Habituation

In habituation, response decrement occurs when stimuli are repeatedly presented; this decreasing response to recurring stimuli allows filtering out redundant information and freeing up attention for higher cognitive functions (Ramaswami, Reference Ramaswami2014; Rankin et al., Reference Rankin, Abrams, Barry, Bhatnagar, Clayton, Colombo and Thompson2009). The behavioral rules for habituation are conserved across phylogeny (Groves & Thompson, Reference Groves and Thompson1970; Rankin et al., Reference Rankin, Abrams, Barry, Bhatnagar, Clayton, Colombo and Thompson2009); however, this conservation does not mean its mechanisms are also conserved. Studies with the gill-withdrawal reflex in Aplysia, a mollusk, found that habituation was mediated by depression of an excitatory synapse (Pinsker et al., Reference Pinsker, Kupfermann, Castellucci and Kandel1970), while studies of olfactory habituation in Drosophila, an arthropod, indicated it was mediated by enhancement of an inhibitory synapse (Das et al., Reference Das, Sadanandappa, Dervan, Larkin, Lee, Sudhakaran and Ramaswamia2011). These contrasting findings emphasize how little we understand the cellular and molecular mechanisms for habituation in any organism (Rankin et al., Reference Rankin, Abrams, Barry, Bhatnagar, Clayton, Colombo and Thompson2009).

Although many responses habituate in C. elegans, habituation to mechanosensory stimuli has been the most thoroughly scrutinized. A tap to the side of a Petri dish on which worms are cultured typically elicits backward movement; the magnitude of this reversal response decreases over repetition of stimuli (Rankin et al., Reference Rankin, Beck and Chiba1990). Detailed analyses of how different factors influence habituation has led to insights into its cellular mechanisms.

Timbers et al. (Reference Timbers, Giles, Ardiel, Kerr and Rankin2013) varied tap intensity and observed faster habituation to weaker taps and slower to stronger taps. Furthermore, older worms habituated to taps of the same intensity faster than younger worms. Interestingly, when the touch receptor neurons were activated with blue light by genetically engineering the expression of light-gated cation channels channelrhodopsin-2 (ChR2), and mechanosensory transduction was bypassed, older and younger worms did not differ in habituation rate. Timbers et al. (Reference Timbers, Giles, Ardiel, Kerr and Rankin2013) hypothesized that older worms experience stimuli as weaker, possibly because aging-related cuticle thickening decreases the sensitivity of the mechanoreceptors, and habituation is consequently altered.

Habituation is also affected by interstimulus interval (ISI). Worms trained at shorter ISIs habituated, and spontaneously recovered from habituation, more rapidly than worms trained at longer ISIs (Rankin & Broster, Reference Rankin and Broster1992). This led to the hypothesis that habituation at different ISIs may be mediated, at least in part, by different processes. This hypothesis was substantiated by studies showing that there are genes differentially mediating habituation at different ISIs (Ardiel et al., Reference Ardiel, McDiarmid, Timbers, Lee, Safaei, Pelech and Rankin2018). Two genes, cmk-1 and ogt-1, the C. elegans orthologs of human CAMK1 (calcium/calmodulin-dependent protein kinase I) and OGT (O-linked N-acetylglucosamine transferase), function in the touch receptor neurons to modulate habituation in an ISI-dependent manner: cmk-1 and ogt-1 mutants habituated faster than wild-type worms at a 10 s ISI, and slower than wild-type worms at a 60 s ISI. Thus, different molecular signaling cascades may be involved in short- and long-ISI habituation. It may be that habituation at long ISIs is more energetically costly and so it is better to retain the memory for longer durations, while habituation at short ISIs may be energetically less expensive.

Historically, habituation has been viewed as an equivalent decrement of all components of a response, however, this is not always the case. Timbers et al. (Reference Timbers, Giles, Ardiel, Kerr and Rankin2013) compared habituation for response probability and distance travelled, and found that aging differentially affected habituation of these two components. Ardiel et al. (Reference Ardiel, McDiarmid, Timbers, Lee, Safaei, Pelech and Rankin2018) further dissected “distance” into speed and duration components. Worms with mutations in cmk-1 and ogt-1 had similar habituation curves for response distance; however, examining speed and duration separately revealed that cmk-1 mutants had a higher response speed and a lower response duration than ogt-1 mutants, and the two components showed different habituation dynamics. These findings suggest that the response components of habituation might be modulated by different underlying mechanisms.

Consistent with findings in other species, long-term memory (LTM) is most effectively produced by spaced/distributed training. LTM for habituation was observed 24 hours after 4 blocks of 20 taps at a 60 s ISI separated by 1 hour resting periods (Beck & Rankin, Reference Beck and Rankin1995; Rose et al., Reference Rose, Kaun and Rankin2002). In a later study, memory from this training lasted at least 48 hours, a remarkably long time given the short lifespan of C. elegans. By contrast, massed training with the same number of stimuli, 80 taps, at a 60 s ISI with no rest periods produced an intermediate-term memory of habituation that was exhibited 12 hours but not 24 hours later (Li et al., Reference Li, Timbers, Rose, Bozorgmehr, McEwan and Rankin2013).

Distinct molecular mechanisms underlie memories for habituation lasting different durations and produced by different training protocols. In short-term habituation, the mechanosensory neurons and the sensory-interneuron synapses were hypothesized to be the major cellular sites of habituation (Wicks & Rankin, Reference Wicks and Rankin1995), and, because the mechanosensory neurons are glutamatergic, a role for glutamate at this locus was implicated (Rankin & Wicks, Reference Rankin and Wicks2000). Intermediate-term memory for habituation was hypothesized to be caused by an increase in FLP-20, a FMRFamide-like neuropeptide, released by the mechanosensory neurons; this intermediate-term memory was correlated with an increase in synaptic vesicle density in the so-called PLM mechanosensory neurons (and likely the ALM neurons; Li et al., Reference Li, Timbers, Rose, Bozorgmehr, McEwan and Rankin2013), suggesting an increase in FLP-20-containing vesicles recruited to the presynaptic terminals. By contrast, the molecular mechanisms for LTM, as in many other species, involves protein-synthesis-dependent changes in glutamate signaling (Rose et al., Reference Rose, Kaun, Chen and Rankin2003) and activity of a transcription factor, cAMP response element-binding protein (CREB; Timbers & Rankin, Reference Timbers and Rankin2011). This memory is eliminated by protein synthesis inhibition (heat shock; Beck & Rankin, Reference Beck and Rankin1995) and reconsolidation blockade (Rose & Rankin, Reference Rose and Rankin2006). These data indicate that there is not a single “memory” process, but that memory for short-, intermediate-, and long-term habituation can be produced with different paradigms and are mediated by nonoverlapping cellular and molecular signaling pathways, perhaps to minimize interference among different forms of memory.

Environmental factors also affect mechanosensory habituation. Kindt et al. (Reference Kindt, Quast, Giles, De, Hendrey, Nicastro and Schafer2007) found that C. elegans habituated more rapidly and to a greater degree when trained without food (E. coli patches) on the Petri plates. This occurs because food texture triggers dopamine release, which activates a dopamine D1-like receptor DOP-1 in the touch receptor neurons to modulate mechanosensation and delay habituation. One possible explanation is that the presence of food helps worms remain alert to incoming stimuli, and the absence of food reduces responding to save energy and prioritize foraging.

Environmental contextual cues can add an associative component to habituation. Worms habituated with a distinct chemosensory cue, either the taste of salt or the odor of a volatile chemical, retained the memory better if they were rehabituated later in the same context (Lau et al., Reference Lau, Timbers, Mahmoud and Rankin2013). The effect of the context on habituation memory is mediated by an NMDA-type glutamate receptor subunit NMR-1, as nmr-1 mutants showed normal short- and long-term habituation, but failed to display enhanced memory in the presence of the contextual reminder. Thus, nonassociative learning can have associative components, suggesting that the configuration of environmental cues and stimulation patterns may convey specific information that animals store in memory for future use.

Habituation has also been studied in chemosensation in C. elegans. Decreased responding to repeated chemosensory stimuli could depend on habituation, sensory adaptation, or both. Response decrement by adaptation only recovers over time; this sensory fatigue is mediated by different cellular processes than habituation. Habituation is an attentional process that can be differentiated from adaptation by testing for dishabituation and by determining the rate of spontaneous recovery after training (Rankin et al., Reference Rankin, Abrams, Barry, Bhatnagar, Clayton, Colombo and Thompson2009).

Bernhard and van der Kooy (Reference Bernhard and van der Kooy2000) dissociated chemosensory adaptation and habituation in C. elegans. Preexposure to 0.001% or to 100% diacetyl led to a decreased approach response to diacetyl (normally an attractant). If this exposure was followed by being spun in a centrifuge, only worms preexposed to 0.001% diacetyl showed dishabituation, suggesting the decreased approach was habituation, whereas worms preexposed to 100% diacetyl did not dishabituate, suggesting the decrement was adaptation. Thus, habituation primarily mediated decrement at low concentration, whereas adaptation occurred at high concentration. Habituation to diacetyl smell is glr-1 dependent (Bernhard & van der Kooy, Reference Bernhard and van der Kooy2000); by contrast, habituation to benzaldehyde smell is not (Morrison & van der Kooy, Reference Morrison and van der Kooy2001). Additionally, worms habituated to benzaldehyde could be dishabituated by centrifugation, and this phenomenon was also glutamate dependent, as mutants lacking a vesicular glutamate transporter do not exhibit dishabituation (Nuttley et al., Reference Nuttley, Harbinder and van der Kooy2001).

Like mechanosensory habituation, chemosensory habituation is also affected by the presence of food. Nuttley et al. (Reference Nuttley, Atkinson-Leadbeater and van der Kooy2002) found that habituation to benzaldehyde was inhibited in the presence of food. Exogenous serotonin mimicked the inhibitory effect of food on habituation to benzaldehyde, and the presence of food had no effect on the habituation in serotonin-deficient cat-4 and tph-1 mutants. These studies illustrate that convergent and divergent molecular mechanisms underlie habituation within and across sensory modalities.

Sensitization

Sensitization is facilitation of a response when the animal faces unexpected, often intense stimuli. The facilitated response may serve to enhance the success of avoidance or escape behaviors. In Aplysia and leeches, the neuromodulator serotonin was found to play a key role in sensitization (Byrne & Hawkins, Reference Byrne and Hawkins2015). Research in C. elegans, however, showed that other molecular mechanisms underlie sensitization.

In C. elegans, the pair of polymodal nociceptor ASH neurons in the head of the worm detect aversive stimuli, including harsh touch, volatile chemicals, and osmotic pressure. Detection of these stimuli causes the worm to crawl backward away from the source of stimulation. Transgenic worms with genetically encoded ChR2 in ASH respond to photoactivation of ASH by reversing. If a worm experienced a tap stimulus followed by optogenetic stimulation of the ASH neurons, the ASH-mediated escape response shows sensitization (Chew et al., Reference Chew, Tanizawa, Cho, Zhao, Yu, Ardiel and Schafer2018). Intriguingly, FLP-20, the same neuropeptide implicated in intermediate-term habituation, also regulates this form of sensitization. Mechanosensory neuron-released FLP-20 binds to FRPR-3 neuropeptide receptors to mediate ASH response sensitization. One major cellular site of action of FRPR-3, the neuroendocrine interneuron RID, was identified.

Chemosensory responses can also be sensitized by previous experience. Avoidance of the noxious odor of 2-nonanone can be sensitized if worms are preexposed to it (Yamazoe-Umemoto et al., Reference Yamazoe-Umemoto, Fujita, Iino, Iwasaki and Kimura2015). This form of sensitization is also neuropeptide dependent, as mutations in genes encoding key components of neuropeptide biosynthesis egl-3 (proprotein convertase) and egl-21 (carboxypeptidase E) showed no increased avoidance in this paradigm. Interestingly, although dopamine is not involved in producing sensitization, it contributes to maintaining the movement direction away from the source of aversive stimuli (Yamazoe-Umemoto et al., Reference Yamazoe-Umemoto, Fujita, Iino, Iwasaki and Kimura2015). Together, neuropeptides and dopamine interact to increase the effectiveness of the avoidance response.

Habituation to Nociceptive Stimuli as a Way to Shift Behavioral Strategy

C. elegans habituates to repeated stimulation of the ASH nociceptor neurons (Ardiel et al., Reference Ardiel, Giles, Yu, Lindsay, Lockery and Rankin2016; Hart et al., Reference Hart, Kass, Shapiro and Kaplan1999). Why do animals habituate to potentially dangerous nociceptive stimuli? Additional studies examined behavioral changes of different response components during habituation as well as assessed the behavior during the ISI. Habituation to repeated optogenetic activation of ASH occurs differently in components of the reversal response: pronounced decrement was observed in duration and latency, whereas the probability of responding showed very little decrement (Ardiel et al., Reference Ardiel, Giles, Yu, Lindsay, Lockery and Rankin2016). While habituating to the stimuli, worms also displayed a previously uncharacterized sensitization of forward locomotion between stimuli (Ardiel et al., Reference Ardiel, Yu, Giles and Rankin2017). With repeated stimulation, locomotion during the ISI gradually decreased backward movement and increased fast forward movement. Thus, integration of a suite of coordinated behavioral changes occurred during repeated optogenetic activation of ASH neurons such that worms remained fairly responsive to ensure an escape response to a potentially toxic stimulus; meanwhile, worms reduced the magnitude of the response and increasingly engaged in accelerated forward locomotion to evacuate from where they repeatedly encountered aversive stimuli. The coordination of response habituation and locomotion sensitization is mediated by a family of neuropeptides, Pigment-Dispersing Factors PDF-1 and PDF-2, that are orthologs of a neuropeptide previously implicated in arousal in arthropods (Ardiel et al., Reference Ardiel, Yu, Giles and Rankin2017). Without these peptides or their receptor PDFR-1, the forward locomotion sensitization did not occur. Thus, habituation of components of the response to aversive stimuli combined with sensitization in other behaviors serves to adaptively alter the behavioral strategy to help animals respond to, and disperse away from, potential danger.

Summary of Nonassociative Learning

The behavioral characteristics of nonassociative learning in C. elegans largely align with those in other species; although some of the underlying molecular mechanisms appear to be conserved, others showed divergence. Based on studies of habituation in C. elegans, behavioral responses are made up of multiple components (probability, duration, speed, latency, and ongoing behaviors) that undergo differential plasticity to produce a large range of behaviors. Habituation and sensitization independently modify innate responses, as well as synergistically increase the effectiveness of behavior in the experimental context. These results show that nonassociative learning is not a single, global phenomenon that governs all behavioral aspects; rather, different behavioral changes may occur in one or several behavioral components, depending on the animal’s experience, to produce optimal responses for specific situations.

Gustatory Learning

C. elegans is innately attracted to NaCl. Na+ and Cl− ions are mainly detected by the pair of ASE gustatory neurons in an asymmetrical way: the left ASE neuron, ASEL, primarily senses Na+ and the right one, ASER, primarily senses Cl− (Pierce-Shimomura et al., Reference Pierce-Shimomura, Faumont, Gaston, Pearson and Lockery2001). Wen et al. (Reference Wen, Kumar, Morrison, Rambaldini, Runciman, Rousseau and Van Der Kooy1997) found that untrained worms showed equal preference for Na+ and Cl− in a choice test; however, if one of the ions was paired with the presence of food and the other ion with a repulsive substance, or with the absence of food during training, their choice was biased toward the ion paired with food and away from the ion paired with the repulsive substance.

Worms placed on a Petri dish with NaCl and no food later avoided NaCl; this avoidance could be reversed by exposure to NaCl and food together or starvation without NaCl (Saeki et al., Reference Saeki, Yamamoto and Iino2001). Similarly, if NaCl was paired with repulsive chemicals, worms then avoided NaCl (Hukema et al., Reference Hukema, Rademakers and Jansen2008). NaCl avoidance can be impaired by drugs disrupting transcription and translation or by mutations affecting CREB, suggesting that memory for salt avoidance is protein synthesis dependent, and perhaps long-lasting (Peymen et al., Reference Peymen, Watteyne, Borghgraef, Van Sinay, Beets and Schoofs2019). Thus, worms can promptly modify their response to salt based on their recent experience and retain LTM for salt avoidance.

Three parallel neuropeptide signaling pathways are involved in salt avoidance. Tomioka et al. (Reference Tomioka, Adachi, Suzuki, Kunitomo, Schafer and Iino2006) showed that an insulin signaling pathway is critical for learned NaCl avoidance. The insulin-like peptide INS-1 acts on DAF-2 insulin receptor specifically in ASER to mediate salt learning. By contrast, the vasopressin/oxytocin-related neuropeptide NTC-1 signals through its receptor, NTR-1, in the ASEL sensory neuron, to facilitate salt conditioning (Beets et al., Reference Beets, Janssen, Meelkop, Temmerman, Suetens, Rademakers and Schoofs2012). Myoinhibitory peptide MIP-1/NLP-38 mediates learned salt avoidance by functioning in several neurons (Peymen et al., Reference Peymen, Watteyne, Borghgraef, Van Sinay, Beets and Schoofs2019). Taken together, peptidergic neuromodulation functions in parallel in different cells to underlie salt avoidance learning. Such collaborative molecular actions may ensure that the learned avoidance is strong enough to counteract unconditioned appetitive behavior and allow for tuning of learned behavior at different cellular sites.

In mammals, an important mediator of associative learning is the NMDA-type glutamate receptor. In C. elegans, Kano et al. (Reference Kano, Brockie, Sassa, Fujimoto, Kawahara, Iino and Maricq2008) showed that NMDA receptors were involved in the maintenance of salt conditioning memory. Learned salt avoidance diminished faster in worms with mutations in NMDA receptor subunits nmr-1 and nmr-2. NMDA receptors gate Ca+ influx, and CMK-1, a calcium/calmodulin-dependent kinase, function in the ASE sensory neurons to induce salt avoidance (Lim et al., Reference Lim, Fehlauer, Das, Saro, Glauser, Brunet and Goodman2018). Lau et al. (Reference Lau, Timbers, Mahmoud and Rankin2013) showed that nmr-1 was critical for context conditioning, suggesting a critical role of NMDA receptors in mediating classical conditioning across paradigms and phylogeny.

Olfactory Learning

C. elegans is attracted to odors of volatile chemicals such as diacetyl produced by bacteria (food). Conditioning worms with diacetyl paired with aversive acetic acid led to avoidance of diacetyl odor (Morrison et al., Reference Morrison, Wen, Runciman and van der Kooy1999). Using the same paradigm, Morrison and van der Kooy (2002) found that worms missing a glr-1 AMPA type glutamate receptor subunit failed to learn avoidance, suggesting a role for glutamate. Worms also learned to avoid an attractant, 1-propanol, when it was paired with HCl, a repulsive acid (Amano & Maruyama, Reference Amano and Maruyama2011). Short-lived memory (less than 3 hours) for conditioned propanol avoidance is produced in a massed-training paradigm, whereas LTM lasting at least 24 hours is produced in a spaced-training paradigm. Mutations in NDMA receptor subunit nmr-1 impaired both short-term memory and LTM for conditioned propanol aversion, whereas worms with a mutation in the transcription factor CREB, crh-1, showed a selective deficit in LTM. The role of glutamate signaling and CREB in learning and memory in animals ranging from worms to humans shows the evolutionarily ancient role of these molecules in experience-dependent plasticity.

Experience-dependent changes in unconditioned chemosensory avoidance have also been observed. Worms normally avoid butanone. However, exposure to butanone and food leads to approach toward butanone (Torayama et al., Reference Torayama, Ishihara and Katsura2007). Two AWC olfactory neurons (labelled AWC^ON and AWC^OFF) have distinct gene expression patterns and sense different volatile chemicals; butanone is detected by the AWC^ON neuron (Bargmann, Reference Bargmann2006). C. elegans homologs of Bardet–Biedl syndrome genes, bbs-1, osm-12/bbs-7, and bbs-8, function specifically in AWC^ON to mediate butanone appetitive learning (Torayama et al., Reference Torayama, Ishihara and Katsura2007). Using this odor–food pairing, Kauffman et al. (Reference Kauffman, Ashraf, Corces-Zimmerman, Landis and Murphy2010) showed that spaced training with butanone produced LTM that required CRH-1/CREB. Nishijima and Maruyama (Reference Nishijima and Maruyama2017) showed that after aversive 1-nonanol and appetitive KCl were presented together, worms switched their response to the odor from avoidance to approach. This appetitive olfactory memory shared features with aversive olfactory memory. Memories last for different time periods depending on massed or spaced training protocols. They were affected similarly by nmr-1 and crh-1 mutations suggesting that shared molecular mechanisms mediate both aversive and appetitive olfactory learning. Overall, this research shows that C. elegans modifies unconditioned chemosensory preferences as a result of experience: a compound that predicts food becomes attractive, one that predicts lack of food or something unpleasant becomes repulsive. Thus, chemosensory responses are sculpted by experience in ways that enhance the worm’s ability to successfully find food and avoid environments without food.

Worms are attracted to the smells of some pathogenic bacteria, but they learn to avoid the pathogen after they have experienced illness produced by the bacteria; thus, they learn an association between the odors and pathogenicity. When given a choice between E. coli OP50, the laboratory food for C. elegans, and Pseudomonas aeruginosa PA14, a pathogenic bacterium that infects the worm’s intestine and releases toxins, naïve worms initially showed a preference for PA14; however, a four-hour exposure to PA14 caused worms to lose this attraction (Zhang et al., Reference Zhang, Lu and Bargmann2005). Zhang et al. (Reference Zhang, Lu and Bargmann2005) showed that the learned pathogen avoidance is dependent on serotonin synthesized in the ADF neurons and a serotonin-gated chloride channel, MOD-1. Pathogen aversive learning in C. elegans is regulated by a complex insulin signaling network, and at least five insulin-like neuropeptides, including INS-6, INS-7, INS-4, INS-16, and INS-11 (Chen et al., Reference Chen, Hendricks, Cornils, Maier, Alcedo and Zhang2013; Lee & Mylonakis, Reference Lee and Mylonakis2017; Wu et al., Reference Wu, Duan, Yang, Liu, Caballero, Fernandes de Abreu and Zhang2019), have been shown to mediate learned pathogen avoidance. Interestingly, insulin has also been implicated in conditioned taste aversion in Lymnaea snails (Mita et al., Reference Mita, Yamagishi, Fujito, Lukowiak and Ito2014). Additionally, sites outside of the nervous system have been found to participate in this form of learning. INS-11, an intestine-secreted neuropeptide, regulates this learned avoidance. Zhang and Zhang (Reference Zhang and Zhang2012) found that DBL-1, a TGF-β homolog, acting on its receptor in the hypodermis, was necessary for learned avoidance. Collectively, olfactory aversive learning to pathogens is mediated by multiple neuromodulators, and the mechanisms identified suggest that learning can be modulated by tissues both inside and outside of the nervous system.

Thermosensory Learning

Temperature regulates essential biological processes, including circadian rhythm and metabolism in C. elegans. Mori and Ohshima (Reference Mori and Ohshima1995) found that the neural circuit required for thermotaxis was comprised of the pair of AFD thermosensory neurons and interneurons AIY, AIZ, and RIA. Additionally, AWC olfactory neurons were found to be secondary thermosensory neurons (Biron et al., Reference Biron, Wasserman, Thomas, Samuel and Sengupta2008; Kuhara et al., Reference Kuhara, Okumura, Kimata, Tanizawa, Takano, Kimura and Mori2008). Well-fed worms learned the association between a specific temperature and the presence of food and will thermotax toward their previous cultivation temperature (Hedgecock & Russell, Reference Hedgecock and Russell1975; Mori & Oshima, Reference Mori and Ohshima1995). AFD responds to temperature increase and decrease from a baseline set by previous cultivation, and interestingly, the memory for previous cultivation temperature is stored in these sensory neurons, as laser-ablating AFD dendrites disrupted thermosensory memory (Clark et al., Reference Clark, Biron, Sengupta and Samuel2006). Ohnishi et al. (Reference Ohnishi, Kuhara, Nakamura, Okochi and Mori2011) demonstrated that AFD-released glutamate inhibits AIY through a glutamate-gated inhibitory receptor and causes worms to migrate toward a colder temperature, whereas AWC-released glutamate activates AIY through an excitatory glutamate receptor and causes worms to migrate toward a warmer temperature. Thus, the learned temperature preference is regulated by the balance between AFD-AIY and AWC-AIY synaptic signaling. Perhaps having two reciprocal cellular sites to encode thermotaxis allows worms to differentiate the response to relative cold and warm temperatures based on their thermal experience rather than on absolute temperatures.

Mohri et al. (Reference Mohri, Kodama, Kimura, Koike, Mizuno and Mori2005) observed that worms thermotaxed toward temperatures paired with food and away from temperatures paired with starvation and showed that the effect of feeding state on thermosensory associative learning is mediated by monoamines, as exogenous serotonin and octopamine mimicked the effects of on- and off-food, respectively. Kodama et al. (Reference Kodama, Kuhara, Mohri-Shiomi, Kimura, Okumura, Tomioka and Mori2006) showed that INS-1 antagonizes DAF-2 in AIZ, and this insulin signaling pathway modulates temperature-feeding state conditioning by downregulating AIZ activity. It is interesting to note that although insulin signaling is important for both salt starvation conditioning and thermotaxis to cultivation temperature, it plays opposite roles in these processes. By contrast, TAX-6, a homolog of calcineurin, is critical for both thermosensory and gustatory associative learning (Kuhara & Mori, Reference Kuhara and Mori2006).

Learning of Other Features

C. elegans also learns about other environmental features. For example, although worms have a natural preference for a specific level of oxygen in the environment, that oxygen preference can be altered by experience (Cheung et al., Reference Cheung, Cohen, Rogers, Albayram and De Bono2005). When allowed to freely roam in an 0%–21% oxygen gradient, naïve worms prefer 5%–12% oxygen; however, worms previously maintained at 1% oxygen with food sought low oxygen (0%–7%), suggesting that worms learn the association between oxygen level and food, and use oxygen level as a cue for food. Prior experience with high or low levels of oxygen also reprograms the response to pheromones (Fenk & de Bono, Reference Fenk and de Bono2017). The same pheromone attracted worms maintained at 21% oxygen, but repelled worms maintained at 7% oxygen, and this cross-modal behavioral modification is thought to be mediated by the RMG interneurons.

Worms can also integrate multiple sensory experiences in learning. Normally, both male and hermaphrodite worms exhibit learned salt avoidance after salt-starvation conditioning. Surprisingly, male worms showed a form of sex-specific learning in this paradigm. If there were hermaphrodites nearby during conditioning, male worms continued to be attracted to salt even after it was paired with starvation, suggesting that they used salt as a cue to find mates, perhaps to prioritize reproductive success (Sakai et al., Reference Sakai, Iwata, Yokoi, Butcher, Clardy, Tomioka and Iino2013). Sammut et al. (Reference Sammut, Cook, Nguyen, Felton, Hall, Emmons and Barrios2015) found that the neuropeptide PDF-1 in a pair of male-specific interneurons, MCM, regulates sex-specific conditioning.

Summary of Associative Learning

Research has shown that worms can learn and remember a wide range of environmental features that predict the presence or absence of food or danger and modulate their behavior according to their experience in a way that allows them to find optimal environments and escape from adversity. Various forms of associative learning are mediated by both shared and distinct neural and molecular pathways, and a number of genes and molecules show functional convergence between worms and other species (see Kriete and Hollis, Chapter 3; Guan, Chapter 25), suggesting that many learning and memory mechanisms may be evolutionarily conserved.