Role of Serotonin in Tritonia diomedea, The

role of Serotonin in Tritonia diomedea, The

Brown, Glen D

The Role of Serotonin in Tritonia diomedea1

SYNOPSIS. The within-swim pattern of cycle periods in Tritonia swimming changed when the behavior was repeatedly elicited suggesting that an excitatory process reaches a ceiling or wanes over repeated trials. Exposure to subthreshold stimuli enhanced swimming in response to a subsequent super-threshold stimulus, perhaps using a similar excitatory process. In reduced preparations, subthreshold stimuli increased action potential activity in identified serotonergic neurons. Finally, stimulating serotonergic neurons enhanced a fictive swimming pattern, much like subthreshold stimuli enhanced the swimming behavior. Both within-swim and across-swim changes in the swimming behavior may be caused by increased activity in identified serotonergic neurons. Comparative study suggests that ancestral serotonergic systems facilitated network oscillations for the production of rhythmic behaviors such as feeding and locomotion. This concept of serotonin as oscillatizer is used to explain the role of serotonergic neurons in Tritonia. Implications for human mental health are discussed.


Tritonia diomedea swimming has been studied for more than three decades as a model for understanding the neural basis of rhythmic behavior (Willows, 1967; see Getting, 1989 for review). For almost that long, there has been an interest in the neural basis of simple types of experience-dependent changes in the swimming response (Abraham and Willows, 1971; Brown, 1998). In the present study, a short-term sensitization process that enhances certain aspects of the swimming behavior will be reconsidered in behavioral and cellular experiments with a focus on an identified serotonergic system.

In previous behavioral studies, when the Tritonia swimming response was repeatedly elicited, the latency from stimulus application to swimming was reduced to about half its original length on the second application of the stimulus (Abraham and Willows, 1971; Brown et al., 1996). Swimming threshold was also lowered by prior exposure to swim-eliciting stimuli. When a weak stimulus (tap water) and a strong stimulus (supersaturated saline) were presented alternately, the tap water was more likely to elicit swimming than when it was presented alone.

In another study, swimming latency and threshold were also enhanced by experience with a swim-eliciting stimulus applied to one site five minutes before testing at another site suggesting a heterosynaptic mechanism (Brown, 1994b). The length of the swim (cycle number) and the cycle period were also enhanced in this study. No significant change in cycle period was found when this experiment was repeated (Frost et al., 1998), but the trend in the latter study together with the earlier data suggests that cycle period can be enhanced somewhat by a prior stimulus. The same excitatory process also appears to temporarily reverse long-term habituation (Brown, 1998). Thus, what appears to be a unitary sensitization process produces multiple short-term enhancements in Tritonia swimming.

Short-term decrements in the swimming behavior also occur in repeated-trials experiments. Swimming cycle number, latency, cycle period, and flexion speed all showed some type decrement when a swim– eliciting stimulus was presented repeatedly at short intervals (Abraham and Willows, 1971; Brown et al., 1996). These short-term decrements appeared to be due to a non– adaptive limitation in the swimming network rather than habituation (Brown, 1998; see Discussion). (Habituation and a non-associative inhibitory process, can also produce decrements in the swimming behavior but will not be considered here.) Both the behavioral enhancements associated with sensitization and the short-term decrements are present by the second trial in repeated– trials experiments, and both recover in about an hour.

Excitatory and inhibitory processes are also apparent during a single swimming episode (Hume et al., 1982). Thus, we examined the pattern of cycle periods within swims for each swim in a repeated trials experiment (Fig. 1A). Synaptic depression and heterosynaptic facilitation at one type of synapses may account for both within– swim and across-swim changes in behavior (see Discussion).

Increased activity in the serotonergic dorsal swim interneurons (DSI) has been proposed as a possible memory trace for sensitization in Tritonia (Brown, 1994b; McClellan et al., 1994). In the second part of our study, the link between sensitization and DSI activity is explored further using subthreshold stimuli (Figs. 1B, C, 2). Finally, I consider evidence suggesting that a primitive use of serotonergic neurons was to reconfigure networks to produce oscillations (see also Jacobs and Fornal, 1999), a function that may be retained in modern animals such as Tritonia and humans.


Procedures used in behavioral experiments have been described in detail elsewhere (Brown et al., 1996; Brown, 1998). Briefly, slugs are isolated for at least three hours before experiments begin. Salt solutions used as stimuli were tinted to improve visibility and delivered through a hand-held syringe.

The “head” stimulus (0.15 ml of 1.5 ml NaCl, applied to the oral veil and epithelium anterior to the rhinophores) was subthreshold for swimming. The “tail” stimulus (0.15 ml of 3 M NaCl except as noted, applied to the triangle made up of the posterior set of gill tufts-left and right-and the posterior tip of the animal) always elicited swimming in the intact animal.

In head and tail preparations (see also Brown, 1994a) dissection began by cutting away the dorsal epithelium just above the central nervous system. Cuts were made around the longitudinal axis so that the body wall rostral to the brain was severed from the caudal portion. In head preparations, the esophagus and other internal organs were cut away posterior to the brain so that all posterior body wall nerves were severed. In tail preparations, the anterior part of the slug was dissected away in much the same way.

Intact nerves often pulled the brain into the remaining tissue so that it was usually necessary to dissect away more of the body wall and internal organs. Care was taken not to cut the remaining nerves.

In both preparations, the ventral surface was dissected away to remove cilia and immobilize the preparation. The brain was pinned out on a raised platform of Sylgard. A barrier of dissecting pins was erected between the brain and the remaining tissue to isolate the brain from movements that occurred during fictive swimming.

To test whether a fictive swimming pattern could be generated after dissection, we recorded extracellularly from the cut end of a nerve root and intracellularly from one or more flexion neuron in the pedal ganglion. The remaining skin was then stimulated with a strong saline solution. If the pattern was generated, then the intracellular electrode was removed and the preparation was rested for three hours. Otherwise, the preparation was discarded. Only about half of the preparations were viable by this criterion.


Repeated trials

Twenty slugs were filmed while receiving 10 swim-eliciting tail stimuli at a 2 min inter-trial interval (see Brown et al., 1996). The period of the first three cycles for each swim was measured from these films. All three cycles were significantly longer by the tenth trial as compared to the first (see Fig. 1A).

The pattern of cycle periods was also different by the end of the experiment (Fig. 1A). On the first trial, the second period was slightly shorter than the first, and cycle period lengthened thereafter as expected (Hume et al., 1982). By the tenth stimulus however, the first cycle was the shortest. Thus, the excitatory process that shortens the cycle period either reaches a ceiling or wanes after repeated trials.

The four components of cycle period, ventral flexion (VF), ventral pause (VP), dorsal flexion (DF), and dorsal pause (DP) were also measured for the first cycle of swimming on each trial. The dorsal-ventral duty cycle, (DF + DP)/(VF + VP), showed a small but statistically significant change over trials from 0.98 +/- 0.03 on the first trial to 1.06 +/- 0.02 on the last trial (t = 2.54, P

Subthreshold stimuli

Five subthreshold head stimuli were presented to each of eight slugs at five minute intervals. A tail stimulus followed ten minutes later to elicit swimming. Swimming cycle number in this group was significantly enhanced compared to eight slugs receiving only the tail stimulus (Fig. 1B; 9.1 +/- 0.7; 6.4 +/- 0.4, P

DSI activity

In head preparations, we placed intracellular recording electrodes in the DSI and stimulated the remaining tissue with subthreshold stimuli. The head stimulus that enhanced the response (Fig. 1B) also raised the baseline firing rate of DSI in head preparations (Fig. 2; n = 10).

In tail preparations, intracellular electrodes were placed in one DSI and one to three flexion neurons. After tossing a coin, the DSI was either stimulated with tonic current through the intracellular electrode or not. After two minutes of DSI stimulation and a five second delay or after a two– minute and five second delay, respectively, the remaining tissue was stimulated with a tail stimulus. Stimulating a DSI increased the number of fictive swimming cycles recorded in the flexion neurons (Fig. 1C; DSI stimulation: 9.0 +/- 0.6, n = 3; DSI recording: 5.2 +/- 0.6, n = 5; P


C2 synapses and behavior

The changing strength of synapses made from swimming interneuron cerebral cell 2 (C2) onto its many followers may explain some of the variability found both within and across swimming episodes in Tritonia. C2 synaptic strength is known to be reduced by a use-dependent depression (Snow, 1982). Although C2 makes both excitatory and inhibitory synapses, only the fast excitatory synapses are prone to depression. Depression at C2 synapses may produce the progressive lengthening of cycle period during swimming (Fig. 1A; Hume et al., 1982) and contribute to the eventual termination of a swimming episode.

If the recovery from depression takes longer than the inter-trial interval in a repeated-trials experiment, then this would also limit the duration of the response and increase the cycle period in an abnormal way across trials. Depression appears to recover enough in the two-minute interval between stimuli to support short swimming episodes almost indefinitely and recovers fully after one hour (Brown, 1997, 1998).

Observations of Tritonia swimming in the wild have not been reported in the literature, but the short-term reduction in cycle number across trials does not appear to be an adaptive change in behavior (Brown, 1998). Based on my own experience and communications with other divers who have observed Tritonia in their natural habitat, the time between swim-eliciting stimuli appears to be on the order of many minutes at least and more likely hours or days. Thus, the two minute interval used in many repeated-trials experiments (e.g., Fig. 1A) does not appear to be realistic biologically, and C2 synaptic depression is unlikely to affect slugs across swimming episodes in nature.

Besides synaptic depression, a facilitation of C2 synapses is present after the first activation of the fictive swimming pattern in the isolated brain preparation (Brown, 1994a, b). Stimulation of a single DSI also strengthened C2 synapses in these studies much as it enhanced the fictive pattern overall in tail preparations (Fig. 1C). This synaptic facilitation, which lasts for only a few seconds, appears to be due to direct effects of DSI on C2 presynaptic terminals (Katz and Frost, 1997). Like C2 and other Tritonia central pattern generator (CPG) interneurons, DSI fire discrete bursts of action potentials in an oscillating manner during swimming (Getting, 1977). Bursts of action potentials in C2 overlap with but are slightly phase lagged from bursts in the DSI, which should allow for the facilitation of C2 synapses on a cycle by cycle basis.

DSI fire at elevated rates after the termination of fictive swimming or in response to subthreshold stimuli in reduced preparations (Getting, 1977; Fig. 2). Facilitation at C2 synapses by this ongoing DSI activity may lead to the changes associated with sensitization (Fig. 1B; Brown, 1994b, 1998). According to this model, C2 synapses should be strongest on the second cycle of the first swimming episode in repeated trials experiments (see Fig. 1A). After many trials, the within-swim enhancement of cycle period disappears so the additional assumption that the facilitation of C2 synapses reaches a ceiling or even wanes must be made.

This model makes a number of other predictions. For example, if the DSI were stimulated at a high rate before swimming in naive slugs, C2 synapses should be facilitated, and the first cycle should be the shortest. Similarly, subthreshold stimuli could help to reveal enhancement of the first cycle period. In general, the length of the swim and the cycle period should always correlate with the strength of C2 synapses.

However, this model is overly simplified. First, C2 itself has shown at least three other kinds of plasticity, including a use-dependent facilitation, a depolarization-induced facilitation, and a change in excitability (Snow, 1982; Katz and Frost, 1997). Second, behavioral and cellular studies suggest that multiple types of modifications take place in the afferent pathway during repeated trials experiments, including changes in the central and peripheral nervous systems (Getting, 1976; Snow, 1982; Brown, 1994a, b, 1998). Third, DSI activity has other effects on the swimming network besides facilitation of C2 synapses (McClellan et al., 1994; Brown, 1994b). The DSI may even facilitate other behaviors besides swimming (discussed below). Finally, increased activity in many neurons besides DSI correlates with sensitization (Brown, 1994a, b). Thus, there are likely to be multiple modifications in the swimming network that tend to either enhance or inhibit the swimming behavior.

Nevertheless, C2 and its synapses provide a conceptually attractive place to influence the response since C2 appear to be specialized for swimming (Taghert and Willows, 1978). A two factor theory such as the balance between depression and facilitation at synapses made by one pair of neurons may at least partially explain a number of behavioral observations.

Serotonin as oscillatizer

The cellular model of sensitization discussed above starts with increased DSI activity (Fig. 2), which produces a number of modifications in the swimming network including the facilitation of C2 synapses (Brown, 1994b). Multiple changes in the network then lead to enhancements in multiple components of the swimming behavior (e.g., Fig. 1B).

DSI appear to use serotonin as a neurotransmitter (McClellan et al., 1994), and serotonin has been associated with sensitization in other species (Catarsi et al., 1990; Aggio et al., 1996; Weiger, 1997). In another opisthobranch, Aplysia californica, increased activity in identified serotonergic neurons has been proposed as a mechanism for short-term sensitization (Mackey et al., 1989). However, single-unit recordings from the cat suggest that serotonergic activity does not correlate well with alertness or sensitization in vertebrates (Jacobs and Fornal, 1997, 1999).

Instead of-or in addition to-sensitization, increased DSI activity (Fig. 2) may facilitate other Tritonia behaviors besides swimming. Activity in serotonergic neurons plays a key role in motor-pattern selection in diverse phyla (Jacobs and Fornal, 1997; Weiger, 1997). In particular, the activity of serotonergic neurons correlates with or is known to facilitate CPG activity in a number of species (Jacobs and Fornal, 1999; Weiger, 1997). Different groups of serotonergic neurons or different firing rates may configure the same network for different kinds of oscillations.

In Tritonia, steady firing (beating) modes correlate well with the expression of another rhythmic behavior, crawling, which is already known to be influenced by serotonergic neurons (Audesirk et al., 1979). In cats, serotonergic neurons also have steady beating and rhythmic bursting modes, both of which correlate with rhythmic motor patterns such as feeding and locomotion (Jacobs and Formal, 1999). Thus, even serotonergic activity that is not itself oscillatory often appears to facilitate oscillatory behaviors.

I speculate that a primitive function of serotonergic neurons was to facilitate oscillations in networks of neurons. The production of a rhythmic output may be one of the oldest functions of the nervous system since oscillations are important for feeding, locomotion, and biological clocks. Primitive serotonergic systems may have configured ancestral nervous systems to produce the neural oscillations necessary for these behaviors.

The view of serotonin as oscillatizer reconciles the roles of DSI as mediator and modulator of Tritonia swimming (Katz and Frost, 1995; Weiger, 1997). DSI oscillatizes the Tritonia swimming network directly– but in a minor way-as a bursting neuron using serotonin as a classical neurotransmitter and indirectly-in an essential way-as a neuromodulator. Serotonin may help reconfigure the swimming network from a feed-forward reflexive circuit into the recurrent oscillator used for swimming (see Getting, 1989).

Making the synapses of a key neuron stronger, as in the case of DSI influencing C2 connections, may be one way serotonin oscillatizes a network. C2 has recurrent connections with other neurons in the network, and one possibility is that C2 provides feedback information to the rest of the network. C2 does not appear to fire except during swimming though this issue has not been explored exhaustively (Taghert and Willows, 1978).

DSI activity or the concentration of serotonin is not postulated to control or determine behavior in Tritonia or elsewhere. Serotonin is not necessarily a switch to turn on oscillations. There may be no such switch, no controller inside the controller, no slugunculus. Many factors inside and outside the animal determine whether a certain behavior like swimming will be expressed. Serotonin appears to ready networks of neurons to produce oscillations. How and when oscillations occur is decided by the nervous system as a whole, or rather by the entire action-perception cycle of which the nervous system is an important part.

Serotonin seems to use many different mechanisms and to shape all parts of a network for oscillations. Serotonergic modulation even extends to the level of the muscles and other effectors used in rhythmic behaviors in Tritonia and elsewhere (Audesirk et al., 1979; Satterlie and Norekian, 1996). Serotonergic neurons project widely in vertebrates, including the forebrain, midbrain, and spinal cord, and transplanted serotonergic neurons restored rhythmic motor activity in a mammal (Feraboli-Lohnherr et al., 1997). Serotonin shapes sensory as well as motor systems (see Mackey et al., 1989), which may be important, for example, to provide cycle-by-cycle feedback about rhythmic movements.

While increased activity in serotonergic neurons does not correlate well with reflex or ballistic behaviors, some non-rhythmic, tonic behaviors have been associated with activity in serotonergic neurons (Jacobs and Fornal, 1997, 1999). Although activation is stronger for rhythmic behaviors, this appears to be evidence against the proposed general theory of serotonergic function. However, even if muscles do not need to oscillate, other parts of the network may (e.g., Sugihara et al., 1995). Nevertheless, exceptions to the general role of oscillatizer for serotonergic neurons may exist even if serotonin was a primitive oscillatizer since serotonin may have been co-opted or co– evolved for other purposes.

In humans, the serotonin transporter gene has been implicated in obsessive-compulsive disorder (OCD). Repetitive thoughts and actions associated with OCD can be controlled with serotonin reuptake inhibitors (SRIs; Bengel et al., 1999; see Jacobs and Fornal, 1999). Indeed, repetitive, OCD– like symptoms associated with a variety of disorders such as autism and Tourette’s syndrome can be treated with an SRI (Hollander et al., 1999). It would be of interest to determine whether or not the subset of people with depression that can be treated with SRIs also have compulsive behaviors, repetitive thoughts, or other OCD-like symptoms. Thus, comparative studies of serotonin function may have implications for human mental health. For example, long walks may partially substitute for SRIs.

Another theory of primitive serotonergic function has serotonin as a neurohormone, where levels of serotonin determine certain fear-avoidance or escape behaviors (Weiger, 1997). This is not unlike the function proposed for serotonin in sensitization, and none of these theories are mutually exclusive. Many fight or flight responses require rhythmic movements, and indeed it is difficult to separate the behavior from the brain state. However, it seems logical that movement predates sensitization in evolutionary history.

Do we run because we are scared or are we scared because we run? Does a Tritonia crawl upstream for an hour after swimming to return to food and mates (Murray et al., 1992)? Is that why the DSI have a beating mode? Could at least part of the lowered threshold and reduced latency for swimming associated with sensitization also be part of the normal crawling state? Answering these questions may help us determine how the role of serotonin as oscillatizer and its role as neurohormone are related.


I thank Beth Collins for help with behavioral experiments. Barry Jacobs, Terry Sejnowski, and Myra Emerson provided useful comments and suggestions. These results appeared previously (Brown, 1994a). Supported by a predoctoral fellowship from the National Institutes of Health.

1 From the Symposium on Swimming in Opisthobranch Mollusks: Contributions to Control of Motor Behavior presented at the Annual Meeting of the Society for Integrative and Comparative Biology, 4-8 January 2000, at Atlanta, Georgia.


Abraham, E D. and A. 0. D. Willows. 1971. Plasticity of a fixed action pattern in the sea slug Tritonia diomedea. Comm. Behav. Bio. 6:271-280.

Aggio, J., A. Rakitin, and H. Maldonado. 1996. Serotonin-induced short- and long-term sensitization in the crab Chasmagnathus. Pharmacol. Biochem. Behav. 53:441-448.

Audesirk, G., R. E. McCaman, and A. 0. Willows. 1979. The role of serotonin in the control of pedal ciliary activity by identified neurons in Tritonia diomedea. Comp. Biochem. Physiol. C: Comp. Pharmacol. 62C:87-91.

Bengel, D., B. D. Greenberg, G. Cora-Locatelli, M. Altemus, A. Heils, Q. Li, and D. L. Murphy. 1999. Association of the serotonin transporter regulatory region polymorphism and obsessive-compulsive disorder. Molec. Psych. 4:463-466.

Brown, G. D. 1994a. The cellular analysis of nonassociative learning in Tritonia. Ph.D. Diss., The University of Iowa, Iowa City, Iowa.

Brown, G. D. 1994b. Heterostimic enhancement of a not so fixed action pattern. Netherlands J. Zool. 44:184-199.

Brown, G. D. 1997. Isolated-brain parallels to simple types of learning and memory in Tritonia. Physiol. & Behav. 62:509-518.

Brown, G. D. 1998. Nonassociative learning processes affecting swimming probability in the seaslug Tritonia diomedea: Habituation, sensitization, and inhibition. Behav. Brain Res. 95:151-165.

Brown, G. D., W. N. Frost, and P A. Getting. 1996. Habituation and iterative enhancement of multiple components of the Tritonia swim response. Behav. Neurosci. 110:478-485.

Catarsi, S., M. Garcia-Gil, G. Traina, and M. Brunelli. 1990. Seasonal variation of serotonin content and nonassociative learning of swim induction in the leech, Hirudo medicinalis. J. Comp. Physiol. A 167:469-474.

Feraboli-Lohnherr, D., D. Orsal, A. Yakovleff, Y. Gimenez, M. Ribotta, and A. Privat. 1997. Recovery of locomotor activity in the adult chronic spinal rat after sublesional transplantation of embryonic nervous cells: Specific role of serotonergic neurons. Exper. Brain Res. 113:443-454.

Frost, W. N., C. L. Brandon, and D. L. Mongeluzi. 1998. Sensitization of the Tritonia escape swim. 69:126-135.

Getting, P. A. 1976. Afferent neurons mediating escape swimming of the marine mollusc, Tritonia. J. Comp. Physiol. A 110:271-286.

Getting, P. A. 1977. Neuronal organization of escape

swimming in Tritonia. J. Comp. Physiol. 121: 325-342.

Getting, P A. 1979. A network oscillator underlying swimming in Tritonia. In J. Jacklet (ed.), Neuronal and cellular oscillators pp. 215-236. Marcel Dekker, New York.

Jacobs, B. L. and C. A. Fornal. 1997. Serotonin and motor activity. Curt Opin. in Neurobiol. 7:820825.

Jacobs, B. L. and C. A. Fornal. 1999. Activity of serotonergic neurons in behaving animals. Neuropsychopharmacology 21:9S-155.

Hollander, E., S. Novotny, A. Allen, B. Aronowitz, C. Cartwright, and C. DeCaria. 1999. The relationship between repetitive behaviors and growth hormone response to sumatriptan challenge in adult autistic disorder. Neurophychopharmacology 22: 163-167.

Hume, R. L, P. A. Getting, and M. A. del Beccaro. 1982. Motor organization of Tritonia swimming. I. Quantitative analysis of swim behavior and flexion neuron firing patterns. J. Neurophysiol. 47:6074.

Katz, P S. and W. N. Frost. 1995. Intrinsic neuromodulation in the Tritonia swim CPG: Serotonin mediates both neuromodulation and neurotransmission by the dorsal swim interneurons. J. Neurophysiol. 74:2281-2294.

Katz, P S. and W. N. Frost. 1997. Removal of spike frequency adaptation via neuromodulation intrinsic to the Tritonia escape swim central pattern generator. J. Neurosci. 17:7703-7713.

Mackey, S. L., E. R. Kandel, and R. D. Hawkins. 1989. Identified serotonergic neurons LCB-1 and RCB1 in the cerebral ganglia of Aplysia produce presynaptic facilitation of siphon sensory neurons. J. Neurosci. 9:4227-4235.

McClellan, A. D., G. D. Brown, and P A. Getting. 1994. Modulation of swimming in Tritonia: Excitatory and inhibitory effects of serotonin. J. Comp. Physiol. A 174:257-266.

Murray, J. A., R. S. Hewes, and A. 0. D. Willows. 1992. Water-flow sensitive pedal neurons in Tritonia: Role in rheotaxis. J. Comp. Physiol. A 171: 373-385.

Satterlie, R. A. and T P. Norekian. 1996. Modulation of swimming speed in the pteropod mollusc, Clione limacina: Role of a compartmental serotonergic system. Invertebr. Neurosci. 2:157-165.

Snow, R. W. 1982. Characterization of the synaptic actions of an interneuron in the central nervous system of Tritonia. J. Neurobiol. 13:251-266.

Sugihara, I., E. J. Lang, and R. Llinas. 1995. Serotonergic modulation of inferior olivary oscillations and synchronicity: Multiple-electrode study in the rat cerebellum. Eur. J. Neurosci. 7:521-534.

Taghert, P H. and A. 0. D. Willows. 1978. Control of a fixed action pattern by single, central neurons in the marine mollusk, Tritonia diomedea. J. Comp. Physiol. 123:253-259.

Weiger, W. A. 1997. Serotonergic modulation of behaviour: A phylogenetic overview. Biol. Rev. Camb. Philos. Soc. 72:61-95.

Willows, A. 0. D. 1967. Behavioral acts elicited by stimulation of single identifiable nerve cells. Science 157:570-574.


Computational Neurobiology Laboratory, The Salk Institute for Biological Studies, La Jolla, California 92037

2 E-mail:

Copyright Society for Integrative and Comparative Biology Sep 2001

Provided by ProQuest Information and Learning Company. All rights Reserved