From Scholarpedia
Kim Blackwell and Joseph Farley (2008), Scholarpedia, 3(7):4090. doi:10.4249/scholarpedia.4090 revision #124045 [link to/cite this article]
Jump to: navigation, search
Curator and Contributors

1.00 - Joseph Farley

Figure 1: Hermissenda, Friday Harbor Labs, San Juan Island, WA
Hermissenda is a sea slug used for research on the neurobiology of learning and memory. In a classical conditioning paradigm, it learns to associate light with vestibular stimulation. Changes to membrane properties are correlated with learning behavior. Because of its simple nervous system and classical conditioning behavior, it is a valuable animal model for studying the mechanisms used by neurons to store memories.


General Information

Physical Appearance

Hermissenda crassicornis is a brightly colored sea slug known as the opalescent nudibranch (Figure 1). It grows up to 9 cm in length. The dorsal surface of the body has a bright orange stripe down the midline. Flanking this orange stripe are two blues stripes. Two tentacles, with a continuation of the blue stripe, are located at the front. On the dorsal surface are two rhinophores located at the rostral portion, and multiple cerata containing nematocysts at the tips. The entire ventral surface of the body is comprised of muscle, and referred to as a “foot” because of its role in movement and adherence to a substrate. The base of the foot contains both mucous secreting cells and cilia, which together are involved in locomotion.

Scientific Classification

  • Kingdom: Animalia
  • Phylum: Mollusca
  • Class: Gastropoda
  • Order: Opisthobranchia
    • Suborder: Nudibranchia
    • Superfamily: Aeolidioidea
  • Family: Glaucidae
  • Genus: Hermissenda
  • Species: crassicornis


Hermissenda are found in intertidal and subtidal areas along the Pacific coast of North America, as far north as Alaska and as far south as Mexico. They are observed also along the Northwest Pacific, such as South Korea and Japan. They prefer hydroids as food, but also feed on bryozoans, sea pens, anemones, and colonial ascidians (sea squirts). Though their colouration warns most creatures away from stinging nematocyst loaded cerrata, they do have some predators, including mosshead warbonnet fish, and the opisthobranchs Navanax, and Pleurobranchaea.

Reproduction and Development

In both the laboratory and their natural environment, reproduction of Hermissenda occurs year round. Hermissenda are hermaphroditic, each organism having both male and female reproductive organs. Nonetheless, a single Hermissenda cannot fertilize itself - copulation requires two animals. Sometime after copulation, egg masses are laid. Animals hatch from large fertile egg masses within 5-6 days as free-swimming shelled larvae (veligers), and feed on plankton. The veliger stage is an obligatory one lasting at least 34 days. This extended veliger stage, during which animals have to feed, is one of the major bottlenecks in the laboratory mariculture production of the animals. Veligers that successfully metamorphose are those that have eyes, a relatively large shell (~ 310 um), enlargement of the foot and propodium, reduced swimming activity, settling at the bottom of their aquatic environment (rather than higher up in a water column), and the emergence of a tooth at the base of the shell aperture. Thirty four days post-hatching, competent veligers begin crawling on their preferred food source of hydroids. Within the next 12-24 hrs metamorphosis is initiated, and the animal undergoes a radical change in appearance and behavior: the velum is lost, the larva crawls out of its shell, pairs of tentacle and cerata buds emerge, and the operculum is lost. By 4-5 days post metamorphosis, the juvenile animal is progressively taking on the form, diet, and behavioral patterns of the adult. Reproductive maturity is typically reached about 1.0-1.5 months following metamorphosis, and the cycle begins again. Generation time (egg-to-egg) can be as short as ~ 2.5 months (though it requires about 163 days in the lab). This sequence of life history stages is the same as other opisthobranchs: hatching, competency to metamorphose, velum loss, shell and operculum detachment and loss, and development of the visceral mass into the foot.


Modern History of the Preparation

Early systematic work on Hermissenda derives from Agersborg’s careful descriptions of its behavior and gross anatomy in the early 1900's. Modern work on the preparation began in the 1960’s. Several researchers, notably Michael Dennis, in Donald Kennedy’s laboratory at Stanford University recognized the simplicity of its “camera” eyes, the occurrence of lateral inhibition between photoreceptors, and the potential for insights into how simple eyes could form images. This group began to characterize the synaptic connections between the photoreceptors and their role in the formation of a spatial representation of the visual field. They found that each eye consists of a lens, a cup of black screening pigment to ensure that light enters the eye from just the lens, and an underlying retinal mosaic consisting of only five photoreceptor cells. Surprisingly, an eye this primitive is capable of detecting the position of small light sources in the visual field.

In the early 1970's, Daniel Alkon, in Michaelangelo Fuortes's laboratory at the National Institutes of Health (NIH), established the preparation as a useful one for sensory neurophysiology with his studies of multiple ionic conductances underlying sensory generator potentials in photoreceptors and statocyst hair cells, detailed synaptic interactions among the 5 photoreceptors in each eye and photoreceptor synaptic interactions with second order visual neurons. The statocyst is a gravity- and body-position sensing organ, analogous to the vestibular organs of vertebrates. The hair cells within each statocyst, as well as between ipsilateral and contralateral statocysts, were shown by Alkon and a colleague, Peter Detwiler, to have an elaborate network of synaptic interactions.

These neurophysiology studies together with behavioral studies showing true classical (Pavlovian) conditioning, performed at the NIH and the Marine Biological Laboratory, led to the development of Hermissenda as a model system for studying the cellular bases of associative learning. Most of the work on associative learning and memory in Hermissenda has focussed on neural changes responsible for alterations in light-directed locomotion (phototaxis), and light-evoked reflexive behaviors, produced by pairings of light and vestibular stimulation. Other learning paradigms have been developed as well, notably changes in feeding and food-directed behaviors to chemosensory cues, due to pairings of food extracts and rotation/turbulence.

Behavior and Learning

Locomotion in Hermissenda is a complex motor act that can be elicited by a variety of stimuli, including food, pheromones, noxious events, gravitational stimuli (negative geotaxis), and light (phototaxis). It occurs only in a forward direction (i.e., head first). Like other taxes, locomotion is often directed towards or away from an eliciting stimulus, depending upon the quality and intensity of the stimulus, its motivational valence, and the behavioral state of the animal. In the apparent absence of any of these stimuli, animals also spontaneously locomote, though this generally lacks the directionality of elicited locomotion. As in some other opisthobranchs, locomotion in Hermissenda appears to involve both peristaltic-like waves of contraction and relaxation of the pedal musculature (pedal wave) as well as pedal ciliary beating (mucociliary movement). The relative contributions of these two processes to pedal locomotion is still under investigation.

When Hermissenda's foot is viewed from below while the animal is locomoting, the pedal wave is clearly visible as a rearward movement of the lateral margins of the foot. Stimulation of several individual pedal motoneurons results in clear contractions of those portions of the pedal musculature that they innervate. When this occurs unilaterally, it can result in the animal turning toward the stimulated side as portions of the ipsilateral pedal musculature contract. Observations of pedal cilia when animals are inverted often reveal them to be spontaneously active, presumably due to intrinsic motility mechanisms. As shown for other opisthobranchs, pedal ciliary activity is likely to be modulated by the central nervous system (CNS).

In addition to the orienting movements towards light, Hermissenda will suppress movement and contract its foot (the clinging response) in response to vestibular stimulation , such as turbulence, as well as to local tactile stimulation. In contrast to some other nudibranchs and opisthobranchs Hermissenda do not swim, not even to escape predators. However, rhythmic alternating left-right flexions can be elicited in the laboratory using application of salt crystals.

Hermissenda can learn to associate two different environmental stimuli, analogous to classical or Pavlovian conditioning. Hermissenda learns to associate light, which serves as the conditioned stimulus (CS), with turbulence, the unconditioned stimulus (US). After repeated pairings of the CS (light) and the US (turbulence), presentation of the light alone evokes a conditioned response (CR). The CR that emerges due to pairings (excitatory conditioning) consists of delayed phototaxis and foot contraction, resembling the unconditioned response (UR) elicited by the US. Similar to excitatory classical conditioning in vertebrates, the temporal interval between the CS and US presentations is critical. Specifically, the CS and US must overlap together in time (temporal contiguity) for strong (excitatory) conditioning to occur.

Other characteristics of vertebrate classical conditioning are observed in Hermissenda classical conditioning.

  • Pairing specificity: The suppression of phototaxis and the light-elicited contraction response are unique outcomes of stimulus pairings; they are not produced by exposure to just light alone or turbulence alone (i.e., minimal contributions of non-associative learning processes.)
  • CS specificity: Pairing light with turbulence reduces the normal positive phototactic response, but produces little systematic change to gravitational or chemosensory stimuli (not involved in training).
  • Contingency sensitivity: Learning proceeds more slowly, and the asymptotic strength of conditioning is weaker, if unpaired stimuli (CSs and/or USs) are interspersed with pairing trials.
  • Forgetting: The conditioned response weakens and may eventually be lost if training ceases for several days or more.
  • Savings: Following forgetting, retraining is faster than original training.
  • Extinction: Presentation of light alone trials (non-reinforced CS presentations) results in disappearance of the conditioned response.
  • Massed vs. distributed training trial effects: Optimal conditioning occurs if the training trials are spaced out in time, rather than being compressed together.
  • CS stimulus intensity effects: Optimal conditioning occurs if the light (CS) is of intermediate intensity; the use of dimmer or brighter light intensities results in attenuated conditioning.

In another variation on classical conditioning, termed inhibitory conditioning, light and turbulence are presented on each training trial in an explicitly unpaired (EU) fashion, i.e., they are always separated by a relatively long time interval (e.g. 4 min) and thus never overlap. Animals exposed to EU training exhibit enhanced phototaxis, increased light-elicited foot expansion, show slower acquisition of excitatory conditioning when light is subsequently paired with turbulence, and other behavioral changes indicative of inhibitory conditioning. Collectively, these changes reflect the animal having learned that light signals the absence of turbulence, just as animals exposed to pairings can be said to have learned that light signals the imminent occurrence of turbulence.

The rather remarkable generality of the above associative learning phenomena for Hermissenda, other invertebrates, and vertebrates (including mammals) encourages the view that they may be the outcome of a limited set of cellular and molecular mechanisms that are highly conserved across evolution.

Overall Nervous System Organization

The CNS consists of a set of fused ganglia arranged in a ring around the esophagus, hence the name circumesophageal ganglia. This set of ganglia include paired cerebropleural, pedal, buccal, and optic ganglia (Fig. 2 Figure 1)
Figure 2: Hermissenda circumesophageal nervous system (CNS, brain). The CNS has been stained with toluidine blue, and is shown in situ, straddling the esophagous atop the buccal mass. Right: cpg, right cerebropleural ganglion; ped, right pedal ganglion. Left: p1, p2, p3, left pedal nerves 1, 2, & 3; the left eye and statocyst are also indicated
. Similar to Tritonia, the cerebropleural ganglion replaces the separate cerebral and pleural ganglia seen in other opisthobranchs. Associated with the CNS are the two eyes, with connecting optic ganglia and two statocysts. The latter three organs are located in the crease between the cerebropleural and pedal ganglia. Emanating from each of the pedal ganglia are three pedal nerves (p1, p2, p3), which control locomotor behavior and reflexive foot and body wall movements. Sensory and motor nerves innervating the tentacles and rhinophores are derived from neurons in the cerebropleual ganglia. The buccal nerves and connectives originate within the buccal ganglia, and are involved in movements of the buccal mass.

Visual System

Hermissenda has one pair of eyes, each located at the junction between the cerebropleural and pedal ganglia. Each eye has five photoreceptors, three of type B and two of type A. The photoreceptors are sensory cells for the CS stimulus and transduce light energy into depolarization. The apical pole of each photoreceptor has an elaborate microvillar or rhabdomeric membrane, abutting the lens of the eye, that is specialized for phototransduction. Similar to other invertebrate photoreceptors, Hermissenda photoreceptors depolarize in response to light. In addition, the photoreceptors generate action potentials, a process called “firing” or "spiking". The differences between type A and type B photoreceptors are reminiscent of the difference between vertebrate rods and cones. Type A photoreceptors are sensitive to bright lights, whereas type B photoreceptors are sensitive to dim lights. Type A photoreceptors respond rapidly to an increase in light, then rapidly adapt to the light, and then, when light is terminated, they quickly cease firing. This rapid cessation of firing is due to their fast dark adaptation and ionic conductance changes that quickly repolarize them. In contrast, type B photoreceptors fire less strongly in response to an increase in light, they light adapt less than type A photoreceptors, and they continue to fire for several minutes following termination of light. One exception to the parallels between type A vs. B cells and vertebrate rods vs. cones concerns the wavelengths of light to which they are most sensitive. Rods and cones have different spectral sensitivies, while Hermissenda type B and A photoreceptors have the same spectral sensitivity (that of the photopigment rhodopsin).

Each photoreceptor gives off a single neurite (combination axon and dendrite), where spikes are generated; all five neurites unite to form the optic nerve at the base of the eye. The optic nerve passes through the optic ganglion, enters the pleural portion of the cerebropleural ganglion, and extends medially ~ 150 um before ending in an arborization of fine nerve endings, where the majority of synaptic interactions occur. The photoreceptors synaptically inhibit each other via fast inhibitory post synaptic potentials. Type B cells are mutually inhibitory while type A cells show little or no interactions (Figure 3). However, the laterally located type A receptor is reciprocally inhibitory with the type B cells in the same eye.

Figure 3: Network of Hermissenda photoreceptors and interaction with Hair Cells. Type B photoreceptors are mutually inhibitory, and Lateral A photoreceptors are mutually inhibitory with Lateral B photoreceptors. Caudal Hair cells inhibit photoreceptors (Red lines) and photoreceptors inhibit hair cells (Blue lines), not all connections are shown.

This reciprocal lateral inhibition is crucial to processes of spatial contrast enhancement, and underlies the ability of the animal's eyes to respond differentially to objects moving in different directions and at different rates within the visual field. Lateral inhibition between the photoreceptors also facilitates the animal's detection of light-dark boundaries, and plays an important role in the animal’s phototactic behavior.

Vestibular System

The vestibular system consists of one pair of statocysts, each of which has 13 hair cells, arranged in a sphere, and filled with statolymph fluid. Several hundred statoconia (tiny pebbles) are buoyant within the statolymph. Movement that produces acceleration, such as gravity or turbulence, cause the statoconia to collide with and bend the cilia of the hair cells (Figure 3). These cilia are "true" cilia (9+2 arrangement of microtubules), and intrinsically motile. Ciliary beating and bending opens membrane ion channels, leading to a depolarizing generator potential. Action potentials generated during depolarization causes release of the neurotransmitter GABA. Hair cells that lie at opposite poles of the statocyst reciprocally inhibit one another, with the strength of the synaptic inhibition being a graded function of the distance between the cells. It is strongest for cells separated by 180 degrees, and diminishes as that separation decreases. In addition, adjacent hair cells are electrically coupled, which tends to synchronize the trains of action potentials arising from one side of the statocyst. Both processes probably serve as spatial contrast enhancement mechanisms to sharpen the animal's representation of body space.

Chemosensory System

One pair of tentacles acts not only as somatosensory organs but also chemosensory organs. The rhinophores also are sensitive to dissolved chemicals, touch, and also water movement.

Motor system

Locomotion is controlled by pedal nerves nerves p1 and p2 through which all motoneuron axons travel. Multi-unit recording from pedal nerves shows relatively high levels of spontaneous activity in the dark. Low frequency, periodic bursts of synchronized activity are apparent, and are indicative of a central pattern generator (CPG). Light stimulation increases burst amplitude and frequency; however, in classically conditioned animals, the light-induced burst activity decreases below the dark activity level.

The isolated CNS, and even a single pedal ganglion, is capable of generating rhythmic multiple unit activity similar to that observed during locomotion. This suggests that the basic neuronal elements controlling locomotion are located within the pedal ganglia. Each of the two pedal ganglia contains its own CPG (though the neurons composing the CPG in Hermissenda have not yet been identified) and can generate rhythmic activity even in the absence of feedback about muscular movements. Synchronization and coordination of their activity occurs via the pedal connective (subesophageal commissure).

Several motoneurons (MNs) have been identified. The MN1 cell does not contribute to burst activity, but is involved in turning. LP1 is a motoneuron found in the left pedal ganglia, but does not have an obvious homolog in the right pedal ganglion. It is involved in cerata movement. MN4, P7 and P9 have axons in the pedal nerves, and exhibit light modulated activity. They are considered putative motoneurons because their effect on locomotion has not been established. VP1 and VP2 are motor neurons located on the ventral surface of the pedal ganglia, with axons in nerve P2. VP1 is responsible for pedal ciliary movement, and VP2 activation produces anterior foot and ventral tentacle movement. Both motoneurons are activated by light. VCMN is a foot contraction motoneuron located on the ventral surface of the pedal ganglia.

Intersensory Integration

A major theme of research with Hermissenda is that intersensory convergence occurs at multiple sites within the CS and US pathways involved in classical conditioning: at sensory receptors, interneurons, and motoneurons. Several of these convergences are essential for the initial induction of the neural plasticity underlying associative learning, and may also be critical for its maintenance (memory) and expression in behavior.

The first sites of intersensory convergence are the primary sensory neurons of the CS (light) and US (vestibular) neural pathways, the ocular photoreceptors and statocyst hair cells, respectively. Hair cells and photoreceptors form reciprocal monosynaptic inhibitory synapses (Figure 3). Type B photoreceptors receive both direct (monosynaptic) and indirect (polysynaptic) influences from the statocyst hair cells. Type B and A photoreceptors are monosynaptically inhibited by caudal hair cells, which release GABA that then acts through GABA-A receptors to produce fast synaptic inhibition. GABA also acts on a slower time scale, through G-protein coupled GABA-B type receptors, to produce a small delayed depolarization. Additional polysynaptic routes also exist through which hair cells modulate activity in type B cells. These involve serotonin stimulation of type B photoreceptors.

A variety of classes of cerebropleural interneurons contribute to phototaxis in various ways. By definition, these neurons are polysensory and respond to at least one stimulus modality other than vision. Four distinct and important groups are the optic ganglion interneurons, the second-order visual central (Type I) interneurons, the Type II and III interneurons, and the serotonergic immunoreactive interneurons.

Figure 4: Identified Neurons from Sensory to Motor. A - Type A photoreceptor, B - Type B photoreceptor, HC - Hair Cell, SE shows coupled S and E optic ganglion cells. Interconnections between photoreceptors and hair cells are omitted from this diagram. Lines projecting across midline indicate contralateral projections. Unconnected lines imply monosynaptic connections not yet identified.

Optic Ganglion Neurons. The 14 neurons in each optic ganglion have been sorted into several classes (C, D, E and S). All are synaptically inhibited by ipsilateral type B, but not A, photoreceptors. Optic ganglion cells are also a site for intersensory convergence; they are synaptically inhibited by trains of action potentials in ipsilateral statocyst hair cells. With one exception (S-E cell complex), optic ganglion cells do not feed back onto photoreceptors, nor do they interact within the same optic ganglion. But some classes (C and D) do interact across the Hermissenda brain. The role of the optic ganglion cells in visually-guided behavior of Hermissenda is still poorly understood. The C and D neurons may mediate cross-brain comparisons of visual input to the two eyes important for position and movement perception. The S-E optic ganglion cell complex may play a role in the induction of learning-related plasticity in type B photoreceptors. This complex consists of two distinct (S and E) cells that have been regarded as a single functional unit, due to their electrical coupling. In the absence of visual or vestibular stimulation, the S-E cell complex is responsible for positive synaptic feedback onto all three ipsilateral type B photoreceptors and inhibitory synaptic input to ipsilateral caudal hair cells (Figure 3). Type B photoreceptors and ipsilateral hair cells synaptically inhibit the S-E cell complex, which responds to the termination of inhibition with a rebound depolarization.

Second-Order Visual Central (Type I) Interneurons. One class of cerebropleural (CP) interneurons receives synaptic excitation from three distinct sensory pathways: the visual (EPSPs from ipsilateral type B photoreceptors), chemosensory, and statocyst. A second class of visual neurons receives synaptic inhibition from the same three sensory pathways (IPSPs from ipsilateral B cells). Thus, type B photoreceptors have a dual synaptic action on central visual interneurons. The same B cell projects to more than one CP interneuron, directly exciting some while inhibiting others. Those CP interneurons, however, are not affected monosynaptically by other photoreceptors. Thus, a divergent “labeled-line” organization seems to apply for different individual photoreceptor subtypes and the CP interneurons. For example, if a lateral type B photoreceptor monosynaptically excites (or inhibits) a CP interneuron, that interneuron is not monosynaptically affected by either type A photoreceptor, or a medial type B photoreceptor. If a medial type B photoreceptor monosynaptically excites (or inhibits) a CP cell, that interneuron is unaffected by either type A, or the lateral type B, cell.

Type II and III Interneurons. These spontaneously active interneurons are light sensitive but do not receive monosynaptic projections from photoreceptors. Both type IIe and IIi interneurons receive polysynaptic input from photoreceptors. Type IIe interneurons are excited by light, and project primarily to the ipsi-lateral pedal ganglion. Type IIi interneurons are inhibited by light and project to the contralateral cerebropleural ganglion. The pathway from these interneurons to the motor neurons is polysynaptic. Additional interneurons, such as type III, project directly to motor neurons in the pedal ganglia.

Serotonin Immunoreactive Interneurons. Three clusters of serotonergic immunoreactive (5-HT-IR) cerebropleural interneurons are found in Hermissenda. These are common to all nudibranch species that have been examined, and may be homologues of the dorsal swim interneurons in Tritonia and other similar interneurons in many opisthobranchs. These cells may function as a general arousal system in opisthobranchs, serving to increase the excitability of diverse sensory and motor pathways when animals encounter aversive stimuli. One of these clusters, the CPT triplet, was shown to project to photoreceptors and optic ganglion cells. Electrophysiological studies indicate that the CPT triplet of interneurons appears to integrate information from both the visual system and cutaneous stimulation arising from the foot; these cells are excited by light and cutaneous stimulation of the middle or tail regions of the foot. The CPT cells project to VP2 motoneurons where they contribute to reflexive foot movements, but evidently not to ciliary movement.

The functional significance of these multiple intersensory convergences is still being studied. Several are critical for the occurrence and expression of associative learning and memory. Others may provide the underpinnings for behavioral choice. For example, the chemosensory pathway is capable of inducing profound synaptic inhibition of both the visual and vestibular pathways. This may reflect the fact that the animal’s need for food will generally supercede the animal's orientation with respect to light and gravity.

Changes in photoreceptor properties with learning

Photoreceptors are the first site of convergence of the CS and US stimuli, and a key locus of memory storage. When activated by turbulence, hair cells in the statocyst release the inhibitory neurotransmitter GABA onto the terminal branches of the photoreceptors. Thus, during excitatory classical conditioning, light stimulated photoreceptors receive GABAergic input from the vestibular system. In addition, both light and strong vestibular stimulation inhibit are the S-E cell complex in the optic ganglion. Following their termination, the S-E cell is transiently disinhibited and an increase in positive synaptic feedback onto ipsilateral type B photoreceptors occurs. This pairing-specific increased positive synaptic feedback potentially contributes to the prolonged depolarization of the type B photoreceptors due to pairings. Inactivation of the S-E cell complex during in vitro paired conditioning attenuates the cumulative depolarization of type B photoreceptors that is a short-term neural correlate of conditioning.

Both excitatory classical conditioning, and in vitro conditioning, in which the central nervous system receives paired light and statocyst hair cell stimulation, produce long term changes in photoreceptor properties. Type B photoreceptors show an increase in excitability, seen as increases in input resistance, light-evoked generator potentials and firing frequency, and an enhanced long lasting depolarization after light termination. Type A photoreceptors show a decrease in excitability, seen as a decrease in input resistance and a decrease in the light response. The increase in excitability of type B photoreceptors is due to reductions in transient voltage-dependent and sustained calcium-dependent potassium currents; whereas the decrease in excitability in type A photoreceptors is due to an increase in the delayed rectifier potassium current. Inhibitory classical conditioning produces changes in the photoreceptors that are the opposite of those produced by pairings of visual and vestibular stimulation. The light-evoked generator potentials and firing frequency of type B photoreceptors are reduced by EU-training, while those of the type A photoreceptors are enhanced. These changes in photoreceptor light responses and spiking lead to increased synaptic transmission between photoreceptors, as well as between photoreceptors and select interneurons. The alterations in synaptic transmission contribute to decreased phototaxis and enhanced light-evoked foot contraction, through separate neural circuits that are still being characterized.

The photoreceptors within an eye are interconnected with synapses. The three type B photoreceptors are mutually inhibitory, and the type A photoreceptors are inhibited by the type B photoreceptors. Both classical conditioning and in vitro conditioning produce changes in the strength of the synaptic connections. The amplitude of individual inhibitory synaptic potentials evoked in one member of a pair of photoreceptors by action potentials in the other are increased following pairings of light and rotation. Most of this synaptic facilitation appears due to changes in the amplitude and kinetics of the presynaptic action potential.

Signaling Pathways Underlying Changes in Photoreceptor Properties

Figure 5: Signaling Pathways underlying classical conditioning. Molecules are color coded as follows: receptors - magenta; membrane associated molecules - pink; membrane bound enzymes are yellow; diffusible messengers - pale green; calcium release channels - bright green; enzymes ultimately responsible for long-term memory storage - light blue. Solid lines indicate known pathways; dashed lines indicate hypothesized or multistep pathways. abbreviations: PLC - phospholipase C, PLA2 - Phospholipase A2, DAG - diacylglycerol, AA - arachidonic acid, PKC - protein kinase C, MAPK - mitogen activated protein kinase, RyR - Ryanodine Receptor, IP3R - IP3 receptor, 5HT - serotonin. Stimulation of the Cs pathway by light begins with the absorption of photons by rhoposin. Stimulation of the US pathway by vestibular stimulation begins with mechanosensory-induced depolarization of the hair cells.

Long term memory storage, consisting of changes in ionic and synaptic currents, only occurs when light is paired with turbulence. Several intracellular events have been proposed as the mechanism sensitive to temporal pattern of paired stimuli in the photoreceptors. One mechanism is an elevation in calcium, produced by both light stimulation and turbulence induced GABA release onto photoreceptors. The pairing of light-induced calcium increases in type B photoreceptors with stimulation of GABA receptors by vestibular-produced hair cell stimulation contributes to the induction of learning changes in type B cell excitability. Rebound firing of the S-E ganglion cells contributes to B cell depolarization and calcium influx. Blocking the elevation in calcium prevents the changes in photoreceptor excitability correlated with memory storage. A second mechanism is protein kinase C, which is activated by classical conditioning and modulates ionic channels. Activation of protein kinase C requires calcium and diacylglycerol; PKC is further stimulated synergistically by arachidonic acid, which is produced by phospholipase A2. Light stimulation activates phospholipase C, which produces diacylglycerol and an elevation in calcium. GABA release from hair cells produces an elevation of arachidonic acid. Thus, the pairing of turbulence and light synergistically activates PKC which leads to changes in photoreceptor properties. An additional signal transduction pathway that may be important involves mitogen activated protein kinase (MAPK), which has been proposed to be activated by pairings of serotonin-stimulation and light. Both PKC and other as yet unidentified kinases contribute to MAPK activation.

Computational Models of Hermissenda Classical Conditioning

The goal of computational models is to demonstrate how paired stimuli produce changes in neuronal properties, neuronal activity, and motor output. Models of the photoreceptors include voltage-dependent, calcium-dependent, light-induced, and leakage currents. Simulations confirm prior experimental results that a reduction in potassium currents produces an increased generator potential, enhanced long lasting depolarization, increased firing frequency, and spike broadening. Simulations also reveal that the increased excitability - both input resistance and long lasting depolarization - require reductions in the light induced potassium current, a change not yet demonstrated experimentally. Another type of computational model evaluates the signaling pathways underlying learning in the photoreceptor, and includes phototransduction and calcium dynamics. This model demonstrates that the calcium elevations produced by both light and GABA cannot explain the sensitivity to temporal pattern. This reinforces the importance of second messenger signals such as diacylglycerol and arachidonic acid that activate protein kinase C, and serotonin that leads to MAPK, and justifies further experiments to identify which interactions among signaling pathways are sensitive to temporal interval of CS and US stimuli.


Alkon DL, Fuortes MGF. 1972. Response of photoreceptors in Hermissenda. Journal of General Physiology 60:631-649.

Alkon DL: Neural organization of a molluscan visual system. J Gen Physiol. 61: 444- 461, 1973.

Alkon DL, Bak A: Hair cell generator potentials. J Gen Physiol. 61: 619-637, 1973.

Alkon DL: Intersensory interactions in Hermissenda. J Gen Physiol. 62: 185-202, 1973.

Alkon DL, Lederhendler I, Shoukimas JJ: Primary changes of membrane currents during retention of associative learning. Science 215: 693-695, 1982.

Alkon DL, Sakakibara M, Forman R, Harrigan J, Lederhendler II, Farley J. 1985. Reduction of two voltage-dependant K+ currents mediates retention of a learned association. Behavioral and Neural Biology 44:278-300.

Blackwell KT. 2004. Paired turbulence and light do not produce a supralinear calcium increase in Hermissenda. J Comput Neurosci 17:81-99

Blackwell KT. 2006. Subcellular, Cellular, and Circuit Mechanisms Underlying Classical Conditioning in Hermissenda crassicornis. The Anatomical Record (Part B: New Anatomist)

Crow T, Xue-Bian JJ, Siddiqi V, Neary JT. 2001. Serotonin activation of the ERK pathway in Hermissenda: contribution of calcium-dependent protein kinase C. J Neurochem 78:358-364.

Crow T, Tian LM 2002 Facilitation of monosynaptic and complex PSPs in type I interneurons of conditioned Hermissenda. J Neurosci 22:7818-7824

Crow T. 2004. Pavlovian conditioning of Hermissenda: current cellular, molecular, and circuit perspectives. Learn Mem 11:229-238.

Detwiler PB, Fuortes MG. 1975 Responses of hair cells in the statocyst of Hermissenda. J Physiol. 251:107-29.

Farley J. 1987 Contingency learning and causal detection in Hermissenda: I. Behavior. Behav Neurosci. 101:13-27.

Farley J, Richards WG, Grover LM. 1990. Associative learning changes intrinsic to Hermissenda type A photoreceptors. Behavioral Neuroscience 104:135-152.

Goh Y, Alkon DL. 1984. Sensory, interneuronal, and motor interactions within Hermissenda visual pathway. J Neurophysiology 52:156-169.

Lederhendler I, Alkon DL. 1986 Implicating causal relations between cellular function and learning behavior. Behav Neurosci. 100:833-8.

Lederhendler II, Gart S, Alkon DL 1986 Classical conditioning of Hermissenda: origin of a new response. J Neurosci 6:1325-1331

Muzzio IA, Gandhi CC, Manyam U, Pesnell A, Matzel LD. 2001. Receptor-stimulated phospholipase A(2) liberates arachidonic acid and regulates neuronal excitability through protein kinase C. J Neurophysiol 85:1639-1647.

Olds JL, Anderson M, McPhie DL, Staten L, Alkon DL: Imaging of memory-specific changes in the distribution of protein kinase c in the hippocampus. Science 245: 866-869, 1989.

Richards WG, Farley J. 1987. Motor correlates of phototaxis and associative learning in Hermissenda crassicornis. Brain Res Bull 19:175-189.

Schuman EM, Clark GA. 1994. Synaptic facilitation at connections of Hermissenda type B photoreceptors. J Neuroscience 14:1613-1622

Internal references

  • Eugene M. Izhikevich (2006) Bursting. Scholarpedia, 1(3):1300.
  • Howard Eichenbaum (2008) Memory. Scholarpedia, 3(3):1747.
  • John Dowling (2007) Retina. Scholarpedia, 2(12):3487.
  • Arkady Pikovsky and Michael Rosenblum (2007) Synchronization. Scholarpedia, 2(12):1459.
  • Paul S. Katz (2007) Tritonia. Scholarpedia, 2(6):3504.
  • Kathleen Cullen and Soroush Sadeghi (2008) Vestibular system. Scholarpedia, 3(1):3013.
Personal tools
Focal areas