A translational and multidisciplinary approach to studying the Garcia effect, a higher form of learning with deep evolutionary roots Free

In the 1950s, John Garcia characterized a novel higher-order form of learning that today bears his name: the Garcia effect, a form of conditioned taste aversion (CTA; see Glossary) that results from experiencing nausea several hours after the consumption of a novel food (Fig. 1A; Garcia and Koelling, 1966; Garcia et al., 1955, 1967, 1985). This discovery profoundly impacted the prevailing theories about learning and memory and laid the foundation for subsequent clinical investigations, which revealed the relevance of the Garcia effect to conditions such as eating disorders and food aversion during chemotherapy (Bernstein, 1978).

For a long time, the Garcia effect was investigated only in small mammals (i.e. rats and mice) (Garcia et al., 1955, 1966, 1967, 1985). However, seven decades after its discovery, our research group characterized the Garcia effect in an invertebrate model organism, the pond snail Lymnaea stagnalis (Rivi et al., 2021b). This Commentary begins by retracing the key experiments of Garcia and collaborators and considering their clinical significance. We then focus on summarizing recent studies of the Garcia effect in the pond snail, underscoring its conservation across taxa. Additionally, we propose that further research on the Garcia effect should take a multidisciplinary approach encompassing ecological and translational studies, integrating in silico, molecular and behavioural studies (Fig. 1B). This effort aims to unravel the conserved mechanisms underlying this form of learning in diverse model organisms, offering insights into its adaptive significance and potential therapeutic applications.

Most people have experienced eating a novel food and, sometime later, feeling sick. What triggered that sickness? There are various potential causes, such as improper preservation or preparation of the dish, or an unrelated infection. Despite these alternative (and plausible) explanations, the malaise will likely be attributed to the novel food consumed; subsequently, the mere thought of the taste of that new food causes a sense of nausea that deters the person from consuming it again. Why is it so easy to ‘blame’ the new food instead of considering other possible factors? The answer to this question came with experiments in the 1950s that led to a paradigm shift in thinking both about how humans and other mammals learn, and about the conditions under which learning occurs (Freeman and Riley, 2009). The initial insights into the Garcia effect date back to observations made during World War II by the ecologist C. S. Elton, who noted that rodents, after consuming poisoned bait, avoided that specific bait food for weeks even after they recovered, suggesting the formation of a long-lasting ‘bait-shyness' (see Rzóska, 1953). At the beginning of 1951, John Garcia came to realize that this phenomenon held significance beyond its potential application in pest control (Garcia et al., 1955). He observed that rats provided with water in plastic bottles before experiencing radiation-induced sickness refrained from drinking water from those bottles but still consumed water from glass bottles (Garcia et al., 1955). This led to the hypothesis that the plastic bottles altered the taste of the water, and this taste was associated by the rats with later feelings of sickness (Garcia et al., 1966, 1985). To test this, Garcia exposed rats to a saccharin solution during radiation treatment and found that irradiated rats exhibited a significant reduction in saccharin consumption compared with the control group that had consumed saccharin but had not been subjected to radiation (Garcia et al., 1985). This was the first demonstration of the Garcia effect, paving the way for discoveries that – in many respects – changed the history of neuroscience (Freeman and Riley, 2009; Gradowski, 2024; Lin et al., 2014; Reilly, 2009).

Garcia's lab also noticed that the irradiated rats formed a rapid and long-lasting avoidance behaviour only to water and did not avoid the compartments where radiation exposure occurred, suggesting that not all stimuli are equally prone to forming associations with the sickness status (Garcia et al., 1985). These findings introduced the concept of selective learning and demonstrated the taste specificity of the Garcia effect. Additionally, Garcia and colleagues noticed that rats developed a taste aversion after ingesting saccharin when subsequently administered apomorphine, a nausea-inducing drug, even with a delay of up to 75 min between the two events (Garcia et al., 1966, 1985). Thus, despite the extraordinary nature of these studies, Garcia's findings violated what were believed at the time to be the laws of learning (e.g. the necessity of close temporal contiguity between the conditional and unconditional stimuli; see Glossary). Consequently, these discoveries were met with scepticism, generating a plethora of questions from the scientific community (Gradowski, 2024). However, the robustness of the Garcia effect and its reproducibility made it evident that this was a special form of taste-aversion learning (i.e. conditioned taste aversion) with a fundamental role in the evolution of behaviour. From an evolutionary perspective, the ability to form taste–illness associations and retain them as long-term memory (LTM) over a long delay between tasting and nausea, even after a single exposure, is highly advantageous, enabling animals to avoid the potentially lethal consequences of repeated ingestion of toxins (Garcia, 1981).

From the mid-1970s onward, the principles derived from Garcia's studies were translated to humans, where this phenomenon is typically referred to as the ‘Sauce Béarnaise effect’ (Stensmyr and Caron, 2020). Additionally, researchers began to apply these principles to a diverse range of issues, utilizing the Garcia effect not solely for understanding the phenomenon itself but also as a tool to investigate and address various behavioural issues. Thus, the role of the Garcia effect in food aversions and preferences during pregnancy, changes in food preferences associated with cancer progression, and side-effects linked to chemotherapy have been demonstrated across many studies (Bernstein, 1978; Davidson et al., 2014; Jcobsen et al., 1993). However, the number of studies on the Garcia effect in rodents has undergone a considerable decline, with roughly only 15 studies conducted in the last 15 years (Schier et al., 2011, 2012; Tracy and Davidson, 2006; Tracy et al., 2004). Importantly, these studies deviate significantly from Garcia's original research, showing that a substance perceived solely within the stomach, rather than in the oral cavity, can strengthen subsequent aversion to oral taste.

Interestingly, current research trends indicate a new interest in the Garcia effect, particularly in the context of leveraging its principles to mitigate interactions between predators and domesticated farm animals (Andrewartha et al., 2023; Cassaigne et al., 2023; Parker, 2006; Rowland et al., 2004; Sandner, 2004; Snijders et al., 2021). In other words, the use of the Garcia effect is returning to its roots, with a renewed interest in pest management. Consequently, a growing number of studies are underscoring the potential efficacy of employing the Garcia effect as an effective, non-lethal, simple and ethically acceptable method of predator management, including that of invasive predators (Lang et al., 2023).

Since its discovery in rodents, research on the Garcia effect has been exclusively focused on mammals, including humans (Schier et al., 2011, 2012; Tracy and Davidson, 2006; Tracy et al., 2004). However, we recently (and quite unexpectedly) demonstrated a Garcia-like effect for the first time in an invertebrate model organism, L. stagnalis (Rivi et al., 2021a). Lymnaea stagnalis is a wonderful model organism for investigating the shared molecular, cellular and behavioural mechanisms involved in learning and memory formation (Batabyal et al., 2022; Benatti et al., 2020; Fodor et al., 2020; Ito et al., 1999; Kagan et al., 2022; Rivi et al., 2020, 2021a, 2022a,b, 2023a,b,c), as it exhibits various forms of associative learning – including both classical and operant conditioning (see Glossary) – and can also consolidate the acquisition of new behaviours (i.e. learning) into LTM (Braun and Lukowiak, 2011; Rivi et al., 2023d,e,f). In the next sections, we provide an overview of the Garcia effect in L. stagnalis, highlighting its conserved characteristics and the types of stimuli that can be used to induce sickness in snails. We also compare memory capabilities between laboratory-inbred and freshly collected snails and explore the impact of different fasting durations on learning and memory processes.

To cause a Garcia effect in a strain of laboratory-bred and reared snails, the animals were exposed to heat shock stress (30°C for 1 h) 1 h after introducing them to a novel appetitive taste (a carrot slurry). The link between heat shock stress and nausea is discussed in more detail below. This protocol resulted in suppression of an appetitive response to the novel taste (i.e. a Garcia effect), persisting for at least 24 h as a LTM. However, in an additional cohort fed only carrots for a week before experiencing the same protocol, the Garcia effect was not formed, showing that food novelty is a necessity for the formation of the Garcia effect in Lymnaea as it is in mammals. Furthermore, Lymnaea also shows taste specificity of the Garcia effect: even after aversion to the novel taste is acquired, sucrose (i.e. another appetitive stimulus, but one not paired with the heat shock) still elicits feeding behaviour. Consistent with other studies, this research demonstrates that snails can recognize sucrose and carrot juice as different appetitive stimuli and discriminate between them, choosing what to eat and what not to eat based on past experiences (Ito et al., 2013; Sugai et al., 2006).

Additional studies revealed that a single pairing of the novel taste and heat shock, even with a separation of up to 48 h, is enough for the establishment of a Garcia effect lasting 24 h (Rivi et al., 2021b). This suggests that in snails, as in mammals, the Garcia effect can be formed even with a long interval between the novel appetitive stimulus and the nausea (Kagan et al., 2022, 2023). In a further parallel with mammalian work, the strongest feeding suppression in Lymnaea occurs when the inter-stimuli interval is between 1 and 4 h (Andrews and Braveman, 1975; Domjan and Bowman, 1974; McLaurin et al., 1963; Revusky, 1968; Rivi et al., 2021b; Smith and Roll, 1967). In summary, these data are all consistent with the ‘requirements’ presented by the Garcia group as necessary to demonstrate the Garcia effect: food novelty, taste specificity and a long time interval between the presentation of the novel taste and the nausea-inducing stimulus (Rivi et al., 2021b).

Of course, it is extremely difficult to show nausea in a snail, and the first studies aimed at investigating the Garcia effect in L. stagnalis used heat shock to induce a sort of visceral sickness. How can we be sure that this is something akin to nausea? Overall, heat shock could be considered as a multi-faceted stressor that may cause metabolic changes and induce adaptative reprogramming of energy metabolism (Bensaude et al., 1996). Previous studies from rodents demonstrated that heat exposure leads to deleterious effects on the homeostasis of the central nervous and immune systems, and induces alterations in epithelial barrier function and cell structure, which affect gut microbiota colonization (He et al., 2019; Lan et al., 2019; Qu et al., 2021; Quinteiro-Filho et al., 2010). Moreover, in humans, heat stress can lead to nausea and vomiting hours later (Becker and Stewart, 2011). The exposure of lab-inbred snails to temperatures of 30°C for longer than 3 h is lethal (McDonald, 1973), as this represents a severe stressor for these organisms, which have been maintained at ∼20°C for generations (Fernell et al., 2021; Rivi et al., 2022c). In L. stagnalis, at the transcriptional level, there is a significant increase in the expression of the heat shock proteins (HSPs) HSP70 and HSP40 in the central ring ganglia (see Glossary) of snails that form the Garcia effect (Rivi et al., 2021b). Importantly, the exposure of snails to a HSP blocker 1 h before the heat shock prevents both the heat-induced upregulation of HSPs and the occurrence of behaviour that is consistent with the Garcia effect (Rivi et al., 2021b, 2023a). Sickness is a requirement for the formation of the Garcia effect; in L. stagnalis, the heat shock-induced upregulation of HSPs (Rivi et al., 2021c, 2023g) seems to be necessary for the sickness induction and, therefore, for the formation of the Garcia effect (Rivi et al., 2021b, 2022a,c). Thus, although determining whether exposure to 30°C for 1 h makes snails sick is extremely challenging, the results from these studies suggest that in snails, as in mammals, the heat shock results in somewhat similar sickness. However, additional studies are required to explore further the sickness induced by heat shock exposure in snails.

Interestingly, the protocol outlined above does not induce the Garcia effect in freshly collected, outbred L. stagnalis (Rivi et al., 2022a,c). This may be because wild Lymnaea have experienced temperature fluctuations and extreme heat throughout their ecological and evolutionary history (Rivi et al., 2022a), which might give them a higher heat tolerance and resilience (Rivi et al., 2022c, 2023d). However, the heat shock stressor also does not produce a Garcia effect in their first-generation offspring reared under laboratory conditions. Thus, it has been hypothesized that tolerance to the heat shock stressor has been inadvertently selected out of the laboratory strain by rearing snails in the laboratory for decades. As the formation of the Garcia effect is contingent upon animals experiencing a visceral sickness, if snails do not undergo any ‘sickness’ following the heat shock stressor, they will not display the Garcia effect. It is possible that wild Lymnaea do not experience a sickness-type stimulus in response to heat shock; in fact, this hypothesis is supported by the observation that freshly collected snails have higher basal mRNA levels of HSPs compared with laboratory-bred ones, suggesting that they may possess greater thermal tolerance (Rivi et al., 2023e).

More recently, another potentially sickness-inducing stimulus, the bacterial toxin lipopolysaccharide (LPS) has been successfully used to produce a Garcia effect (Rivi et al., 2023d) in both lab-inbred and wild-outbred populations of Lymnaea (Rivi et al., 2023c,h). LPS is commonly used in rodent studies to induce systemic immune system activation, leading to significant physiological and behavioural alterations (e.g. anhedonia, lethargy, reduced appetite, anxiety and drowsiness) collectively known as ‘sickness behaviour’ (see Glossary; Bull et al., 1994; Dantzer and Kelley, 2007; De La Garza, 2005; Exton et al., 1995; Maes et al., 2012; Rademacher et al., 2021). Moreover, in piglets, the medullary ‘vomiting centre’ is activated following LPS treatment, and the diarrheogenic activity of LPS has been described in mice (Choudhary et al., 2023; Girod et al., 2000; Lasselin et al., 2021; Liang et al., 2005). Together, these results suggest that LPS can be considered as a sickness-inducing stimulus.

Injecting L. stagnalis with 25 μg of LPS induces an inflammatory response, leading to elevated aerial respiration and impaired cognitive function (Rivi et al., 2022d). At the molecular level, LPS injection leads to a significant upregulation of the mRNA levels of key targets for immune response (such as Toll-like Receptor 4, TLR4) in the central ring ganglia of snails (Rivi et al., 2023h). This aligns with mammalian studies indicating that LPS, by stimulating TLR4, induces the release of proinflammatory cytokines, triggering a robust immune response, resulting in sickness behaviour (Bassi et al., 2012; Lu et al., 2008). Importantly, although LPS treatment per se does not affect the expression levels of neuroplasticity genes in L. stagnalis, its combination with the conditioning procedure results in a significant upregulation of the expression levels of key (conserved) targets for learning and memory formation, including the glutamate ionotropic receptor NMDA type subunit 1 (LymGRIN1), 2A (LymGRIN2A) and 2B (LymGRIN2B), the glutamate ionotropic receptor AMPA type subunit 1 (LymGRIA1), as well as the transcription factor cAMP response element-binding protein 1 (LymCREB1) in the central ring ganglia (Batabyal et al., 2021; Rivi et al., 2020, 2023h).

The LPS-induced upregulation of immune-related genes can be prevented by exposing snails to acetylsalicylic acid (aspirin, an anti-inflammatory drug) before the LPS injection, and the suppression of the inflammatory cascade and the associated sickness-like behaviour can prevent the formation of the Garcia effect (Rivi et al., 2022d, 2023b).

As previously reported, L. stagnalis represents a valid model organism to study how memory is formed, consolidated, recalled and extinguished, as it exhibits various forms of associative learning (Rivi et al., 2021a). Pond snails can be operantly conditioned to not perform aerial respiration in a hypoxic environment through a tactile stimulus to their pneumostome area (Lukowiak et al., 1996). Typically, lab-bred snails require two 30 min training sessions to form LTM for this operant conditioning (Braun and Lukowiak, 2011). However, certain stressors and bioactive compounds – such as heat shock, predator scent and flavonoids – enhance LTM, so that a single 30 min session is sufficient to form a LTM lasting for at least 24 h (Il-Han et al., 2010; Rivi et al., 2021c, 2023a,g; Teskey et al., 2012). It has been recently demonstrated that snails that formed a Garcia effect (following the paired presentation of a novel taste and the LPS injection) showed LTM formation for the operant conditioning of aerial respiration when trained with a single 30 min training session in the presence of the taste they became averse to (Rivi et al., 2023f). In other words, following the Garcia effect formation, the taste serves as a ‘sickness’ risk signal, acting as a memory-enhancing stressor for another conditioning procedure. Thus, this study illustrates the complex interactions between multiple memory stores, paving the way for future studies aimed at exploring the underlying molecular mechanisms for memory formation and consolidation.

As mentioned above, it is possible to use the Garcia effect as a tool to investigate other behaviours, and the Garcia effect in Lymnaea has been used for this purpose, being incorporated into a protocol to assess the link between fasting and memory formation (Rivi et al., 2023i). Growing evidence indicates that brief periods of fasting can enhance neuronal resilience against injuries, inflammation and degeneration, and improve brain plasticity and memory performance (Gudden et al., 2021), whereas prolonged fasting may have adverse effects on brain function and lead to memory impairment (Frintrop et al., 2019; McCormick et al., 2008). The molecular mechanisms underlying the divergent impacts of short- and long-term fasting on cognitive function remain relatively unexplored because of the complexity of mammalian brains and behaviours. In snails, the Garcia effect procedure has proven to be a valuable behavioural paradigm for overcoming these challenges at both the behavioural and molecular level (Rivi et al., 2023i).

In a study using the Garcia effect in L. stagnalis to investigate the link between fasting and learning, snails fed ad libitum and those food deprived for 1 or 5 days before being subjected to the Garcia effect procedure were graded based on their memory performance by referring to a validated grading system (Batabyal et al., 2021; Rivi et al., 2021b, 2023i; Rosenegger et al., 2004). Consistent with previous Lymnaea studies (Aonuma et al., 2018; Ito et al., 2015, 2017; Mita et al., 2014; Totani et al., 2019, 2020), snails that were food deprived for 1 day were the ‘best’ learners, the ad libitum-fed snails were ‘average’ learners, whereas the severely hungry individuals (i.e. those food deprived for 5 days) did not show the Garcia effect. Similar data have been obtained with more conventional studies on the ‘standard’ CTA in L. stagnalis (Ito et al., 2013, 2015, 2017). One day of food deprivation induces a transcriptional upregulation of the molluscan insulin-like peptide (LymMIP-II) and genes involved in neuroplasticity, whereas severe fasting significantly downregulates the expression of LymMIP-II and upregulates the expression of key genes involved in energy homeostasis, not only in snails but also in rodents and mammals (Gagnon et al., 2023; Jørgensen et al., 2021; Şentürk et al., 2024). However, the injection of bovine insulin into snails experiencing severe fasting before being trained for the Garcia effect procedure allows LTM to form; these snails show upregulation of LymMIP-II as well as of key targets for neuroplasticity, including LymGRIN1 and LymCREB1. In contrast, injecting moderately fasting snails with the insulin receptor antibody results in a significant upregulation of genes involved in energy homeostasis and animal survival and prevents the formation of the Garcia effect.

Apart from indicating the existence of a specific energetic state needed for the formation of the Garcia-effect LTM, these findings are also consistent with the hypothesis that insulin might have a crucial role in determining the memory phenotype. Thus, this study (Rivi et al., 2023i) demonstrates that the use of the Garcia effect paradigm in a relatively simple organism like L. stagnalis is a valuable platform on which to study the relationship between different lengths of fasting and memory performance and the role of insulin-mediated synaptic plasticity, paving the way for studies in mammals.

A great advantage of using L. stagnalis as a model organism is that it may be the only model system where specific neurons have been identified as necessary sites for LTM formation after operant conditioning of aerial respiration (RPeD1 neuron; Lukowiak et al., 2003) and CTA (cerebral giant cells; Azami et al., 2006). Using this knowledge, in conjunction with the ability to remove the soma (including the nucleus) of an individual neuron while leaving behind a functional neurite (Van Minnen et al., 1997), it is possible to show that an individual neuron is a necessary site of LTM formation in L. stagnalis (Scheibenstock et al., 2002; Sunada et al., 2017). Because the Garcia effect is a form of CTA, it will be possible to directly study the changes occurring in the relevant neurons following Garcia-effect training, including whether these changes are necessary for LTM formation. It may also be possible to develop an in vitro semi-intact preparation in which to study the changes in the neuronal activity of the cerebral giant cells before and after Garcia-effect training (McComb et al., 2005). If this were achieved, it would be the first instance where the electrophysiological changes in a single neuron could be shown to be causal for a Garcia-effect memory. Although we do not believe this can yet be accomplished in a mammalian preparation, experiments like these would allow us to develop strategies to either enhance or mitigate the formation of such memory at the level of a single neuron.

Although invertebrates possess relatively simple nervous systems with a limited number of neurons in separate ganglia, they can perform sophisticated and complex behaviours, as well as higher-order forms of learning. In particular, molluscs and arthropods exhibit remarkable behavioural repertoires and cognitive abilities, with a nervous system that is much simpler than that of mammals (Agin et al., 2006; Allen et al., 1986; Anderson and Mather, 2007; Baldwin, 1896; Benard et al., 2006; Chichery and Chichery, 1987; Coolen et al., 2005; Gatto et al., 2022; Guiraud et al., 2018; Hammer and Menzel, 1995; Hepburn et al., 1973; Heyes, 1994; Orvis et al., 2022; Pahl et al., 2013; Ponte et al., 2022; Rivi et al., 2023b; Thorpe, 1944; Worden and Papaj, 2005).

In the last five decades, invertebrate research has significantly contributed to our knowledge of the mechanisms underlying memory formation, consolidation, reconsolidation and extinction. Despite this, with the sole exception of L. stagnalis, the Garcia effect has not been investigated in invertebrate models. However, many invertebrate organisms are known to form LTM for the ‘standard’ CTA, including crickets (Gryllus bimaculatus), flies (Drosophila melanogaster), worms (Caenorhabditis elegans), crayfish (Procambarus clarkii) and bees (Apis mellifera) (Amano and Maruyama, 2011; Arzuffi et al., 2000; Chapman, 1982; Guiraud et al., 2018; Lyu and Mizunami, 2022; Wu et al., 2005). Thus, as the Garcia effect is a ‘special’ form of CTA, it is possible that these models, like L. stagnalis, are capable of forming a Garcia effect. We strongly believe that the use of invertebrates to explore the conserved behavioural and molecular mechanisms underlying the Garcia effect is only just beginning. Research on invertebrate models may be extremely useful for rapidly and efficiently studying the conserved mechanisms underlying the Garcia effect and addressing the clinical conditions in which it is involved, bridging the gap between preclinical animal assays and clinical studies. From an ethical and economic perspective, the use of invertebrate models avoids the implications of mammalian work. Mammals need only be used to validate the results obtained from invertebrates, which would also help to greatly reduce the costs of preclinical studies (Tascedda et al., 2015), while still allowing us to gain important information on the conserved mechanisms underlying the Garcia effect.

Here, we report what we think to be the best practice to provide novel insight into the conserved mechanisms underlying the Garcia effect. To promote the ‘translatability’ of the results obtained in invertebrates to mammals, proteomic and metabolomic analyses are necessary (Cristina et al., 2022). A valuable technique that allows us to correlate molecular effects with behavioural outcomes is CRISPR/Cas9, which has been recently validated in L. stagnalis and other invertebrates (Abe and Kuroda, 2019). This technique allows researchers to turn off key genes in the pathways of interest so that one can evaluate the phenotypic, molecular and behavioural consequences of gene manipulation. In this way, CRISPR/Cas9 allows us to uncover causal relationships between genetic alterations, molecular changes and behavioural outcomes.

To bridge the gap between the molecular data and the physiological and behavioural outcomes, the use of network analysis (see Glossary) represents another promising tool. This approach, by combining data from bioinformatic and molecular studies, offers insights into system-level phenomena such as the Garcia effect (Charitou et al., 2016). From a comparative and evolutionary standpoint, network analysis would allow investigation of the conserved mechanisms underlying the Garcia effect, illustrating relationships among genes, transcripts, proteins and metabolites through visual network construction. Furthermore, by combining network theory with bioinformatic techniques (e.g. molecular docking) and mathematical models (such as machine-learning algorithms), it should be possible to identify key components involved in this higher-order form of learning, explore associations between behaviours and genes, and uncover the strength and directionality between the variables, revealing the cues that are most strongly linked to aversive experience. This should allow us to model the network of factors involved in the Garcia effect to predict the occurrence and intensity of taste aversion under different conditions.

Research across invertebrates and vertebrates using the Garcia-effect framework could provide novel insights into designing ecological studies on foraging patterns and social learning. The potential application of the Garcia effect could be extended to conservation biology, where it might aid in devising strategies for managing wildlife populations by studying their learned responses to specific environmental cues or threats that induce a sickness state. Such responses can also be socially learned (Johnston et al., 2012; Mason, 1984), and investigating principles of social learning with regard to the Garcia effect would be a new avenue of research. For example, individuals might communicate sickness states to their conspecifics through social cues, leading to a widespread Garcia effect in a group within a short time where not all individuals necessarily need to encounter the sickness stimulus themselves.

In this Commentary, after reviewing the ground-breaking experiments that led to the characterization of the Garcia effect in rodents and its clinical implications in humans, we summarized the results of studies performed in L. stagnalis, which provided the first evidence of the Garcia effect in an invertebrate model system. These experiments paved the way for new avenues of research that may unravel the conserved mechanisms underlying this higher form of learning with deep evolutionary roots. When we conducted our first studies on the effects of novel food presentation with heat shock in L. stagnalis, we did not expect to provide the first characterization in an invertebrate model organism of the Garcia effect, a field of research that was pioneered by John Garcia more than 70 years ago. The results described in this Commentary are the first in what we expect to be an exciting scientific adventure, and we hope that other researchers worldwide will provide their contribution to this field.

We extend our gratitude to our colleagues Prof. Quentin Pittman, Prof. Enzo Ottaviani, Diana Kagan, Bevin Wiley and David Chau for their insightful discussions during the research process. Additionally, we express our appreciation to Dr Iain Philips for his efforts in collecting the wild-strain snails and to Dr Petra Hermann for providing the laboratory-reared population of snails. We acknowledge the Consorzio Interuniversitario Biotecnologie (CIB, Trieste, Italy) for the research grant awarded to Dr Veronica Rivi, enabling her to work as a visiting researcher at the Hotchkiss Brain Institute (University of Calgary, AB, Canada). We are grateful to the three anonymous reviewers for their valuable feedback on the manuscript.