Notice: Undefined offset: 6624 in /var/www/ on line 5961
Lymnaea learning and memory - Scholarpedia

Lymnaea learning and memory

From Scholarpedia
Paul R. Benjamin and György Kemenes (2010), Scholarpedia, 5(8):4247. doi:10.4249/scholarpedia.4247 revision #138129 [link to/cite this article]
Jump to: navigation, search
Post-publication activity

Curator: Paul R. Benjamin


Lymnaea is an important model system for learning and memory studies

Lymnaea is an attractive model for neuroscientists interested in understanding the behavioral, neural and molecular basis of associative (Pavlovian) learning and memory. The behavior of this pond snail is surprisingly dynamic and snails can be trained to respond to a wide variety of sensory cues. Feeding and respiration have proved to be particularly useful for Lymnaea learning studies because the neural circuitry underlying these behaviors is well-described (Benjamin, 2008). Electrical changes induced by conditioning occur in individually identified neurons that are key elements of the motor circuitry underlying the two behaviors. Single-trial conditioning has been developed successfully in both classical and operant conditioning. Single-trial paradigms are particularly useful for studying the temporal sequence of electrical and molecular changes that occur at various time-points after training. The molecular analysis of memory formation has shown that it involves conserved molecular cascades present in all organisms. One of the most exciting aspects of the Lymnaea work is that quantitative changes in gene expression induced by conditioning can be studied at the level of single identified neurons.

Behavioral mechanisms

Classical conditioning

The feeding system of Lymnaea has been used extensively to investigate reward-based classical conditioning. Feeding in Lymnaea is a rhythmic motor behavior where repeated rasping movements of a toothed tongue or radula scrape at the surface of floating pond-weed or other food substrates leading to the ingestion of food. Sucrose acts a convenient food stimulus in the laboratory and is used as a rewarding stimulus in conditioning experiments. Tactile, chemical and visual cues have all been used in classical conditioning experiments (Benjamin & Kemenes, 2009). Lymnaea is a generalist in its feeding habits and it is advantageous for it to learn about potentially useful foods using a variety of sensory modalities. In tactile conditioning experiments, a neutral stimulus, lip touch (the CS or conditioned stimulus) is paired with a strong feeding stimulus like sucrose (the US or unconditioned stimulus). After repeated CS+US pairing, application of touch to the lips induces a sequence of rhythmic feeding movements that is not observed in control groups of snails in which the CS and US are separated in time (unpaired group) or the CS and US are applied alone (Kemenes & Benjamin, 1989). By explicitly pairing touch with sucrose the snail has learned that touch means food. This type of tactile reward-learning shares important characteristics with associative conditioning in vertebrates. It shows stimulus generalization, so that training to one site on the body (the lips) transfers to another site on the body (the tentacles). Success in tactile training depends on internal and external variables. For example, snails learn only if they are hungry (internal variable) and learn better if they are placed in a novel clean water environment (external variable). Tactile conditioning requires multiple trials (5 trials per day for 3 days) but another type of reward conditioning, chemical conditioning, is successful even with only a single trial (Alexander et al., 1984). In this example, amyl acetate is the CS (Fig. 1Ai) and sucrose the rewarding US (Fig. 1Aii) and pairing these two chemical stimuli results in a feeding response to the CS when applied in the post-training test (Fig. 1Aiii). Again background variables are important, so one-trial conditioning only works in snails starved for 4 or 5 days; ‘old’ snails learn less well than young ones and then only in multi-trial experiments. A single pairing of amyl acetate and sucrose in starved snails results in a long-term memory (LTM) trace that lasts for at least 19 days, a remarkable example of associative learning. Single-trial conditioning of feeding also is possible with visual cues (Andrew & Savage, 2000). Here the snail is rewarded with sucrose when it approaches a black panel along a gutter. After conditioning, approaching the black panel elicited more rasping movements. Lymnaea has a lens that is capable of forming an image on the retina underwater and reaching behavior is evoked by a black and white check pattern. Successful reward conditioning was obtained with this checkered pattern that was discriminated from a grey pattern of equal luminance, but in this experiment multiple trials were necessary (4 trials per day for 4 days). A memory trace for this visual stimulus was recorded 4 days after conditioning.

Aversive classical conditioning of feeding (conditioned taste aversion, CTA) is also possible in Lymnaea (Ito et al., 1999). In this type of conditioning sucrose (CS) is paired with KCl (US), an aversive stimulus that inhibits feeding. After 8 trials, trained animals show a weaker feeding response to sucrose compared with controls and this memory lasted for over a month. A second type of aversive classical conditioning involves the conditioning of whole-body withdrawal responses to a visual CS (Sakakibara, 2006). Lymnaea does not normally respond to a light flash (CS) but after body rotation (US) light evokes a whole-body withdrawal response. This type of learning requires the interaction between two sensory organs, the eyes and the statocysts (the snail equivalent to the vestibular organ in vertebrates) and so has been called ‘visuo-vestibular conditioning’. It requires at least 10 training trials over 3 days and results in long-term memory trace lasting for at least 1 week.

Figure 1: Classical and operant conditioning in Lymnaea: behavioral paradigms A, Food-reward classical conditioning. Ai, Before training, amyl acetate fails to evoke feeding. Aii, During training, amyl acetate (the CS) is paired with sucrose (the US). Aiii, After training, amyl acetate evokes feeding. B, Aversive operant conditioning. Bi, Under hypoxic conditions, rhythmic aerial respiration occurs. Bii, When a tactile stimulus (the US) is applied during aerial respiration (the operant behavior), the pneumostome closes. Biii, After training, the number of respiratory movements is reduced.

Operant conditioning

The aerial respiratory behavior of Lymnaea has been used extensively to study the behavioral mechanisms of aversive operant conditioning and to investigate the influence of environmental factors that are known to influence learning in a number of other systems. When the environment is made hypoxic the snails float to the water surface and perform rhythmic opening and closing movements of their pulmonary opening, the pneumostome, and this respiratory behavior is used as the operant for behavioral conditioning (Lukowiak et al., 1996) (Fig. 1Bi). Gentle tactile stimulation of the pneumostome area evokes pneumostome closure and stops respiratory behavior (Fig. 1Bii). During operant conditioning, a tactile stimulus is applied to the pneumostome, repeatedly, each time the aerial respiration is attempted. Conditioning reduces the number of openings (Fig. 1Biii), latency to first opening and total breathing time compared with pre-training or when compared with yoked controls. A variety of training protocols have been used that determines the duration of the memory trace. Training for 0.5 hr is sufficient to produce a memory trace that is 3 hr but not 4 hr in duration and this has been classed as intermediate memory (ITM). A single 1 hr training session is sufficient to establish a memory that persists for 24 hr and this has been classed as LTM. The duration of LTM can be increased beyond 24 hr by repeated sessions of spaced training. An interesting recent aspect of the work on operant conditioning was to study the role of environmental ‘context’ in determining the expression of LTM (Haney & Lukowiak, 2001). One group of snails were trained under standard hypoxic conditions but the context of training was altered in a second group by training the snails in standard conditions but with the addition of extracts of a food odorant, derived from carrot (the context). The presence of the odor did not prevent learning and formation of LTM but expression of the memory trace depended on testing the snails in the context in which they were trained. Thus snails trained with the odor context did not demonstrate LTM unless they were tested in the presence of the odor and vice versa for snails trained under standard conditions. Context also influences the extinction of the LTM trace (McComb et al., 2002). An operant training procedure of two 45 min training sessions followed by a third 18 hr later produced an extended duration of LTM lasting for at least 5 days. If after the last training session the snails were subjected to three 45 min extinction ‘training’ sessions (breathing with no reinforcing touch stimulus) then LTM was not observed on the next day. Extinction does not occur, however, if the extinction training was carried out in a context that was different from the context of the associative training. Another type of chemical signal that influences operant conditioning is the scent of a crayfish predator. ‘Stress’ induced by this chemical has been shown to enhance learning as part of a range of changed behavioral responses induced by the presence of a predator (Orr & Lukowiak, 2008). Two types of effects were seen depending on the duration and spacing of the training. In one experiment, snails were trained for 30 min to produce an ITM lasting for 3 hr. Exposure to crayfish effluent (CE) during training produced a memory that persisted for 48 hr consistent with formation of LTM. A second type of experiment training consisted of two periods of 30 min training separated by a one hr interval. Here the presence of CE during training extended the LTM duration from 1 day to at least 8 days. A context-dependent element was found in these experiments so that memory recall depends on the presence of CE. However, when snails trained in CE were challenged during recall with a background of carrot or standard hypoxic conditions (no added chemical) the LTM memory trace was still expressed indicating context generalization. This contrasts with the earlier described experiments where training with a carrot context required the presence of the carrot to get memory recall, showing a requirement for context specificity. All the paradigms described so far described for operant conditioning in Lymnaea require multiple trials but recently a single-trial aversive paradigm has been developed. Instead of a gentle poke, a chemical aversive stimulus, KCl, is added to the training disk shortly after breathing (Martens et al., 2007). The LTM trace persisted for at least 24 hr.

Neural mechanisms

Initial studies aimed to correlate conditioning-induced changes at the behavioral level with systems level changes in electrical activities that underlie feeding or respiratory movements. Subsequent studies successfully located electrical changes at specific neuronal locations within the circuits. This was aided by the previous analysis of the networks underlying these rhythmic behaviors, particularly the identification of the individual neuronal components of the feeding and respiratory central pattern generator (CPG) interneurons, the modulatory interneurons that control the CPG circuits and the sensory pathways that carry the CS and the US signals. A key point here is that fundamental electrical changes underlying learning and memory occur in the same circuit that mediates the basic motor behavior often at multiple locations. We now study the synaptic and cellular (non-synaptic changes, such as changes in excitability) mechanisms that play a vital role in network plasticity and considerable progress is being made at this more detailed level of analysis. Two general approaches have been used to study the neural mechanisms of associative conditioning. One approach involves first training the intact snail behaviorally and then recording electrical changes in a semi-intact preparation made from the same trained animals. In these semi-intact preparations, intracellular electrical recording of neurons is possible whilst applying the natural stimuli to retained sensory structures. The second approach is to train a semi-intact preparation, in vitro, and examine the changes induced by training at various stages in memory formation including acquisition and consolidation of the memory trace.

Classical conditioning

Most progress has been made in understanding the neural mechanisms underlying reward-based classical conditioning of the feeding network. Using the single-trial chemical conditioning paradigm (pairing amyl acetate with sucrose), an electrophysiological correlate of the conditioned response was recorded as a sequence of CS-driven bursts (‘fictive’ feeding activity) in motoneurons recorded in semi-intact preparation made from behaviorally-conditioned snails (Kemenes et al., 2002). This CPG-driven feeding activity in the motoneurons depends on activity of neurons at all levels of the feeding network, so the conditioned fictive feeding activity recorded in motoneurons is a systems ‘readout’ of the memory trace in the whole feeding system. More detailed experiments found that conditioning affects central but not peripheral processing of chemosensory information, with the cerebral ganglia being an important site of plasticity (Straub et al., 2004). Cerebral plasticity originates from a special type of feeding interneuron type called the cerebral buccal interneuron (CBI). These CBIs act as ‘command cells’ for the feeding network so that when they fire the feeding CPG is activated followed by rhythmic ingestion movements. Neural firing in the CBIs is enhanced in response to the CS after conditioning. This increase in CBI firing is due to an increase in strength of the excitatory synapse between the CS chemosensory neurons and the CBIs (Fig. 2A). How does this come about? Considerable experimental evidence suggests that it is due to pre-synaptic facilitation of the sensory neuron to CBI synapse by the modulatory neuron type known as the cerebral giant cell (CGC). The CGCs are known to have a basic role in normal activation of feeding by food but independent of this function they also play a role in learning (Kemenes et al., 2006). The CGCs are persistently depolarized by about 10 mV after behavioral conditioning and this change in membrane potential increases the strength of the CGCs synaptic effects on feeding neurons by a process that involves an increase in intracellular CGC calcium concentration. Artificial depolarization of the CGCs in naive semi-intact preparations increases the response of the CBI to the CS, mimicking the effects of conditioning. Recently it has been shown that the conditioning-induced depolarization of the CGCs is due to an increase in the size of the persistent sodium current of its soma membrane (Nikitin et al., 2008). The measured increase in the size of this current is sufficient to depolarize the CGC by the required amount and this was confirmed by computer modelling. Surprisingly, the depolarization of the CGCs does not cause a change in the firing rate of the CGCs or their spike shape and this is due to balancing increases in two other currents, a delayed rectifier potassium current (ID)and a high voltage activated calcium current (IHVA)(Vavoulis et al., 2010). These results indicate that the cellular changes that occur in the CGCs after conditioning are sufficient to explain the enhancement of the CS effects on feeding following conditioning but they cannot be the whole story because the onset of CGC depolarization is at 16-24 hr after training whereas a behavioral memory trace is present at 2 hr after training, so an alternative mechanism must also be present to explain the early memory trace. Significantly, there is a second type of electrical change in the feeding circuit that occurs as early as 1 hr after one-trial in vitro conditioning. This is a conditioning-induced reduction in tonic (continuous) inhibitory synaptic input to the feeding CPG. By reducing the frequency of this ‘background’ inhibitory input the threshold for the CS to activate the feeding network is reduced making it more likely that the feeding network (CPG) will respond to the CS (Benjamin et al., 2009; Marra et al., 2010). This inhibitory input originates from one of the CPG interneuron type known as the N3t (N3tonic). How these two types of mechanism (pre-synaptic facilitation, reduction in tonic inhibition) interact to produce a conditioned response in the feeding network is shown in Fig. 2A.

Electrical correlates of tactile reward conditioning were also recorded at different sites in the feeding network. After 15 behavioral training trials over 3 days, touching the lips of the snail in the intact snail induces a pattern of feeding movements significantly greater than controls. Similar significant changes were recorded at the level of motoneurons in semi-intact preparations made from the same snails (Staras et al., 1998). Electrophysiological correlates of behavioral differential classical conditioning were also obtained (Jones et al., 2001). In these experiments, the lips and the tentacle were used as CS+ (reinforced conditioned stimulus) or CS- (non-reinforced conditioned stimulus) sites for behavioral tactile conditioning. In a second experimental group the CS+/CS- negative sites were reversed. Following successful behavioral conditioning, the touch stimulus evoked CPG-driven fictive feeding in CS+ but not the CS- sites in both experimental groups. More detailed changes could also be recorded in other parts of the feeding network. For instance, conditioning-induced increases in spike activity were recorded in the touch sensitive CS pathway between cerebral and buccal ganglia (Staras et al., 1999). CPG interneurons, such as the N1M (N1 medial), were also affected by learning. A long-lasting sequence of inhibitory synaptic inputs that occur in the N1M in response to lip touch change to a strong depolarizing synaptic input after in vitro conditioning and this drives a sustained plateauing pattern in the N1M cell (Benjamin & Kemenes, 2009). This is an example of synaptic plasticity affecting an important CPG element of the feeding network. A non-synaptic neuronal change also occurs after tactile conditioning that has been shown also to play an important role in tactile conditioning. This occurs in one of the CBI cell types known as the CV1. This neuron is capable of driving a feeding rhythm via its excitatory synaptic connection with the N1M cell and spiking activity in this cell normally accompanies unconditioned feeding responses stimulated by sucrose. After tactile conditioning, the CV1 cells are considerably more active following touch in conditioned snails compared with controls and show the typical rhythmic activity seen with sucrose (Jones et al., 2003). Underlying this increase in tactile responses after conditioning is a membrane depolarization of about 11 mV that persists for as long as the electrophysiological and behavioral memory trace. The depolarization makes the CV1a more responsive to the CS and accounts for the activation of the feeding response following conditioning. The significance of this result is emphasized by experiments in which the membrane potential of CV1 was manipulated to either reverse the effect of behavioral conditioning or to mimic the effects of conditioning in naive snails. These experiments showed that the persistent depolarization of the CV1 cells is both sufficient and necessary for the conditioned tactile response in the feeding network. It is interesting that there were no learning-induced changes in the CGC cells after tactile conditioning and no change in the CV1 occurs after chemical conditioning. This difference in the neuronal site of the non-synaptic change may be important in determining the specificity of the conditioned response in the two types of conditioning.

Experiments aimed at investigating the neural basis of aversive classical conditioning of the feeding systems (CTA) were carried out in isolated brains dissected from conditioned and control animals. The synaptic connection between the CGCs and the N1M was examined and a significant change in the size of IPSPs recorded in the N1M was recorded following spike induced in the CGCs by artificial depolarization (Kojima et al., 1997). Since the CGCs are known to play a critical gating role in feeding behavior and the N1M has a pivotal member of the CPG, this enhanced IPSP may be an important cellular correlate of the reduced response to sugar observed in behavioral CTA. The N1M IPSPs examined in these experiments probably originates from the N3t cells of the feeding CPG indicating that these cells are important in both reward and aversive conditioning of the Lymnaea feeding system. Neural analysis of visuo-vestibular conditioning has concentrated on the sensory processing the CS and US in the eye and statocyst, respectively, and this work forms the basis for future mechanistic studies of memory formation using this paradigm.

Figure 2: Parsimonious cellular models of classical and operant conditioning in Lymnaea A, Cellular mechanisms of memory after single-trial chemical classical conditioning of feeding. After conditioning, the modulatory interneuron CGC is persistently depolarized (bold blue outline), which facilitates (upward arrow) the CS to cerebral-buccal interneurons (CBIs) connection and thereby gates-in the activation of N1M, a key CPG interneuron of the feeding system. The spike frequency of the dual-role interneuron N3t is reduced after conditioning (*), reducing (downward arrow) the suppressive control from N1M to allow it to be more readily activated by the CS. B, Cellular mechanisms of memory after multi-trial operant conditioning of respiration. After conditioning, the respiratory CPG interneurons RPeD1 and IP3 show a reduced rate of firing (*) and the excitatory input from RPeD1 to IP3 becomes weaker (downward arrow).

Operant conditioning

Neural changes associated with this type of learning have focused on the respiratory CPG circuit, a three cell network that shows rhythmic activity under the hypoxic conditions used in the operant conditioning experiments. The main emphasis has been on analysing systems-level changes in neural activity in the isolated CNS derived from operantly conditioned snails. Specifically, spontaneous patterned activity in the IP3 (input 3) neuron, which is involved in pneumostome opening, shows a significant reduction compared with yoked controls. Furthermore, a higher percentage of RPeD1 (right pedal dorsal 1) CPG interneurons, which is involved in the triggering of CPG activity, are silent in conditioned versus control preparations (Spencer et al., 1999). A reduction in the ability of the RPeD1 cell to induce IP3 activity was also observed. These results (summarized in Fig. 2B) can explain the reduction of behavioral respiration that follows operant conditioning. More direct evidence for the role of RPeD1 in operant conditioning comes from somal ablation experiments. Removal of the RPeD1 soma in intact animals prevented LTM but had no effect on ITM without affecting the ability of the snail to carry out respiratory behavior; this suggested that the RPeD1 soma is necessary for LTM formation (Scheibenstock et al., 2002). Removal of the RPeD1 soma 1 hr after conditioning had no effect on LTM, indicating that somal ablation did not interfere with memory access or retrieval. Interesting, both extinction and reconsolidation also required the presence of the RPeD1 soma (Sangha et al., 2003a,c). Finally, more recent in vitro conditioning experiments in the semi-intact preparations showed that artificially reducing the level of RPeD1 firing in the interval between training sessions increased the duration of the memory trace to produce LTM from the ITM produced in control preparation (Lowe & Spencer, 2006), providing direct evidence that the reduction in RPeD1 firing seen in the isolated brain after behavioral training plays an important role in LTM formation. There is also electrophysiological evidence that the electrophysiological correlate of operant conditioning is altered by stress-related environmental stimuli and the context in which training takes place, so that the conversion of ITM- training to LTM by exposure to crayfish extracts during behavioral training has an electrophysiological correlate in the parallel extension in the duration of reduced RPeD1 firing (Orr & Lukowiak, 2008). Context generalization was also seen in the electrophysiological experiments.

Molecular mechanisms

One major feature that is emerging from molecular analysis of learning in Lymnaea is the conserved nature of the molecular processes involved. They are not specific to snails adding to the general importance of the results obtained. The opportunity to study a wide range of molecules involved in Lymnaea memory arises from the recent transcriptome (Feng et al., 2009) and proteomic analysis (Rosenegger et al., 2010) of the CNS.

Phases of Memory

All the types of associative conditioning paradigms described here are capable of inducing LTM lasting from hours to many days. LTM requires new RNA and new protein synthesis and the use of protein synthesis inhibitors (eg anisomycin, ANI) and RNA synthesis inhibitors (eg Actinomycin D, Act-D) and blocking memory formation using these inhibitors confirmed the formation of LTM in Lymnaea in both classical and operant conditioning paradigms. Injecting ANI into intact snails at various time points after one-trial chemical conditioning and testing for the presence of the 24 hr memory showed that there is a single critical period of sensitivity to protein-synthesis blockers in the first hr after conditioning (Fulton et al., 2005). Subsequent work using both protein synthesis and RNA synthesis blockers showed that was a long term memory trace as early as 4 hr after single- trial conditioning. Recent work shows that consolidation of 24 hr LTM can be blocked by cooling snails immediately after training and before the time period 10-60 min after training when protein synthesis blockers are known to be effective (Fulton et al., 2008). It is suggested that this very early blocking of the 24 hr memory trace by cooling is interfering with enzymic cascades including kinases that are known to be activated by chemical conditioning (see below). In operant conditioning reconsolidation and extinction of LTM have been shown to have similar general molecular requirements to LTM. ITM has been described in operant conditioning and can be distinguished from LTM by its differential sensitivity to protein synthesis and RNA synthesis blockers (Sangha et al., 2003b). ITM is sensitive to protein synthesis blockers but, unlike LTM, ITM is not blocked by RNA synthesis blockers (i.e., it does not depend on new transcription). Similar data are now available for classical conditioning based on recent in vitro conditioning experiments carried out using single-trial chemical conditioning (Marra et al., 2013). 1 hr, 2 hr and 3 hr memories are classified as ITM on the basis of their requirement for protein, but not RNA synthesis. ITM is divided into an early and late phases based on its sensitivity to protein kinase inhibitors. PKC is required for early ITM (1 hr-2 hr after conditioning) but not for late ITM (2-3 hr after conditioning). Classic short-term memory (STM) is shown at 10 min because it required neither de novo protein or RNA synthesis. LTM was first recorded at 4 hr after conditioning. The recent work on one trial classical conditioning also showed lapses in memory recall at the transitions between STM and ITM and early and late ITM. We showed that during these lapses consolidation of memory traces becomes vulnerable to external stimuli providing an mechanism that could be important adaptive process in memory formation in the natural environment (Marra et al., 2013).

Gene-regulating cascades

Regulation of gene expression following one-trial reward chemical conditioning of feeding is known to involve transcription factors like the cyclic AMP-responsive element binding protein (CREB). The highly conserved CREB gene and CREB-like proteins have been identified recently in Lymnaea (Ribeiro et al., 2003, Sadamoto et al., 2004). Consistent with a role for CREB in LTM is the observation that levels of phosphorylated CREB1 are increased in the CGCs following reward conditioning (Ribeiro et al., 2003). In more recent work it was shown that conditioned taste aversion (CTA) learning increased LymCREB1 gene expression (Sadamoto et al. 2010), so it seems that classical conditioning can increase both the level of the expression of the gene of this transcriptional activator and the level of its activation by phosphorylation. Other highly conserved molecular pathways that have been implicated in LTM are the protein kinase A (PKA) (Michel et al., 2008) and mitogen-activated protein kinase (MAPK) signaling pathways (Ribeiro et al., 2005). Inhibition of PKA catalytic subunit activity or MAPK phosphorylation without blocking sensory or motor pathways blocks LTM and levels of both PKA activity and MAPK phosphorylation were increased, with PKA activation first detected at an even earlier time window (5 min). Memory consolidation after retrieval at 6 hours post-training is both PKA and protein-synthesis-dependent, whereas reconsolidation after retrieval at 24 hr depends on protein synthesis but not PKA activity (Kemenes et al., 2006). This interesting finding indicates that depending on the ‘age’ of the consolidated memory, different molecular pathways are activated by memory retrieval and contribute differentially to memory reconsolidation. Like reward conditioning, aversive chemical conditioning (CTA) of feeding involves conserved molecular pathways linked to the PKA and the CREB transcriptional regulatory system, such as the CAAT Element Binding Protein (Hatakeyama et al., 2006). The injection of cAMP or PKA into the soma of the CGC leads to a long-term enhancement of the strength of the synaptic connection between the CGC and the B1 feeding motoneuron and presumably other neurons of the feeding network (Nakamura et al., 1999, Nikitin et al. 2006). An exciting new finding has been that the Type II Calcium Calmodulin dependent kinase (CaMKII) is intrinsically activated (by phosphorylation at T286) in a later post-training time window, around 24 hr after food-reward training. This activation occurs independently of NMDA receptor activation and is a necessary molecular component of late memory consolidation (Wan et al., 2010).

Transmitters and modulators

Figure 3: Time windows of known molecular requirements for LTM after single-trial food-reward classical conditioning in Lymnaea. The large curved arrows indicate the requirement for the activation or synthesis of molecules in the 0h to 6h (acquisition/early consolidation) and 24hr time window (late consolidation) for memory expressed at 24 hr and 42 hr, respectively. In the late memory consolidation phase, only activation of CaMKII is required for memory expression at 42 hr post-training.
An increasing number of neurochemicals, known to be important in chemical communication between neurons, are involved in various phases of memory in Lymnaea. For instance, consolidation of the 24 hr memory trace following one-trial chemical conditioning is dependent on the nitric oxide (NO)-GMP signalling pathway. There is a critical period of sensitivity up to 5 hr after conditioning when blocking this pathway by drug injection prevents behavioral LTM formation (Kemenes et al., 2002). Further evidence for a role in LTM comes from experiments on the CGCs. These neurons express mRNA transcripts from the two related nNOS genes (Lym-nNOS1 and Lym-nNOS2) (Korneev et al., 2005). Six hr after chemical conditioning, Lym-nNOS1 is up-regulated compared with controls. This up-regulation of the NOS coding transcript may be due to an earlier, 4 hr, down-regulation of the Lymnaea antiNOS that is known to be inhibitory on NOS transcript production. The importance of antisense RNA regulation in learning is emphasized by recent results showing enhanced axonal trafficking of antisense RNA to the synaptic regions of the CGCs following classical conditioning ( Korneev et al., 2013). NO is known to modulate the strength of serotonergic transmission between the CGCs and motoneurons in the feeding system (Straub et al., 2007) and this could be involved in conditioning. As well as NO, the monoamine dopamine plays a key role in reward conditioning in Lymnaea. Injection of the specific D1-like receptor antagonist SCH23390, but not the D2 antagonist sulpiride, 10 minutes after one-trial chemical conditioning blocks the 24 hr memory trace. Because the drug was injected after conditioning, it suggests that dopamine is involved with consolidation of LTM (Kemenes et al., 2011).

Figure 3 provides a summary of published data on the known molecular requirements for the early and late phase of the consolidation of long-term memory expressed > 24 hr after single-trial food-reward classical conditioning in Lymnaea (Kemenes et al., 2002; Fulton et al., 2005; Ribeiro et al., 2005; Michel et al., 2008; Wan et al., 2010). It is interesting to note that the requirement for PKA and NO outlast the requirement for protein synthesis indicating that these molecules are required for some as yet unidentified transcription and translation independent processes underlying memory consolidation in the 1h to 6hr post-training time window. It is also unknown at present what downstream molecular mechanisms are involved in the CaMKII dependent late phase of memory consolidation.


  • Alexander, J., Audesirk, T.E. & Audesirk, G.J. (1984) One-trial reward learning in the snail Lymnaea stagnalis. J. Neurobiol. 15, 67-72.
  • Andrew, R.J. & Savage, H. (2000) Appetitive learning using visual conditioned stimuli in the pond snail, Lymnaea. Learn. Mem. 73, 258-273.
  • Benjamin, P.R. (2008) Lymnaea. Scholarpedia 3, 4124. doi:10.4249/scholarpedia.4124.
  • Benjamin, P.R. & Kemenes, G. (2009) Invertebrate models to study learning and memory: Lymnaea. Encycloped. Neurosci. 5, 197-204.
  • Benjamin, P.R., Marra, V., Kemenes, I. & O’Shea, M. (2009) Role of inhibition in associative reward conditioning of feeding in Lymnaea. Society for Neurosci. Abstracts 2009 889.5.
  • Feng, Z-P., Zhang, Z. Etc & Smit, A.B. (2009) Transcriptome analysis of the central nervous system of the mollusc Lymnaea stagnalis. BMC Genomics 10, 451 doi: 10.1186/1471-2164-10-451.
  • Fulton, D., Kemenes, I., Andrew, R.J. & Benjamin, P.R. (2005) A single time window for protein synthesis-dependent long-term memory formation after one-trial appetitive conditioning. Eur. J. Neurosci. 21, 1347-1358.
  • Fulton, D., Kemenes, I., Andrew, R.J. & Benjamin, P.R. (2008) Time window for sensitivity to cooling distinguishes the effects of hypothermia and protein synthesis inhibition on the consolidation of long-term memory. Neurobiol. Learn. Mem. 90, 651-654.
  • Haney, J. & Lukowiak, K. (2001) Context learning and the effect of context on memory retrieval in Lymnaea. Learn. & Mem. 8, 35-43.
  • Hatakeyama, D., Sadamoto, H.,Watanabe, A., et al (2006). Requirements of new protein synthesis of a transcription factor for memory consolidation; paradoxical changes in mRNA and protein of C/EBP. J. Mol. Biol. 356, 569-577.
  • Ito, E., Kobayashi, S., Kijoma, S., Sadamato, H. & Hatakeyama, D. (1999) Associative learning in the pond snail, Lymnaea stagnalis. Zool. Sci 16, 711-723.
  • Jones, N. , Kemenes, G. & Benjamin, P.R. (2001) Selective expression of electrical correlates of differential classical conditioning in a feeding network. J. Neurophysiol. 85, 89-97.
  • Jones, N.G., Kemenes, I., Kemenes G., & Benjamin, P.R. (2003) A persistent cellular change in a single modulatory neuron contributes to associative long-term memory. Curr. Biol., 13, 1064-1069.
  • Kemenes, G. & Benjamin, P.R. (1989) Appetitive learning in snails shows characteristics of conditioning in vertebrates. Brain Res. 489, 163-166.
  • Kemenes, G., Kemenes. I., Michel, M., Papp, A., & Muller, U (2006) Phase-dependent molecular requirements for memory reconsolidation: differential roles for protein synthesis and protein kinase A activity. J. Neurosci 26:6298-6302.
  • Kemenes, I., Kemenes, G., Andrew, R.J., Benjamin, P.R. & O’Shea, M. (2002) Critical time-window for NO-cGMP-dependent long-term memory formation after one-trial appetitive conditioning. J. Neurosci. 22, 1414-1425.
  • Kemenes, I., O'Shea, M. & Benjamin, P.R. (2011) Different circuit and monoamine mechanisms consolidate long-term memory in aversive and reward classical conditioning. Eur. J. Neurosci. 33, 143-152.
  • Kemenes, I., Straub, Nikitin, E.S., Staras, K., O’Shea, M., Kemenes, G., Benjamin, P.R. (2006) Role of delayed non-synaptic neuronal plasticity in long-term associative memory. Curr. Biol. 16, 1269-1279.
  • Kojima, S., Nanakamura,H., Fujito, Y. & Ito, E. (1997) Enhancement of an inhibitory input to the feeding central pattern generator in Lymnaea stagnalis during conditioned taste aversion. Neurosci. Letters. 230, 179-182.
  • Korneev, S.E., Straub, V., Kemenes, I. Korneeva, E.I., Ott, S.R., Benjamin. P.R. & O’Shea, M. (2005) Timed and targeted differential regulation of nitric oxide synthase (NOS) and antiNOS genes by reward conditioning leading to long-term memory formation. J. Neurosci. 25, 1188-1192.
  • Korneev, S.E., Kemenes, I., Bettini, N.L., Kemenes, G., Staras, K., Benjamin, P.R. & O'Shea, M.(2013) Axonal trafficking of an antisense RNA transcribed from a pseudogene is regulated by classical conditioning. Scientific Reports 3, 1027; doi:10.1038/srep01027 (2013).
  • Lowe, M.R. & Spencer, G.E. (2006) Perturbation of the activity of a single identified neuron affects long-term memory formation in a molluscan semi-intact preparation. J. Exp. Biol. 2009, 711-721.
  • Lukowiak, K. Ringseis, E., Spencer, G., Wildering, W. & Syed, N. (1996) Operant conditioning of aerial respiration in Lymnaea stagnalis. J. Exp. Biol, 199, 683-691.
  • Martens, K., Amerell, M., Parvez, K., Hittel, K., Caigny, P. de, Ito, E., & Lukowiak, K. (2007) One-trial conditioning of aerial repiratory behaviour in Lymnaea stagnalis. Neurobiol. Learn. Mem. 88, 232-242.
  • Marra, V., Kemenes, I.,Vavoulis, D.,J-F. Feng, O'Shea, M., Benjamin, P.R. (2010) Role of tonic inhibition in associative reward conditioning in Lymnaea. Frontiers in Behavioral Neuroscience 4, article 161. doi:10.3389/fnbeh.2010.00161.
  • Marra, V., O'Shea, M., Benjamin, P.R. & Kemenes, I. (2013) Susceptibility of memory consolidation during lapses in memory recall. Nature Communications 4;1578 doi: 10.1038/ncomms2591 (2013).
  • McComb, C., Sangha, S., Qadry, S.,Yue, J., Scheibenstock, A. & Lukowiak, K. (2002) Context extinction and associative learning in Lymnaea. Neurobiol. Learn. Mem. 78, 23-34.
  • Michel, M., Kemenes, I., Muller, U & Kemenes, G. (2008) Different phases of long-term memory require distinct temporal patterns of PKA activity after single-trial classical conditioning. Learn. Mem. 15, 694-702.
  • Nakamura, H., Kobayashi, S., Kojima, S., Urano, A., & Ito E (1999) PKA dependent regulation of synaptic enhancement between a buccal motor neuron and its regulatory interneuron in Lymnaea stagnalis. Zool. Sci. 16, 387–394
  • Nikitin, E. S., Kiss, T., Staras, K., O'Shea, M., Benjamin, P. R., & Kemenes, G. (2006) Persistent sodium current is a target for cAMP-induced neuronal plasticity in a state-setting modulatory interneuron. J. Neurophysiol. 95, 453-63.
  • Nikitin, E.S., Vavoulis, D.V., Kemenes, I., Marra, V., Pirger, Z., Michel, M., Feng, J-F., O’Shea, M., Benjamin, P.R. & Kemenes, G. (2008) Persistent sodium current is a non-synaptic substrate for long-term associative memory. Curr. Biol. 18, 1221-1226.
  • Orr, M.V. & Lukowiak, K. (2008) Electrophysiological and behavioral evidence demonstrating that predator detection alters adaptive behaviours in the snail Lymnaea. J. Neurosci. 28, 2726-2734.
  • Rosenegger, D., Wright, C. & Lukowiak, K. (2010) A quantitative proteomic analysis of long-term memory. Mol. Brain, 3, 9.
  • Ribeiro, M., Schofield, M., Kemenes, I., O’Shea, M., Kemenes, G., & Benjamin, P.R. (2005) Activation of MAPK is necessary for long-term memory consolidation following food-reward conditioning. Learn. Mem. 12, 538-545.
  • Ribeiro, M., Serfozo, Z., Papp, A., Kemenes, I., O’Shea, M., Yin, J.C.P., Benjamin, P.R. & Kemenes, G. (2003) Cyclic AMP response element-binding (CREB)-like proteins in a molluscan brain: cellular localization and learning-induced phosphorylation. Eur. J. Neurosci. 18, 1223-1234.
  • Sadamoto, H., Sato, H., Kobayashi, S., Murakami, J., Aonuma, H., Ando, H., Fujito, Y., Hamano, K., Awaji, M., Lukowiak, K., Urano, A. & Ito, E. (2004) CREB in the pond snail Lymnaea stagnalis: cloning, gene expression, and function in identifiable neurons of the central nervous system. J. Neurobiol. 58, 455-66.
  • Sadamoto, H., Kitahashi, T., Fujito, Y. & Ito, E. (2010) Learning-dependent gene expression of CREB1 isoforms in the molluscan brain. Front. Behav. Neurosci. 4:25. doi:10.3389/fnbeh.2010.00025
  • Sakakibara, M. (2006) Comparative study of visuo-vestibular conditioning in Lymnaea stagnalis. Biol. Bull. 210, 298-307.
  • Sangha, S., Scheibenstock, A. & Lukowiak, K. (2003a) Reconsolidation of a long-term memory in Lymnaea requires new protein and RNA synthesis and the soma of right pedal dorsal 1. J. Neurosci. 23, 8034-8040.
  • Sangha, S., Scheibenstock, A., McComb, C. & Lukowiak, K. (2003b) Intermediate and long-term memories of associative learning are differentially affected by transcription versus translation blockers in Lymnaea. J. Exp. Biol. 206, 1605-1613.
  • Sangha, S., Schiebenstock, A., Morrow, R. & Lukowiak, K, (2003c) Extinction requires new RNA and protein synthesis and the soma of the cell right pedal dorsal 1 in Lymnaea stagnalis. J. Neurosci. 23, 9842-9851.
  • Scheibenstock, A., Krygier, D., Haque, Z., Syed, N. & Lukowiak, K. (2002). The soma of RPeD1 must be present for long-term memory formation of associative learning in Lymnaea. J. Neurophysiol. 88, 1584-1591.
  • Spencer, G., Syed, N. & Lukowiak, K.(1999) Neural changes after operant conditioning of the aerial respiratory behaviour in Lymnaea. J. Neurosci. 19, 1836-1843.
  • Staras, K., Kemenes. G. & Benjamin, P.R. (1998) Neurophysiological correlates of unconditioned and conditioned feeding behaviour in the pond snail Lymnaea stagnalis. J. Neurophysiol. 79, 3030-3040.
  • Staras, K., Kemenes, G. & Benjamin, P.R. (1999) Cellular traces of behavioural classical conditioning can be recorded at several specific sites in a simple nervous system. J. Neurosci. 19, 347-357.
  • Straub, V.A., Grant, J., O’Shea, M. & Benjamin, P.R. (2007) Modulation of serotonergic neurotransmission by nitric oxide. J. Neurophysiol. 97, 1088-1099.
  • Straub, V.A., Styles, B.J., Ireland, J.S., O’Shea, M. & Benjamin, P.R. (2004) Central location of plasticity involved in appetitive conditioning in Lymnaea. Learn. Mem. 11, 787-793.
  • Vavoulis, D.V., Nikitin, E.S., Kemenes, I., Marra, V., Feng J-F., Benjamin, P.R. & Kemenes, G. (2010) Balanced plasticity and stability of the electrical properties of a molluscan modulatory interneuron after classical conditioning: a computational study. Frontiers in Behavioral Neuroscience 4, Article 19 doi:10.3389/fnbeh.2010.00019.
  • Wan, H., Mackay, B., Iqbal, H., Naskar, S. & Kemenes, G. (2010) Delayed intrinsic activation of an NMDA-independent CaM-kinase II in a critical time window is necessary for late consolidation of an associative memory. J. Neurosci. 30, 56-63.

Internal references

  • Valentino Braitenberg (2007) Brain. Scholarpedia, 2(11):2918.
  • Eugene M. Izhikevich (2006) Bursting. Scholarpedia, 1(3):1300.
  • Tamas Freund and Szabolcs Kali (2008) Interneurons. Scholarpedia, 3(9):4720.
  • Paul R. Benjamin (2008) Lymnaea. Scholarpedia, 3(1):4124.
  • Howard Eichenbaum (2008) Memory. Scholarpedia, 3(3):1747.
  • Rodolfo Llinas (2008) Neuron. Scholarpedia, 3(8):1490.
  • John Dowling (2007) Retina. Scholarpedia, 2(12):3487.
  • Wolfram Schultz (2007) Reward. Scholarpedia, 2(3):1652.
  • Kathleen Cullen and Soroush Sadeghi (2008) Vestibular system. Scholarpedia, 3(1):3013.,,
Personal tools

Focal areas