive, the postsynaptic potential varies with but is always smaller than the pre-synaptic potential that caused it, so that electrical synapses lack the amplifying ability of chemical synapses. The amplitude of the electrical postsynaptic potential also depends on the conductance of the synapse and the local input resistance of the postsynaptic cell. In turn, the electrical synaptic conductance depends on the density, number, and conductance of the gap junction channels that make up the electrical synapse.
Rectifying Electrical Synapses at the Input and Output of the LG Escape Circuit
As mentioned above, LG receives rectifying electrical synaptic inputs from mechanosensory afferents and interneurons, and it makes rectifying electrical synaptic outputs with the SGs and FF motor neurons, including the MoG motor neuron (Edwards, Heitler, Leise, & Fricke, 1991; Giaume, Kado, & Korn, 1987; Heitler, Fraser, & Edwards, 1991). Unlike the bidirectional synapses between LGs and between MGs, rectifying electrical synapses are directional. They act like electrical rectifiers, conducting much more synaptic current when the voltage across the synapse is polarized in one direction than in the opposite direction (Figure 7A).
The “giant motor” synapses between the giant neurons, LG and MG, and the MoG are the strongest rectifiers of the circuit, and provided the first and classical demonstration of a rectifying electrical synapse (Furshpan & Potter, 1959). They permit single action potentials in either LG or MG to excite MoG one for one, but prevent the MoG spike from having any effect on the unexcited giant interneuron. The property of electrical rectification means that the synapse and the gap junctions that compose it are asymmetric, with presynaptic and postsynaptic sides. At rest, the synapses are “reverse-biased” by about −15 mV, with the rest potentials of the giant neurons at −85 mV and that of MoG at −70 mV. For the synapse to become conductive, the transynaptic potential must be “forward biased” by 28 mV, meaning the presynaptic LG or MG potential must depolarize by 43 mV (Giaume et al., 1987). Because the synapses occur outside the ganglion at the third nerve root, only an action potential produces such a local presynaptic depolarization. Rectifying electrical synapses also occur in each hemiganglion between the giant neurons and SG, and between SG and the FF motor neurons. They allow an action potential in LG or MG to excite SG, and the SG to excite FF motor neurons, with a minimal delay and in the one-for-one fashion needed to produce the tailflip fast abdominal flexion (Heitler et al., 1991).
Coincidence Detection Mechanisms in the LG Circuit
Escape is an emergency reaction to a predator’s attack, and so has a high stimulus threshold and sensitivity to phasic stimuli. These requirements are met in LG through several coincidence detection mechanisms, including a lateral excitatory network among the afferents, rectifying electrical synapses between the afferents and LG, post-excitatory inhibition, and the electrotonic structure of LG.
Lateral Excitatory Network
Coupling between the afferent axons that converge on a single dendritic branch promotes synchrony of their spikes and discriminates against non-synchronous spikes (Figure 5B) (Herberholz, Antonsen, & Edwards, 2002). When a group of afferents are excited by an attack, current flow between them will help to synchronize their spikes. However, if only a few afferents are excited, their current will sink into neighboring afferents and not contribute to the PSP in LG. This “lateral excitatory network” creates a positive feedback mechanism for synchronizing afferent inputs (Antonsen, Herberholz, & Edwards, 2005; Herberholz et al., 2002).
Rectifying Electrical Synapses
Rectifying electrical synapses are highly conducting when forward biased and non-conductive when reverse-biased. This property selects for synchronous inputs to create the largest and most phasic postsynaptic potential; later inputs will find a strongly depolarized postsynaptic terminal and a reverse-biased synapse and be unable to contribute to the PSP (Figure 7B; Edwards, Yeh, & Krasne, 1998).
Postexcitatory Inhibition (PEI)
In addition to exciting LG and mechanosensory interneurons like LG, the phasic afferent input also excites inhibitory interneurons that inhibit LG distally on its dendritic branches (Figure 5B2; Figure 6A; Vu et al., 1997; Vu & Krasne, 1992). This depolarizing inhibition effectively shunts the excitatory postsynaptic current that converges on LG’s initial segment. It begins within four ms of the start of the LG EPSP and curtails its β component to shorten the window for occurrence of an LG spike.
The Electrotonic Structure of LG
LG has a tree-like structure, with numerous small dendrites emerging from thicker branches that grow out of the large initial segment of the axon (Figures 4 and 5). This structure causes the EPSPs evoked by afferents in distal dendrites to be quite large (~60 mV) and phasic, but to attenuate and slow as they conduct proximally (Figure 6A: LG; Edwards et al., 1994). Synaptic currents from inputs on the separate dendrites converge at the branch points of the tree to produce the maximal EPSP. Dendrites without inputs will provide sinks for synaptic current from other dendrites. In addition, the cell’s giant axon provides a sink for synaptic current arriving at the initial segment, and so quickly reduces the EPSP created there (Antonsen et al., 2005).
Proximal and Distal Inhibition
Afferent inputs to LG occur on the distal tips of dendrites, whereas inhibitory inputs occur both distally and proximally, and have distinct effects on the integration of EPSPs (Vu & Krasne, 1992). Descending inhibition of LG derives from higher centers and acts to modulate the excitability of LG (Vu, Lee, & Krasne, 1993; Vu & Krasne, 1993). Both descending inhibition and postexcitatory inhibition of LG occur distally (Vu et al., 1997), whereas recurrent inhibition occurs more proximally at the initial axon segment (Figure 5C; Roberts, 1968). Distal inhibition can target selected excitatory inputs and prevent their effects from exciting the cell. Distal inhibition can also be overcome by increased excitatory input to the same or adjacent branches: the local membrane potential is determined by voltage division between the excitatory and inhibitory reversal potentials (for the afferent input to LG, which is mediated by weakly rectifying electrical synapses, the excitatory “reversal potential” is the presynaptic action potential). When all EPSPs originate distally, proximal inhibition is absolute. Distal excitatory inputs evoke converging depolarizing currents at proximal segments, where proximal inhibition can shunt them to ground. Increasing the number or strength of the excitatory inputs can add little to their excitatory effect at the initial segment in the face of proximal inhibition, which has a “veto” effect on the cell’s response to those inputs (Vu & Krasne, 1992). This functional synaptic organization is likely to be found in other electrotonically extended dendritic structures in which proximal and distal inhibition affect a neuron’s response to dendritic synaptic inputs (Silver, 2010).
LG as a Command Neuron
A “command neuron” is one that releases a pattern of behavior upon firing a single spike or train of spikes. LG is one of the few to meet the “necessity” requirement of a command neuron, but it fails the “sufficiency” requirement (Kupfermann & Weiss, 1978; Olson & Krasne, 1981): the pattern of animal movement evoked by artificially stimulating an LG neuron differs from the pattern evoked by a simulated attack. That said, it is clear that LG and (most likely) MG are each essential for their two forms of escape (Edwards et al., 1999).
Both cells usually only fire once in response to an attack, and that single spike is sufficient to release the appropriate tailflip response. Repetitive firing in either cell is prevented by a very fast, long-lasting recurrent and proximal inhibition produced by a depolarizing inhibitory postsynaptic potential (IPSP) that immediately follows the cell spike. That depolarizing IPSP is blocked by picrotoxin and mimicked by exposure to γ-aminobutyric acid (GABA), but its source has not been identified (Roberts, 1968).
Attacks directed at the cephalothorax in the front of the animal primarily excite the MG neuron, but LG receives EPSPs as well (Liu & Herberholz, 2010). Moreover, an MG tailflip will often cause the tailfan to strike the substrate, which would tend to excite the LG. This excitation is prevented by strong, short-latency inhibition of LG by MG (Figure 6B). Like recurrent inhibition, MG inhibition of LG is mediated by short-latency, picrotoxin-sensitive, long depolarizing IPSPs (Roberts, 1968). The inhibitory interneuron that mediates this inhibition has not been identified. Although some evidence suggests that MG is not similarly inhibited by LG (Herberholz & Edwards, 2005), this bears re-examination. An LG tailflip thrusts the front end of the animal downward, which should excite mechanosensory afferents that drive MG. Such an MG command might disrupt the LG flip or the re-extension that follows, and so is likely to be prevented.
Plasticity and Modulation of Escape Excitability
Habituation and Protection
In adult animals, but not in juveniles, the LG escape habituates to repeated mechanical stimulation of the abdomen or electric shock of sensory nerves (Krasne, 1969). Part of this habituation is caused by synaptic depression at two sites, the cholinergic synapses made by afferents onto LG and the MSIs (Krasne, 1969; Zucker, 1972b), and the neuromuscular junctions made by the MoG motor neuron with the FF muscle (Bruner & Kennedy, 1970). Synaptic depression at both sites results in part from a reduction in transmitter release (Zucker, 1972b).
In juvenile crayfish, the synaptic depression is masked by the action of the parallel electrical synapses from primary afferents, which evoke a superthreshold α EPSP in LG that does not depress (Fricke, 1984). As the crayfish grows, LG’s electrotonic structure lengthens, and the phasic α EPSP is more severely attenuated than the slower, more proximal, depression-prone β EPSP from MSIs (Edwards et al., 1994; Hill, Edwards, & Murphey, 1994). The synaptic depression is unmasked with continued growth of LG as the reliable α; EPSP is reduced in size and fails to reach threshold, while the depression-prone β EPSP evoked by the MSIs remains superthreshold (Fricke, 1984; Edwards, Fricke, Barnett, Yeh, & Leise, 1994). LG’s response to repetitive stimuli then habituates as the β EPSP depresses and falls below threshold.
Because walking, swimming, and the escape tailflips would all repeatedly excite the abdominal mechanosensory afferents, transmission at their synapses must be blocked to prevent both inappropriate excitation of LG and synaptic depression (Bryan & Krasne, 1977a, 1977b). Transmission is blocked by primary afferent depolarization (PAD) at the afferent terminals; as in vertebrates (Eccles, Kostyuk, & Schmidt, 1962), this form of presynaptic inhibition prevents transmitter release and thereby protects the synapses from depression (Kennedy et al., 1974). PAD can be produced by LG or MG stimulation, by passive leg movement, and by stimulation of sensory afferents that innervate the uropods.
The greater part of LG escape habituation is produced by higher centers rather than by synaptic depression in the afferent pathway to LG (Krasne & Teshiba, 1995). Habituation results from descending tonic inhibition of LG that increases LG’s stimulus threshold. The tonic inhibition persists even after the ventral nerve cord between the thorax and abdomen is severed, disconnecting the higher centers from the abdomen (Shirinyan et al., 2006). The descending signal reprograms local inhibitory circuits to raise the stimulus threshold for escape.
Descending control over LG escape is also exerted during walking, during feeding, and when the animal is restrained (Kennedy, Fricke, & Block, 1982; Krasne & Lee, 1988; Krasne & Wine, 1975). In all three instances, the level of tonic inhibition of LG is modulated by the animal’s activity. Tonic inhibition is increased during walking, while presynaptic inhibition of afferent terminals protects the afferent synapses from depression. During feeding, if the crayfish senses that the food is immovable, tonic inhibition is increased to raise the LG escape threshold, but as soon as the food becomes portable, tonic inhibition is removed and the escape threshold falls. In a similar fashion, increased restraint of a crayfish increases tonic inhibition and LG’s threshold, whereas a decrease in restraint removes inhibition, the threshold drops, and “spontaneous” escape tailflips occur.
The excitability of the LG circuit is modulated by at least three monoamines: dopamine, octopamine, and serotonin (5-HT) (Antonsen, 2008; Araki & Nagayama, 2012; Glanzman & Krasne, 1983). LG has at least two 5-HT receptors, including a 5-HT1α receptor that decreases cyclic adenosine monophosphae (cAMP) and protein kinase A (PKA) and is located on the initial axon segment, and a 5-HT2β receptor that increases protein kinase C (PKC) and is located in the dendrites (Edwards & Spitzer, 2006). Other 5-HT receptors that increase PKA are also likely to be present. Serotonergic interneurons have axonal endings in each of the abdominal ganglia on LG’s initial axon segment (Figure 8; Edwards, Yeh, Musolf, Antonsen, & Krasne, 2002). The effect of these serotonergic neurons on LG is unknown, but the effect of bath-applied 5-HT on LG in an isolated ventral nerve cord was intensively studied. Apart from an increase in input resistance, 5-HT was found to have little direct electrophysiological effect on LG, but to have a strong effect on LG’s response to synaptic input from afferents and interneurons (Antonsen & Edwards, 2007).
Remarkably, 5-HT’s effect on LG’s responsiveness depends on how it is applied (Teshiba et al., 2001). A strong (50 μM), rapid (full concentration in a few seconds), and brief (10 min) application of 5-HT reduced both components of LG’s synaptic response to afferent input, whereas a weaker (1 μM) or slower (full concentration in 30 min) application produced a significant increase in LG’s EPSP. Although the effects disappeared immediately after brief exposures, longer (30 min) exposures produced a prolonged (>2 hr) increase in LG’s responsiveness immediately after the exposure ended, regardless of whether the effect during exposure was to facilitate or depress LG’s EPSP. Experiments with second messenger agonists and blockers indicated that the facilitation resulted from activation of a cAMP—PKA pathway, but the inhibition was produced by non-cAMP–PKA pathways (Araki, Nagayama, & Sprayberry, 2005; Lee, Taylor, & Krasne, 2008). Computational modeling suggested that the inhibitory and facilitatory effects of 5-HT resulted from activation of two parallel, competing second messenger pathways in which the inhibitory pathway had a higher threshold but was faster than the facilitatory pathway, and could block it. However, if the facilitatory pathway became active alone, it could block subsequent activation of the inhibitory pathway and also remain persistently active after the 5-HT was removed (Lee et al., 2008). Additional experiments showed that the facilitation is produced by three physiological mechanisms, two at the distal dendrites and afferent electrical synapses that is active directly after 5-HT exposure, and one at the proximal segment that responds later (Antonsen & Edwards, 2007). The conductance of the electrical synapses between the afferents and LG is increased early, as is the input resistance of the distal dendrites. Both of these changes enhance the LG EPSP. Somewhat later, the input resistance at the initial segment is increased, and this too enhances the EPSP.
Crayfish live in communities at the edges of freshwater ponds or streams, or in low-lying areas near fresh water. They dig burrows for shelter and for egg-bearing females to hatch their young. They compete for these resources with other members of the community, and the competition leads to formation of a social dominance hierarchy (Bovbjerg, 1953; Issa, Adamson, & Edwards, 1999). Winners, which are larger or aggressive, become dominant and have ready access to the available resources, whereas losers become subordinate and use resources unclaimed by more dominant animals. Dominance contests between closely matched crayfish are resolved suddenly. Two crayfish in a dominance contest rarely perform escapes while grappling with one another. The contest is resolved suddenly when one animal (the new subordinate) performs NG or MG escapes followed by swimming, and the other (the new dominant) persists in the attack (Herberholz, Issa, & Edwards, 2001). In contrast, LG tailflips rarely occur in either animal. These results indicate that NG and MG threshold must drop suddenly and persistently in the new subordinate, while LG threshold remains high. Indeed, LG stimulus threshold increases significantly in a new subordinate as it grapples with the dominant and is pushed backward, whereas stimulus threshold stays the same or increases slightly in the dominant (Krasne, Shamsian, & Kulkarni, 1997).
Serotonin exposure has been associated with socially dominant behavior in decapod crustacea (Antonsen & Paul, 1997; Kravitz, 1988; Livingstone, Harris-Warrick, & Kravitz, 1980), and it has opposite effects on LG’s EPSP to afferent stimulation in dominant and subordinate animals (Yeh, Fricke, & Edwards, 1996; Yeh, Musolf, & Edwards, 1997). In socially isolated crayfish, and in social dominants, slow exposure to serotonin increased LG’s EPSP, whereas in social subordinates, slow exposure reduced LG’s EPSP. In newly formed dominant/subordinate pairs, the difference in serotonin’s effect on the LG EPSP developed over a 12 day period, with the facilitating effect of serotonin persisting in the new dominants, while it diminished and then completely reversed in their subordinate partners. If the subordinates were then re-paired and became dominant to a new crayfish, the effect of 5-HT returned to being facilitatory again after several days. Similar changes in the modulatory effects of serotonin with social status have been seen among neurons that control leg posture in crayfish (Issa, Drummond, Cattaert, & Edwards, 2012).
The inhibitory and facilitatory modulatory effects of serotonin could be evoked by different serotonin agonists. mCPP, a 5-HT1α receptor agonist, had no effect on social isolate crayfish, but inhibited LG’s response in both new social subordinates and new social dominants. The inhibitory effects developed gradually in both animals, over the same 12 day time-course as serotonin’s effects developed in new subordinates. As with serotonin, the inhibitory effects of mCPP in both dominants and subordinates reversed after the animals were re-isolated for 8 days. α-methyl-serotonin, a 5-HT2β receptor agonist, had facilitatory effects on the LG EPSP in isolate, dominant and subordinate crayfish. Together, the effects of these agonists and serotonin itself suggest that serotonin’s effects are mediated by different 5-HT receptors that change in their relative effectiveness following a change in the animal’s social status. The 5-HT1α receptor, which downregulates cAMP, is likely to mediate the inhibitory effects of 5-HT and mCPP, whereas the 5-HT2β receptor, which upregulates PKC and diacylglycerol (DAG), may contribute to the facilitatory effects of 5-HT and α-Me-5-HT together with a receptor that upregulates cAMP and PKA (Lee et al., 2008; Spitzer, Edwards, & Baro, 2008).
Sudden but persistent changes in the social experience of an animal can evoke similarly sudden and persistent changes in its behavior. These changes are mediated by neural circuits whose activation must be similarly suddenly and persistently affected. Here, with a change in social status, these changes in crayfish are effected by changing the modulatory effects of serotonin, a ubiquitous neuromodulator, on the excitability of the LG command neuron (and possibly MG), probably by changing the relative expression or efficacy of opposing 5-HT receptors. Similar status-dependent changes in serotonergic modulation occur in circuits that mediate the crayfish’s avoidance response to unexpected touch (Issa et al., 2012; Cattaert, Delbecque, Edwards, & Issa, 2010) and escape in fish (Whitaker et al., 2011).
LG and MG and Behavioral Choice
Like other animals, crayfish are often forced to choose between mutually exclusive behavioral responses to conflicting stimuli (Edwards, 1991). An attack on the rear of the animal that excites the LG will not excite the MG, whereas an attack on the front of the animal that excites the MG will not excite LG (Liu & Herberholz, 2010). Walking inhibits LG, and both LG and MG inhibit the abdominal postural motor neurons (Kuwada, Hagiwara, & Wine, 1980). NG escapes are inhibited when the animal is feeding on non-portable food, but become highly excitable as soon as the crayfish senses that the food is portable (Bellman & Krasne, 1983). The crayfish makes similar neuroeconomic calculations during foraging. When following a food odor trail towards its source, a crayfish that encounters a moving shadow will choose between escape and freezing, an alternative predator-avoidance response (Liden, Phillips, & Herberholz, 2010). When the shadow moved slowly, crayfish were more likely to tailflip, whereas they were more likely to freeze when the shadow moved quickly. If the food odor was strengthened, the balance was shifted in favor of freezing, which was less disruptive of the animal’s progress towards the food.
In the 19th century, Huxley’s “The Crayfish” provided a useful perspective on the entire field of zoology through the lens of a common animal that was known to many people, but familiar to only a few. In a similar manner, the study of crayfish escape in the 20th and 21st centuries has provided many useful insights into how a nervous system is organized to respond to attack in the context of all its other ongoing sensory, motor, and regulatory processes, and in the larger environmental and social context in which it lives. Many important questions about crayfish escape and its neurobiological mechanisms remain unanswered. These include: how is the trajectory of a NG escape planned before an expected attack; what is the circuit mechanism for choosing between freezing and triggering an MG escape in response to a looming shadow; how is descending inhibition generated to modulate LG excitability and response habituation; how is inhibition and excitation among and between the circuits for escape and other crayfish behaviors organized to mediate adaptive, context-sensitive behavioral choice; how do visual, tactile, and other sensory inputs help guide swimming behavior during escape? Answers to these and other questions about escape will continue to extend our insights into the neurobiological mechanisms that evolution has provided for animals to respond to their environmental and social challenges.
- Antonsen, B. L. (2008, November). Interactions between serotonergic and dopaminergic modulation in a reflex escape circuit. Paper presented at the meeting of Neuroscience, Washington, DC.
- Antonsen, B. L., & Edwards, D. H. (2003). Differential dye coupling reveals lateral giant escape circuit in crayfish. Journal of Comparative Neurology, 466(1), 1–13.
- Antonsen, B. L., & Edwards, D. H. (2007). Mechanisms of serotonergic facilitation of a command neuron. Journal of Neurophysiology, 98(6), 3494–3504.
- Antonsen, B. L. Herberholz, J., & Edwards, D. H. (2005). The retrograde spread of synaptic potentials and recruitment of presynaptic inputs. Journal of Neuroscience, 25(13), 3086–3094.
- Antonsen, B. L., & Paul, D. H. (1997). Serotonin and octopamine elicit stereotypical agonistic behaviors in the squat lobsters Munida quadrispina (Anomura, Galatheidae). Journal of Comparative Physiology A, 181(501), 510.
- Araki, M., & Nagayama, T. (2003). Direct chemically mediated synaptic transmission from mechanosensory afferents contributes to habituation of crayfish lateral giant escape reaction. Journal of Comparative Physiology A, 189(10), 731–739.
- Araki, M., & Nagayama, T. (2012). IP(3)-mediated octopamine-induced synaptic enhancement of crayfish LG neurons. Journal of Comparative Physiology A, 98(8), 607–615.
- Araki, M., Nagayama, T., & Sprayberry, J. (2005). Cyclic AMP mediates serotonin-induced synaptic enhancement of lateral giant interneuron of the crayfish. Journal of Neurophysiology, 94(4), 2644–2652.
- Bellman, K. L., & Krasne, F. B. (1983). Adaptive complexity of interactions between feeding and escape in crayfish. Science, 221, 779–781.
- Bovbjerg, R. V. (1953). Dominance order in the crayfish Oconectes virilis (Hagen). Physiological Zoology, 26(2), 173–178.
- Bruner, J., & Kennedy, D. (1970). Habituation: Occurrence at a neuromuscular junction. Science, 169(3,940), 92–94.
- Bryan, J. S., & Krasne, F. B. (1977a). Presynaptic inhibition: The mechanism of protection from habituation of the lateral giant fibre escape response. The Journal of Physiology, 271(2), 369–390.
- Bryan, J. S., & Krasne, F. B. (1977b). Protection from habituation of the crayfish lateral giant fibre escape response. The Journal of Physiology, 271(2), 351–368.
- Cattaert, D., Delbecque, J. P., Edwards, D. H., & Issa F. A. (2010, November). Social interactions determine postural network sensitivity to 5-HT. Paper presented at the meeting of The Society of Neuroscience, San Diego, CA.
- Cofer, D., Cymbalyuk, G., Reid, J., Zhu, Y., Heitler, W. J., & Edwards, D. H. (2010). AnimatLab: A 3D graphics environment for neuromechanical simulations. Journal of Neuroscience Methods, 187(2), 280–288.
- Cooke, I. R. C. (1985). Further studies of crayfish escape behaviour: II. Giant axon-mediated neural activity in the appendages. Journal of Experimental Biology, 118(1), 367–377.
- Cooke, I. R. C., & Macmillan, D. L. (1985). Further studies of crayfish escape behaviour: I. The role of the appendages and the stereotyped nature of non-giant escape swimming. Journal of Experimental Biology, 118(1), 351–365.
- Dumont, J. P. C., & Wine, J. J. (1987). The telson flexor neuromuscular system of the crayfish: III. The role of feedforward inhibition in shaping a stereotyped behaviour pattern. Journal of Experimental Biology, 127(1), 295–311.
- Eccles, J. C., Kostyuk, P. G., & Schmidt, R. F. (1962). Presynaptic inhibition of the central actions of flexor reflex afferents. The Journal of Physiology, 161, 258–281.
- Edwards, D. H. (1991). Mutual inhibition among neural command systems as a possible mechanism for behavioral choice in crayfish. Journal of Neuroscience, 11(5), 1210–1223.
- Edwards, D. H. (2009). Excitation and habituation of crayfish escape. Journal of Experimental Biology, 212(6), 749–751.
- Edwards, D. H., Fricke, R. A., Barnett, L. D., Yeh, S. R., & Leise, E. M. (1994). The onset of response habituation during the growth of the lateral giant neuron of crayfish. Journal of Neurophysiology, 72(2), 890–898.
- Fraser, K., & Heitler, W. J. (1989). Thoracic output of crayfish giant fibers: II. The segmental giant neurone. Journal of Comparative Physiology A, 166(1), 125–132.
- Fraser, K., & Heitler, W. J. (1991). Photoinactivation of the crayfish segmental giant neuron reveals a direct giant-fiber to fast-flexor connection with a chemical component. Journal of Neuroscience, 11(1), 59–71.
- Edwards, D. H., Heitler, W. J., & Krasne, F. B. (1999). Fifty years of a command neuron: The neurobiology of escape behavior in the crayfish. Trends in Neurosciences, 22(4), 153–161.
- Edwards, D. H., Heitler, W. J., Leise, E. M., & Fricke, R. A. (1991). Postsynaptic modulation of rectifying electrical synaptic inputs to the LG escape command neuron in crayfish. Journal of Neuroscience, 11(7), 2117–2129.
- Edwards, D. H., & Mulloney, B. (1987). Synaptic integration in excitatory and inhibitory crayfish motoneurons. Journal of Neurophysiology, 57(5), 1425–1445.
- Edwards, D. H., & Spitzer, N. (2006). Social dominance and serotonin receptor genes in crayfish. Current Topics in Developmental Biology, 74, 177–199.
- Edwards, D. H., Yeh, S. R., Barnett, L. D., & Nagappan, P. R. (1994). Changes in synaptic integration during the growth of the lateral giant neuron of crayfish. Journal of Neurophysiology, 72(2), 899–908.
- Edwards, D. H., Yeh, S. R., & Krasne, F. B. (1998). Neuronal coincidence detection by voltage-sensitive electrical synapses. Proceedings of the National Academy of Sciences, 95(12), 7145–7150.
- Edwards, D. H., Yeh, S. R, Musolf, B. E., Antonsen, B. L., & Krasne, F. B. (2002). Metamodulation of the crayfish escape circuit. Brain, Behavior, and Evolution, 60(6), 360–369.
- Fricke, R. A. (1984). Development of habituation in the crayfish due to selective weakening of electrical synapses. Brain Research, 322(1), 139–143.
- Furshpan, E. J., & Potter, D. D. (1959). Transmission at the giant motor synapses of the crayfish. The Journal of Physiology, 145(2), 289–325.
- Giaume, C., Kado, R. T., & Korn, H. (1987). Voltage-clamp analysis of a crayfish rectifying synapse. The Journal of Physiology, 386, 91–112.
- Glantz, R. M., & Viancour, T. (1983). Integrative properties of crayfish medial giant neuron: Steady-state model. Journal of Neurophysiology, 50(5), 1122–1142.
- Glanzman, D. L., & Krasne, F. B. (1983). Serotonin and octopamine have opposite modulatory effects on the crayfish’s lateral giant escape reaction. Journal of Neuroscience, 3(11), 2263–2269.
- Günther, J. (1972). Giant motor neurons in the earthworm. Comparative Biochemistry and Physiology Part A: Physiology, 42(4), 967–973.
- Heitler, W. J, & Fraser, K. (1989). Thoracic output of crayfish giant fibers. I. Pereiopod promotor motor neurones. Journal of Comparative Physiology A, 166(1), 117–124.
- Heitler, W. J., & Fraser, K. (1993). Thoracic connections between crayfish giant fibres and motor giant neurones reverse abdominal pattern. Journal of Experimental Biology, 181(1), 329–333.
- Heitler, W. J., Fraser, K., & Edwards, D. H. (1991). Different types of rectification at electrical synapses made by a single crayfish neurone investigated experimentally and by computer simulation. Journal of Comparative Physiology A, 169(6), 707–718.
- Herberholz, J., Antonsen, B. L., & Edwards, D. H. (2002). A lateral excitatory network in the escape circuit of crayfish. Journal of Neuroscience, 22(20), 9078–9085.
- Herberholz, J., & Edwards, D. H. (2005, November). The control of escape in crayfish through interactions of command neurons. Paper presented at the meeting of the Society for Neuroscience, Washington, DC.
- Herberholz, J., Issa, F. A., & Edwards, D. H. (2001). Patterns of neural circuit activation and behavior during dominance hierarchy formation in freely behaving crayfish. Journal of Neuroscience, 21(8), 2759–2767.
- Herberholz, J., Sen, M. M., & Edwards, D. H. (2004). Escape behavior and escape circuit activation in juvenile crayfish during prey-predator interactions. Journal of Experimental Biology, 207(11), 1,855–1,863.
- Hill, A. A., Edwards, D. H., & Murphey, R. K. (1994). The effect of neuronal growth on synaptic integration. Journal of Computational Neuroscience, 1(3), 239–254.
- Holdich, D. M. (2002). The biology of freshwater crayfish. Oxford: Blackwell Science.
- Huxley, T. H. (1880). The crayfish: An introduction to zoology. New York: D. Appleton & Co.
- Issa, F. A., Adamson, D. J., & Edwards, D. H. (1999). Dominance hierarchy formation in juvenile crayfish, Procambarus clarkii. Journal of Experimental Biology, 202(24), 3497–3506.
- Issa, F. A., Drummond, J., Cattaert, D., & Edwards, D. H. (2012). Neural circuit reconfiguration by social status. Journal of Neuroscience, 32(16), 5638–5645.
- Kennedy, D., Calabrese, R. L., & Wine, J. J. (1974). Presynaptic inhibition: Primary afferent depolarization in crayfish neurons. Science, 186(4162), 451–454.
- Kennedy, D., Fricke, R. A., & Block, G. D. (1982). Inhibition of mechanosensory neurons in the crayfish: II. Inhibition associated with proprioceptive feedback from locomotion. Journal of Comparative Physiology A, 149(2), 251–262.
- Kirk, M. D., Dumont, J. P., & Wine, J. J. (1986). Local inhibitor of the crayfish telson-flexor motor giant neurons: Morphology and physiology. Journal of Comparative Physiology A, 158(1), 69–79.
- Kramer, A. P., Krasne, F. B., & Wine, J. J. (1981). Interneurons between giant axons and motoneurons in crayfish escape circuitry. Journal of Neurophysiology, 45(3), 550–573.
- Krasne, F. B. (1969). Excitation and habituation of the crayfish escape reflex: The depolarizing response in lateral giant fibres of the isolated abdomen. Journal of Experimental Biology, 50(1), 29–46.
- Krasne, F. B., & Lee, S. C. (1988). Response-dedicated trigger neurons as control points for behavioral actions: Selective inhibition of lateral giant command neurons during feeding in crayfish. Journal of Neuroscience, 8(10), 3703–3712.
- Krasne, F. B., Shamsian, A., & Kulkarni, R. (1997). Altered excitability of the crayfish lateral giant escape reflex during agonistic encounters. Journal of Neuroscience, 17(2), 709–716.
- Krasne, F. B., & Teshiba, T. M. (1995). Habituation of an invertebrate escape reflex due to modulation by higher centers rather than local events. Proceedings of the National Academy of Sciences, 92(8), 3362–3366.
- Krasne, F. B., & Wine, J. J. (1975). Extrinsic modulation of crayfish escape behavior. Journal of Experimental Biology, 63(2), 4333–4450.
- Krasne, F. B., & Wine, J. J. (1984). The production of crayfish tailflip escape responses. In R. C. Eaton (Ed.), Neural mechanisms of startle behavior (pp. 179–211). New York: Plenum Press.
- Kravitz, E. A. (1988). Hormonal control of behavior: Amines and the biasing of behavioral output in lobsters. Science, 241(4,874), 1775–1781.
- Kupfermann, I., & Weiss, K. R. (1978). The command neuron concept. Behavioral and Brain Sciences, 1, 3–39.
- Kuwada, J. Y., Hagiwara, G., & Wine, J. J. (1980). Postsynaptic inhibition of crayfish tonic flexor motor neurones by escape commands. Journal of Experimental Biology, 85(1), 343–347.
- Lee, S. H., Taylor, K., & Krasne, F. B. (2008). Reciprocal stimulation of decay between serotonergic facilitation and depression of synaptic transmission. Journal of Neurophysiology, 100(2), 1113–1126.
- Liden, W. H., Phillips, M. L., & Herberholz, J. (2010). Neural control of behavioural choice in juvenile crayfish. Proceedings of the Royal Society B: Biological Sciences, 277(1699), 3493–3500.
- Liu, Y. C., & Herberholz, J. (2010). Sensory activation and receptive field organization of the lateral giant escape neurons in crayfish. Journal of Neurophysiology, 104(2), 675–684.
- Livingstone, M. S., Harris-Warrick, R. M., & Kravitz, E. A. (1980). Serotonin and octopamine produce opposite postures in lobsters. Science, 208(4439), 76–79.
- Miller, L. A., Hagiwara, G., & Wine, J. J. (1985). Segmental differences in pathways between crayfish giant axons and fast flexor motoneurons. Journal of Neurophysiology, 53(1), 252–265.
- Miller, M. W., Vu, E. T., & Krasne, F. B. (1992). Cholinergic transmission at the first synapse of the circuit mediating the crayfish lateral giant escape reaction. Journal of Neurophysiology, 68(6), 2174–2184.
- Mittenthal, J. E., & Wine, J. J. (1973). Connectivity patterns of crayfish giant interneurons: visualization of synaptic regions with cobalt dye. Science, 179(4069), 182–184.
- Newland, P. L., Aonuma, H., & Nagayama, T. (1997). Monosynaptic excitation of lateral giant fibres by proprioceptive afferents in the crayfish. Journal of Comparative Physiology A, 181(2), 103–109.
- Olson, G. C., & Krasne, F. B (1981). The crayfish lateral giants as command neurons for escape behavior. Brain Research, 214(1), 89–100.
- Rayner, M. D., & Wiersma, C. A. G. (1967). Mechanisms of crayfish tail flick. Nature, 213, 1231–1232.
- Reichert, H., & Wine, J. J. (1983). Coordination of lateral giant and non-giant systems in crayfish escape behavior. Journal of Comparative Physiology A, 153(1), 3–15.
- Reichert, H., Wine, J. J., & Hagiwara, G. (1981). Crayfish escape behavior: Neurobehavioral analysis of phasic extension reveals dual systems for motor control. Journal of Comparative Physiology A, 142(3), 281–294.
- Roberts, A. (1968). Recurrent inhibition in the giant-fibre system of the crayfish and its effect on the excitability of the escape response. Journal of Experimental Biology, 48(3), 545–567.
- Roberts, A., Krasne, F. B., Hagiwara, G., Wine, J. J., & Kramer, A. P. (1982). Segmental giant: Evidence for a driver neuron interposed between command and motor neurons in the crayfish escape system. Journal of Neurophysiology, 47(5), 761–781.
- Shirinyan, D., Teshiba, T., Taylor, K., O’Neill, P., Lee, S. C., & Krasne, F. B. (2006). Rostral ganglia are required for induction but not expression of crayfish escape reflex habituation: Role of higher centers in reprogramming low-level circuits. Journal of Neurophysiology, 95(4), 2721–2724.
- Sillar, K. T., Picton, L. D., & Heitler, W. J. (2016). The neuroethology of predation and escape. Chichester, England: John Wiley & Sons.
- Silver, R. A. (2010). Neuronal arithmetic. Nature Reviews Neuroscience, 11(7), 474–489.
- Spitzer, N., Edwards, D. H., & Baro, D. J. (2008). Conservation of structure, signaling and pharmacology between two serotonin receptor subtypes from decapod crustaceans, Panulirus interruptus and Procambarus clarkii. Journal of Experimental Biology, 211(Pt 1), 92–105.
- Teshiba, T., Shamsian, A., Yashar, B., Yeh, S. R., Edwards, D. H., & Krasne, F. B. (2001). Dual and opposing modulatory effects of serotonin on crayfish lateral giant escape command neurons. Journal of Neuroscience, 21(12), 4523–4529.
- Vu, E. T., & Krasne, F. B. (1992). Evidence for a computational distinction between proximal and distal neuronal inhibition. Science, 255, 1710–1712.
- Vu, E. T., Berkowitz, A., & Krasne, F. B. (1997). Postexcitatory inhibition of the crayfish lateral giant neuron: A mechanism for sensory temporal filtering. Journal of Neuroscience, 17(22), 8867–8879.
- Vu, E. T., & Krasne, F. B. (1993). Crayfish tonic inhibition: Prolonged modulation of behavioral excitability by classical GABAergic inhibition. Journal of Neuroscience, 13(10), 4394–4402.
- Vu, E. T., & Krasne, F. B. (1999). Evidence for a computational distinction between proximal and distal neuronal inhibition. Science, 255(5052), 1710–1712.
- Vu, E. T., Lee, S. C., & Krasne, F. B. (1993). The mechanism of tonic inhibition of crayfish escape behavior: distal inhibition and its functional significance. Journal of Neuroscience, 13(10), 4379–4393.
- Whitaker, K. W., Neumeister, H., Huffman, L. S., Kidd, C. E., Preuss, T., & Hofmann, H. A. (2011). Serotonergic modulation of startle-escape plasticity in an African cichlid fish: A single-cell molecular and physiological analysis of a vital neural circuit. Journal of Neurophysiology, 106(1), 127–137.
- Wiersma, C. A. G. (1947). Giant nerve fiber system of the crayfish: A contribution to comparative physiology of synapse. Journal of Neurophysiology, 10(1), 23–38.
- Wine, J. J., & Krasne, F. B. (1972). The organization of escape behavior in the crayfish. Journal of Experimental Biology, 56(1), 1–18.
- Wine, J. J. (1977a). Crayfish escape behavior: II. Command-derived inhibition of abdominal extension. Journal of Comparative Physiology A, 121(2), 173–185.
- Wine, J. J. (1977b). Crayfish escape behavior: III. Monosynaptic and polysynaptic sensory pathways involved in phasic extension. Journal of Comparative Physiology A, 121(2), 187–203.
- Wine, J. J. (1984). The structural basis of an innate behavioural pattern. Journal of Experimental Biology, 112(1), 283–319.
- Wine, J. J., & Mistick, D. C. (1977). Temporal organization of crayfish escape behavior: Delayed recruitment of peripheral inhibition. Journal of Neurophysiology, 40(4), 904–925.
- Wine, J. J., & Krasne, F. B. (1982). The cellular organization of crayfish escape behavior. In D. C. Sandeman, & H. L. Atwood (Eds.), The biology of Crustacea: Neural integration and behavior (pp. 241–292). New York: Academic Press.
- Yeh, S. R., Fricke, R. A., & Edwards, D. H. (1996). The effect of social experience on serotonergic modulation of the escape circuit of crayfish. Science, 271(5247), 366–369.
- Yeh, S. R., Musolf, B. E., & Edwards, D. H. (1997). Neuronal adaptations to changes in the social dominance status of crayfish. Journal of Neuroscience, 17(2), 697–708.
- Zucker, R. S. (1972a). Crayfish escape behavior and central synapses: I. Neural circuit exciting lateral giant fiber. Journal of Neurophysiology, 35(5), 599–620.
- Zucker, R. S. (1972b). Crayfish escape behavior and central synapses: II. Physiological mechanisms underlying behavioral habituation. Journal of Neurophysiology, 35(5), 621–637.