Associative Learning Acquisition and Retention Depends on ...

Neurobiology of Learning and Memory 78, 53–64 (2002) doi:10.1006/nlme.2001.4066

Associative Learning Acquisition and Retention Depends on Developmental Stage in Lymnaea stagnalis Megumi Ono,* Ryo Kawai,† Tetsuro Horikoshi,‡ Takashi Yasuoka,† and Manabu Sakakibara‡ *Graduate School of High-Technology for Human Welfare, Tokai University, 317 Nishino, Numazu 410-0321, Shizuoka, Japan; †Graduate School of Science, Tokai University, Kita-Kaname, Hiratsuka 259-1292, Kanagawa, Japan; and ‡Laboratory of Neurobiological Engineering, Department of Biological Science and Technology, School of High-Technology for Human Welfare, Tokai University, 317 Nishino, Numazu 410-0321, Shizuoka, Japan.

Associative learning dependent on visual and vestibular sensory neurons and the underlying cellular mechanisms have been well characterized in Hermissenda but not yet in Lymnaea. Three days of conditioning with paired presentations of a light flash (conditional stimulus: CS) and orbital rotation (unconditional stimulus: UCS) in intact Lymnaea stagnalis results in a whole-body withdrawal response (WBWR) to the CS. In the current study, we examined the optimal stimulus conditions for associative learning, including developmental stage, number of stimuli, interstimulus interval, and intertrial interval. Animals with a shell length longer than 18 mm (sexually mature) acquired and retained the associative memory, while younger ones having a shell length shorter than 15 mm acquired but did not retain the memory to the following day. For mature animals, 10 paired presentations of the CS and UCS presented every 2 min were sufficient for the induction of a WBWR to the CS. Furthermore, animals conditioned with the UCS presented simultaneously with the last 2 s of the CS also exhibited a significant WBWR in response to the CS. Blind animals did not acquire the associative memory, suggesting that ocular photoreceptors, and not dermal photoreceptors, detected the CS. These results show that maturity was key to retention of associative learning. 䉷 2002 Elsevier Science (USA)

This study was supported by Grants-in-Aid (11168231, 12680783) for Scientific Research, the Ministry of Education, Science, and Culture of Japan, to M.S. and by the Proposed-Based New Industry Creative Type Technology R and D Promotion Program from the New Energy and Industrial Technology Development Organization (NEDO) of Japan in the field of biocybernetics (98S18-001-2) to M.S. Address correspondence and reprint requests to Manabu Sakakibara, Ph.D., Laboratory of Neurobiological Engineering, Department of Biological Science and Technology, School of High-Technology for Human Welfare, Tokai University, 317 Nishino, Numazu 410-0321, Shizuoka, Japan. Fax: ⫹81-55-968-1156. E-mail: sakaki@ fb.u-tokai.ac.jp. 53

1074-7427/02 $35.00 䉷 2002 Elsevier Science (USA) All rights reserved.

54

ONO ET AL.

Key Words: Lymnaea; associative learning; whole-body withdrawal response; learning paradigm; development; interstimulus interval; intertrial interval.

INTRODUCTION Gastropods are well established as models for the study of the molecular and cellular bases of learning and memory. Animals such as Aplysia (Carew, Walters, & Kandel, 1981), Limax (Culligan & Gelperin, 1983), Hermissenda (Alkon, 1974), and Lymnaea (Mogilevskii & Verbnyi Ya, 1994) exhibit well-characterized associative learning. Hermissenda crassicornis is well known to be capable of associative learning of the sequential events of light and rotation. While Lymnaea stagnalis is a well-established animal model exhibiting appetitive classical conditioning (Kojima, Yamanaka, Fujito, & Ito, 1996), our previous study demonstrated that Lymnaea are also capable of classical conditioning pairing light flashes (conditional stimulus: CS) and orbital rotation (unconditional stimulus: UCS) (Sakakibara, Kawai, Kobayashi, & Horikoshi, 1998). Naive Lymnaea do not respond to a flash of light but do respond to a moving shadow, light-off, or orbital rotation with a whole-body withdrawal response (WBWR). After 30 paired presentations of a light flash and orbital rotation for 3 days, the animals exhibit a WBWR (unconditioned response: UCR) to the CS. Lymnaea to which light and rotation stimuli are presented out of phase exhibit behavior that is significantly different from that of those receiving paired stimuli. This form of associative learning in Hermissenda is dependent on interactions between visual and vestibular sensory neurons (Tabata & Alkon, 1982), and the cellular mechanism involved in this learning has been well characterized (Alkon, 1987). The cellular mechanism underlying this form of associative learning in Lymnaea, however, has not yet been studied. To begin to characterize the underlying mechanisms of associative learning in Lymnaea, the current study focused on defining the optimal learning conditions in terms of the developmental stage of the animals, timing of the CS and UCS presentation, and number of pairings as well as on determining the type of photoreceptors involved. Lymnaea have been shown to acquire a conditioned taste aversion that is retained for more than 30 days even at Postnatal Day (PND) 45 (immature: 10 mm shell length) (Yamanaka et al., 1999). This finding suggests that the central nervous system at this stage is sufficiently developed for learning acquisition and memory storage, raising the possibility that the central nervous system of animals having a shell length of 10 mm are also sufficiently developed for the acquisition of associative learning between a flash of light and orbital rotation. The interstimulus interval between the CS and UCS is assumed to be an important property of associative learning in vertebrates (Hoehler & Thompson, 1980) and invertebrates (Lederhendler & Alkon, 1989; Matzel, Schreurs, Lederhendler, & Alkon, 1990). In Hermissenda, the adequate interstimulus interval correlates well with the architecture and activity of photoreceptors and the statocyst hair cells (Lederhendler & Alkon, 1989; Matzel et al., 1990). In our previous study (Sakakibara et al., 1998), we showed that associative learning is observed in Lymnaea using parameters for Hermissenda, but the optimal parameters for Lymnaea have not yet been examined. Recent results indicate that Lymnaea exhibit appetitive classical conditioning with

ASSOCIATIVE LEARNING IN Lymnaea

55

paradigms using visual cues such as a checkerboard (Andrew & Savage, 2000). This raises the possibility that Lymnaea have well-developed eyes and may be capable of form vision. Consistent with this, the number of photoreceptors in Lymnaea eyes is much greater than that in Hermissenda eyes (Stoll & Bijlsma, 1973). Furthermore, Andrew and Savage (2000) raised the question of whether associative learning of light and orbital rotation in Lymnaea involved ocular photoreceptors or dermal photoreceptors. It has been shown that the shadow- or light-off-induced WBWR is mediated via dermal photoreceptors rather than ocular photoreceptors (Stoll, 1972). Only the optic nerve responds to light being switched on (Stoll & Bijlsma, 1973), and dermal photoreceptors are responsible for nonocular orientation behavior (Van Duivenboden, 1982). MATERIALS AND METHODS The materials and methods have been described in detail previously (Sakakibara et al, 1998). In this section, we briefly summarize the methods and describe the differences. Animals Laboratory-reared freshwater pond snails, Lymnaea stagnalis, with shell lengths between 10 and 23 mm were maintained at 20⬚C in well-aerated water, on a 12-h light:12-h dark cycle (on at 08:00) and fed cabbage and goldfish pellets. The relation between shell length and developmental stage has been well established as follows: (a) immature (PND 45–75), shell length of 10 to 15 mm; (b) sexually mature (PND 75–100), shell length of 15 to 20 mm; and (c) adult, shell length longer than 20 mm (Yamanaka et al., 1999). Four animals were trained concurrently, and before training, animals were dark adapted for 15 min. Each animal was kept in a freshwater-filled flask (70 ml) fixed on an orbital shaker (Multi-Mixer type 4600, Lab-Line, Melrose, IL) and placed in a sound- and lightproof incubator (IN800, Yamato, Tokyo) maintained at 20⬚C. To identify the sensory pathway involved in perception of the CS, 10 mature animals were anesthetized with ethanol and had both eyes enucleated with forceps. Blind animals were used for experimentation 1 week after enucleation. Stimulation The CS was delivered using a halogen tungsten lamp source of 50 W (HL50E, HoyaSchott, Tokyo) and guided into an incubator using a fiber-optic cable positioned 35 cm above the animals. An electromechanical shutter (EC-601, Copal, Tokyo), equipped in a lamp house controlled the timing and duration of the light stimulus. The energy of unattenuated light was 700 ␮W/cm2 at 510 nm at the level of the animals. The UCS was delivered using an orbital shaker that produced a 4.0-mm orbital motion at 1960 rpm. Both the timing and the duration of the CS and UCS were controlled using a microcomputer. Training and Evaluation of Associative Learning Experimental animals (conditioned groups, n ⫽ 107) were trained with paired presentations of the CS and UCS, while a control group received explicitly unpaired presentations

56

ONO ET AL.

of the CS and UCS (pseudorandom group, n ⫽ 11). A second control group (naive group, n ⫽ 11) was included in which animals were maintained in a flask on an orbital shaker in an incubator without any stimulus presentations for the same time as the conditioned groups. Animals in the conditioned and pseudorandom groups received the same number of light and orbital rotation presentations, but the timing between the two stimuli was different for the two groups. In the conditioned groups, each animal was presented with a 3-s flash of light. During the last 2 s of the light flash, they were mechanically rotated. The training period consisted of a total of 30 paired stimuli (intertrial interval ⫽ 2 min) presented each day for 3 days. We used the same interstimulus and intertrial intervals as in our previous study (Sakakibara et al., 1998), as was previously established for Hermissenda (Lederhendler & Alkon, 1989). By contrast, animals in the pseudorandom group never experienced simultaneous presentations of light and duration. The behavior of each animals was monitored using a near-infrared sensitive CCD camera (␭max ⫽ 780 nm, CS3450 Tokyo Electronic Indust., Tokyo) under the illumination of an infrared light placed in the incubator. Data were collected and stored on videotape, and timing signals generated using a video timer (VTG-33, Hoei, Tokyo) were superimposed on the recording, accurate to 0.01 s. The WBWR to the presentation of light was recorded three times at 2-min intervals. On days including training sessions, the WBWR to light was tested both immediately before and after the training sessions. Animals that did not exhibit a WBWR to the light stimulus before training were selected to be part of the conditioned group. Following each training session, we categorized animals as “wellconditioned animals” if they responded two or more times to three presentations of the CS. The WBWR latency (the time between the onset of light and the start of the WBWR) was used as a measure of conditioning because both response amplitude and frequency have been shown to decline with stimulus repetition, while the response latency remains constant (Cook, 1970). In the current study, the response latency was measured using the time from the video timer superimposed on the video recording of the WBWR. The start of the WBWR was defined as the video frame in which a certain body feature disappeared into the shell. To facilitate presentation of the data, behavioral scores are graphed using the reciprocal of the response latency values; thus, a response latency of 0.5 s is expressed as 2. The behavioral measures were performed using blind test procedures; the behavioral manipulation was unknown to those analyzing the behavioral scores. Averages of scores obtained after the training session were used in all data analyses except when specific pre- and posttraining comparisons were made. Statistical analyses of the data were performed with a Student’s t test using the data analysis and graphic software ORIGIN (Microcal Software, Northampton, MA). A schematic diagram of the experimental setup is provided in Fig. 1. RESULTS Developmental Stage and Acquisition of Learning We first compared learning ability among the three developmental groups: immature, mature, and adult. Figure 2 illustrates the relation of successive training days and behavioral score among the conditioned mature, conditioned adult, pseudorandom, naive, and blind groups. Because


57

FIG. 1. Schematic illustration of the experimental setup.

only 1 of 31 immature animals responded, data from the immature group are not included in the Figure. The scores of animals at different developmental stages are averaged values obtained from well-conditioned animals. Pseudorandom, naive, and blind animals were selected from the mature group with shell lengths of either 18 or 20 mm. The results of statistical comparisons of scores presented in Fig. 2 and recorded on Day 4 (1 day after training) from conditioned mature, conditioned adult, pseudorandom, and naive animals are shown in Table 1. The scores from immature animals are not included because only 1 of 31 animals retained the memory on Day 4 (Fig. 3). The posttraining score of immature animals was always greater than the pretraining score recorded on each training day (12 averaged behavioral pre- and posttraining scores: 0 to 6.3 for Day 1, 0.8 to 4.5 for Day 2), suggesting that even immature animals had the ability to acquire associative learning. The score, however, was not maintained to the next day (Fig. 3). This suggests that immature animals with shell length shorter than 15 mm acquired the associative learning but did not retain the memory. On the other hand, for mature animals, the posttraining score after the 2nd day was significantly greater than that of both the 1st day ( p ⬍ .005) and pretraining ( p ⬍ .005). Mature animals with 18-mm shell length responded to the CS more than twice, on average, on Day 4 (Table 2). These differences were also significant ( p ⬍ .005) for adult animals (Fig. 2 and Table 1). Furthermore, the behavioral scores for the mature and adult animals increased day by day, indicating memory retention (Fig. 2 and Table 2). There was no significant difference between these two groups (Table 1). Among the three developmental groups, the behavioral score was the highest for the mature animals (Fig. 2). There was no significant difference between animals within this group having shell lengths of 18 or 20 mm. In contrast to the behavior of conditioned mature animals, naive animals with shell lengths of 18 or 20 mm exhibited little response to the CS. Furthermore, none of the blind mature animals responded to the CS after the training paradigm. The behavioral response to shadow in the blind animals was normal, however, suggesting that the dermal photoreceptors, which are responsible for nonocular orientation behavior, remained intact, as has been previously shown for enucleated animals (Stoll, 1972). This result clearly demonstrates that the ocular photoreceptors, and not

58

ONO ET AL.

FIG. 2. The relation of training days and behavioral scores among the conditioned mature, conditioned adult, pseudorandom, naive, and blind groups. Because only 1 of 31 immature animals responded, data from the immature group are not included. The behavioral score is the reciprocal of the response latency to the CS. The scores shown are the posttraining scores from Days 1 to 3 and the postconditioning score from Day 4. Scores for the conditioned groups were obtained from well-conditioned animals (those responding two or more times to three CS presentations each day). Conditioned mature animals scored significantly higher after training than did naive mature animals each day of training and on Day 4, as shown in Table 1. Blind animals did not exhibit any behavior in response to the CS. The behavioral scores of conditioned mature and conditioned adult animals were significantly higher than those of pseudorandom and naive animals. There was no significant difference between conditioned mature and conditioned adult animals. ***p ⬍ .005.

dermal photoreceptors, are involved in detection of the CS. Thus, for the remaining experiments, we used mature animals with shell length of 18 or 20 mm that had not been enucleated.

The Interstimulus Interval In a previous study (Sakakibara et al., 1998), we demonstrated that associative learning can be established using a 3-s light flash (CS) and 2 s of orbital rotation (UCS) with cotermination. In the current study, we examined the efficacy of different UCS presentation times in relation to the CS (Fig. 4). CS–UCS pairings in which the CS preceded the onset of the UCS and terminated with the offset of the UCS evoked a significantly ( p ⬍ .05 [Day 1]) and p ⬍ .005 [after Day 2]) strong UCR as compared to either a CS that preceded the UCS and terminated with its onset or a CS and UCS that started simultaneously and the CS terminated before the termination of the UCS. This result indicates that CS–UCS contiguity as well as the initial interstimulus interval act additively to establish the CS–UCS association in Lymnaea, as has been observed in Hermissenda (Matzel et al., 1990).

59


TABLE 1 Student’s t Test Comparisons of Scores from the Conditioned Mature, Conditioned Adult, Pseudorandom, and Naive Groups Comparison Mature versus naive Mature versus pseudorandom Mature versus adult Adult versus naive Adult versus pseudorandom Pseudorandom versus naive

Statistic t t t t t t

⫽ ⫽ ⫽ ⫽ ⫽ ⫽

⫺6.52492 ⫺6.70952 ⫺1.8807 6.26347 6.60822 1.78861

Significance level p⬍ p⬍ NS p⬍ p⬍ NS

.005 .005 .005 .005

Note. The behavioral scores from the mature conditioned versus the naive and pseudorandom groups were significantly different, and no difference was observed between the conditioned mature and conditioned adult groups. The data from immature animals are not included because only 1 of 31 animals responded to the CS at Day 4. NS, not significant.

The Intertrial Interval In our previous study in Lymnaea (Sakakibara et al., 1998), we demonstrated that massed training (45 pairings/day for 2 days and 90 pairings/day for 1 day) is ineffective for associative learning, while spaced training of 30 pairings/day for 3 days is quite effective. In the current study, we examined the efficacy of 0.5-, 2-, and 8-min intertrial intervals within a day of 30 pairings (Fig. 5). Paired presentation at 2-min intervals resulted in significant ( p ⬍ .01 [Day 1] and p ⬍ .005 [after Day 2]) conditioning and retention at Day 4. The other conditions were not effective. The observed behavioral difference may reflect the differences in time spent in a flask with the 0.5-, 2- and 8-min intertrial interval training paradigms corresponding to 1, 1.5, and 2.25 h, respectively. To exclude this possibility, 11 animals receiving the 2-min intertrial interval paradigm remained in the flask in a quiet and dark incubator for an additional 45 min, making the total time the same as for animals receiving the 8-min

FIG. 3. The ratios of well-conditioned animals to all other animals with the same shell length at Day 4. The numbers of animals responding two or more times to three presentations of the CS each day are shown in white. Note that no immature animals with 10-mm shell length responded to light consistently; immature animals responded to only one of three CS presentations.

60

ONO ET AL.

TABLE 2 Mean Behavioral Scores from 7 Mature Animals with a Shell Length of 18 mm in Response to the CS (three presentations) Mean ⫾ SD Day1 Day2 Day3

Pre Post Pre Post Pre Post

Day4

0.67 1.04 1.22 0.79 1.29 2.13

0 ⫾ ⫾ ⫾ ⫾ ⫾ ⫾

0.32 0.86 1.19 0.68 0.88 0.82

Note. On Day 4, animals were presented with the CS without any presentation of paired stimuli.

intertrial interval paradigm. These animals exhibited no obvious behavioral differences as compared to the other animals receiving the 2-min intertrial interval paradigm. The Number of Pairings and Acquisition of Learning In our previous study (Sakakibara et al., 1998), we presented 30 CS and UCS pairs in a day, as previously done for Hermissenda (Alkon, 1974). Thus, in the current study, we examined the optimal number of paired presentations per day for learning for mature Lymnaea. The posttraining behavioral scores on Day 3 are shown in Fig. 6 for five groups in which the respective total paired presentations per day were 1, 5, 10, 20, or 30. Animals trained with 10 paired presentations per day exhibited the highest behavioral

FIG. 4. The relation of the interstimulus delay and the behavioral score. Only the training paradigm that included the cotermination of the CS and UCS resulted in a significant increase in the posttraining behavioral score after 3 days and memory retention to the following day. The inset shows three different timings of presentation of the UCS. ST, simultaneous termination; SS, simultaneous start; D, delayed. The statistical comparison was made between ST and SS. *p ⬍ .05, ***p ⬍ .005.


61

FIG. 5. The relation of the intertrial interval and the behavioral score. Animals exhibited significant conditioning after training with paired presentations with 2-min intertrial intervals, whereas animals receiving training with other intertrial intervals did not. The statistical comparison was between the 2- and 0.5-min intertrial interval paradigm conditions. **p ⬍ .01, ***p ⬍ .005.

score, significantly greater ( p ⬍ .005) than those with either 1 or 5 paired presentations per day (Fig. 6 and Table 3). DISCUSSION The current study clearly demonstrates that the WBWR in response to the CS (a flash of light) in conditioned Lymnaea involved ocular photoreceptors, and not dermal

FIG. 6. The relation of the number of pairings and the posttraining behavioral score at Day 3. Statistical comparisons are shown in Table 3.

62

ONO ET AL.

TABLE 3 Student’s t Test Comparisons among Different Numbers of Pairings (1, 5, 10, 20, or 30 trials/day) for 3 Days of Training Comparison 1 1 1 1 5 5 5 10 10 20

versus versus versus versus versus versus versus versus versus versus

30 20 10 5 30 20 10 30 20 30

Statistic t t t t t t t t t t

⫽ ⫽ ⫽ ⫽ ⫽ ⫽ ⫽ ⫽ ⫽ ⫽

2.97485 2.85711 5.08161 1.15542 2.33432 2.12167 4.46598 3.63834 2.87067 0.37553

Significance level p⬍ p⬍ p⬍ NS p⬍ p⬍ p⬍ p⬍ p⬍ NS

.005 .01 .005 .05 .05 .005 .005 .01

Note. NS, not significant.

photoreceptors, because enucleated animals did not exhibit any behavioral response to the CS, and the delayed training paradigm, in which light-off and the start of rotation occurred simultaneously, did not produce significant conditioning. The neuronal circuitry involved in this conditioning is at least partially different from the neuronal circuitry involved in the light-off WBWR (shadow response) in which the sensory input has been suggested to be via dermal photoreceptors (Stoll, 1972; Stoll, 1973; Stoll & Bijlsma, 1973). We also found that none of the animals with a 10-mm shell length and only 1 of 18 with a 15-mm shell length (both immature) retained the associative memory through 4 days. The results that improvement of the behavioral score occurred after each day in the training paradigm suggested that even immature animals acquired the learning but did not retain the associative memory. It was not until the animals had matured further that they acquired and retained the associative memory. The WBWR has been shown to be mediated by the collummelar and dorsal longitudinal muscles, both of which are innervated by the motoneuron network in the seven ganglia of the central nervous system, particularly the cerebral and pedal ganglia (Ferguson & Benjamin, 1991). Although the motoneuron network involved in the CR appeared to be developed in the immature animals, the neural circuitry necessary for memory retention did not appear to be fully developed. During memory formation, a modification of the neural circuit controlling the WBWR would be expected to take place at the synapse between the sensory neuron and the motoneuron, but some element of this neural circuit is not fully developed in immature animals. This finding is in contrast to a previous study of conditioned taste aversion, which even immature Lymnaea learn (Yamanaka et al., 1999). This may be because a different neural circuit is involved in the conditioning, especially for the neural circuit governing the associative memory retention. All of the neurons forming the motoneuron network are electrically coupled, and the coupling ratio changes with age, decreasing from 60% at PND 90 (immature) to 30% at PND 400 (adult) (Wildering, van der Roest, de Vlieger, & Janse, 1991). This decrease may stabilize associative memory formation and may reflect a corresponding structural change similar to “focusing,” a decrease in terminal branching volume observed at the type B cell in Hermissenda (Alkon et al., 1985).


63

A previous report indicated that shell length and neural development are well correlated if the shell breadth:length ratio is 0.45 ⫾ 0.05, but both neural development and shell size vary greatly even between animals of similar ages (Croll & Chiasson, 1989). The authors divided the developing animals into four categories to facilitate the description of postembryonic development: Hatchlings had 1.4- to 2.5-mm long shells, juveniles had 2.6- to 15.9-mm shells (PND ⬍ 100), young adults had 16- to 23-mm shells (PND ⫽ 125), and adults had 35-mm shells (PND ⬎ 200). The shell breadth:length ratios in the current study were well within the ranges they proposed: immature animals, 0.48 ⫾ 0.03 (n ⫽ 46); mature animals, 0.52 ⫾ 0.04 (n ⫽ 88); and adult animals, 0.55 ⫾ 0.05 (n ⫽ 19). Although Croll and Chiasson (1989) argued that the variability of neuronal development is large, one clear division shown by the results of the current study is that animals having shell length shorter than 15 mm are not capable of retaining associative memories. The neuronal correlates and conditioning parameters, such as interstimulus delay and interval, have been well studied for Hermissenda (Grover & Farley, 1987; Lederhendler & Alkon, 1989; Matzel et al., 1990), revealing that it is important for the onset of the UCS to precede the onset of the CS by 1 s and the UCS and CS to terminate simultaneously for effective learning of suppression of phototactic behavior. This timing is consistent with the dynamic suppression of type B photoreceptors by both statocyst hair cells and S/E interneurons in the optic ganglion. This generates a long-lasting depolarization, and the B cells remain in an excitatory state much longer, resulting in phototactic suppression. On the other hand, although little is known of the physiology or morphology mediating associative learning of light and rotation in Lymnaea, the results of the current study demonstrate that a 2-min intertrial interval with a 1-s delay of the UCS onset and cotermination with the CS was in good agreement with the optimal parameters used in studies of Hermissenda. This raises the possibility that the same neuronal pathway is involved in Lymnaea conditioning. Contemporary models of associative learning specify that long intertrial intervals (e.g., 8 min) support better acquisition (Gallistel & Gibbon, 2000). By contrast, the current study shows that a 2-min intertrial interval resulted in better acquisition than an 8-min intertrial interval. The mechanism underlying this result is currently unknown and is open for future study. REFERENCES Alkon, D. L. (1974). Associative training of Hermissenda. Journal of General Physiology, 64, 70–84. Alkon, D. L. (1987). Memory traces in the brain. New York: Cambridge Univ. Press. Andrew, R. J., & Savage, H. (2000). Appetitive learning using visual conditioned stimuli in the pond snail, Lymnaea. Neurobiology of Learning and Memory, 73, 258–273. Carew, T. J., Walters, E. T., & Kandel, E. R. (1981). Associative learning in Aplysia: Cellular correlates supporting a conditioned fear hypothesis. Science, 211, 501–504. Cook, A. (1970). Habituation in a freshwater snail (Lymnaea stagnalis). Animal Behavior, 178, 463–474. Croll, R. P., & Chiasson, B. J. (1989). Postembryonic development of serotonin-like immunoreactivity in the central nervous system of the snail, Lymnaea stagnalis. Journal of Comparative Neurology, 280, 122–142. Culligan, N, & Gelperin, A. (1983). One-trial associative learning by an isolated molluscan CNS: Use of different chemoreceptors for training and testing. Brain Research, 266, 319–327. Ferguson, G. P., & Benjamin, P. R. (1991). The whole-body withdrawal response of Lymnaea stagnalis. I. Identification of central motoneurones and muscles. Journal of Experimental Biology, 158, 63–95.

64

ONO ET AL.

Gallistel, C. R., & Gibbon, J. (2000). Time, rate, and conditioning. Psychological Review, 107, 289–344. Grover, L. M., & Farley, J. (1987). Temporal order sensitivity of associative neural and behavioral changes in Hermissenda. Behavioral Neuroscience, 101, 658–675. Hoehler, F. K., & Thompson, R. F. (1980). Effect of the interstimulus (CS–UCS) interval on hippocampal unit activity during classical conditioning of the nictitating membrane response of the rabbit (Oryctolagus cuniculus). Journal of Comparative Physiological Psychology, 94, 201–215. Kojima, S., Yamanaka, M., Fujito, Y., & Ito, E. (1996). Differential neuroethological effects of aversive and appetitive reinforcing stimuli on associative learning in Lymnaea stagnalis. Zoological Science, 13, 803–812. Lederhendler, I. I., & Alkon, D. L. (1989). The interstimulus interval and classical conditioning in the marine snail Hermissenda crassicornis. Behavioral Brain Research, 35, 75–80. Matzel, L. D., Schreurs, B. G., Lederhendler, I., & Alkon, D. L. (1990). Acquisition of conditioned associations in Hermissenda: Additive effects of contiguity and the forward interstimulus interval. Behavioral Neuroscience, 104, 597–606. Mogilevskii, A., & Verbnyi Ya, I. (1994). Influence of different intracellular electrostimulation regimes on the dynamics of the adaptational processes of neurons. Neuroscience and Behavioral Physiology, 24, 386–393. Sakakibara, M., Kawai, R., Kobayashi, S., & Horikoshi, T. (1998). Associative learning of visual and vestibular stimuli in Lymnaea. Neurobiology of Learning and Memory, 69, 1–12. Stoll, C. J. (1972). Sensory systems involved in the shadow response of Lymnaea stagnalis. Proceedings of the Koninklijke Nederlands Akademie van Wetenschappen, C75, 342–351. Stoll, C. J. (1973). On the role of eyes and non-ocular light receptors in orientational behaviour of Lymnaea stagnalis (L.). Proceedings of the Koninklijke Nederlands Akademie van Wetenschappen, C76, 203–214. Stoll, C. J., & Bijlsma, A. (1973). Optic nerve responses in Lymnaea stagnalis to photic stimulation of the eye. Proceedings of the Koninklijke Nederlands Akademie van Wetenschappen, C76, 406–413. Tabata, M., & Alkon, D. L. (1982). Positive synaptic feedback in visual system of nudibranch mollusk Hermissenda crassicornis. Journal of Neurophysiology, 48, 174–191. Van Duivenboden, Y. A. (1982). Non-ocular photoreceptors and photo-orientation in the pond snail Lymnaea stagnalis. Journal of Comparative Physiology, 149, 363–368. Wildering, W. C., van der Roest, M., de Vlieger, T. A., & Janse, C. (1991). Age-related changes in junctional and non-junctional conductances in two electrically coupled peptidergic neurons of the mollusc Lymnaea stagnalis. Brain Research, 547, 89–98. Yamanaka, M., Sadamoto, H., Hatakeyama, D., Nakamura, H., Kojima, S., Kimura, T., Yamashita, M., Urano, S., & Ito, E. (1999). Developmental changes in conditioned taste aversion in Lymnaea stagnalis. Zoological Science, 16, 9–16.