Associative Learning and Memory Processing in Bees

The social life of the honeybee colony forms the ecological framework for the individual animal's behavior and is crucial for each bee's survival, because an individual bee cannot exist on its own (Frisch, 1967; Lindauer, 1967).

Associative Learning

The study of learning and memory formation in bees under natural conditions has focused on latent learning during navigation and on operant learning in the context of food collection. In the laboratory it has focused on appetitive classical conditioning. Bees departing from the hive perform observatory learning flights (Capaldi et al., 2000), and establish a map-like spatial memory for their colony's location relative to landmarks within the framework of their sun compass system (Menzel et al., 2000). When a searching bee discovers a nectaror pollen-producing flower, it quickly learns to associate the surrounding visual and olfactory signals with the reward. It learns olfactory stimuli (e.g., floral odorants) and colors very quickly (within one or a few learning trials). Patterns need more learning trials. Whereas latent learning during navigational tasks may not require a rewarding stimulus, reward learning is a forward-associative process because signals perceived before the reward are associated, whereas those perceived during the reward or during the departure flight are associated less effectively or not at all.

Research on various operant learning phenomena (e.g. reversal and multireversal learning, over-learning, inhibitory learning, context-dependent learning, and reward schedule learning) has found performances similar to those in mammals (Couvillon and Bitterman, 1988; Menzel, 1990). Multiple experience with varying signals but one constant feature (e.g., different kinds of symmetrical patterns) leads to the formation of a concept (the concept of symmetry) that allows the bee to choose new patterns with the same feature as learned targets (Giurfa, Eichmann, and Menzel, 1996). Bees also develop a concept of sameness and difference when they are trained in delayed matching-to-sample tasks, in which they are required to respond to a matching stimulus or a non-matching stimulus (Giurfa et al., 2001). They also transfer the learned rules to new stimuli of the same or a different sensory modality. Thus, not only can bees learn specific objects and their physical parameters, but they also extract rules and apply them to novel situations.

Classical Conditioning

Classical conditioning of reflexes is a convenient way to study the behavioral and neural mechanisms of associative learning. In the honeybee, the proboscis-extension reflex (PER) to a sucrose stimulus at the antennae is a reliable reflex in the context of feeding. A hungry bee will reflexively extend its proboscis (tongue) when the antennae are touched with a drop of sucrose solution. An odor (conditioned stimulus, CS) presented shortly before the sucrose reward (un-conditioned stimulus, US) will be associated with the reward, even under conditions in which the animal is harnessed in a tube or is being prepared for physiological studies (Menzel and Müller, 1996). The associative nature of PER conditioning to odors has been established by demonstrating that only forward-pairing of CS-US sequences is effective. Unpaired CS and US presentations or CS-or US-only presentations do not lead to learning, and in differential conditioning (one odor CS+ paired with US, the other CS- unpaired), bees respond only to the CS+ and not to CS- (Menzel, 1990). The predictive value of the CS depends on the reliability with which it is causally related to the US. In differential conditioning, the reversal to the initially unpaired stimulus CS- is slower after more frequent unreinforced preexposures than after fewer preexposures. The same applies for US- only preexposures in an otherwise reinforced context, indicating that the absence of an expected US leads to inhibitory learning. If naive animals are stimulated with a compound of two odors and one of them is later associated with sucrose reward, the animals will also respond the second odor of the compound, even though this odor was not explicitly experienced during a learning trial (Müller et al., 2000). This form of learning (sensory preconditioning) indicates stimulus associations between equally evaluated stimuli and thus transcends the classical associative paradigm.

The role of reinforcement in the formation of an association is an essential question in learning theories: Are associations formed only by close contiguity between the CS and US? The blocking phenomenon indicates that this is not the case: If a novel CS appears together with a learned stimulus, this novel stimulus will be learned to a lesser extent or not at all (Kamin, 1968). The blocking paradigm is central to most current models of associative learning, and the phenomenon is explained either by the assumption of a competition between the two CSs (the already learned one and the novel one) for attention (Mackintosh, 1975) or for a limitation of reinforcing function that depends on the expectation or prediction of reinforcement (Rescorla and Wagner, 1972). Since attention, expectation and prediction are cognitive faculties, it is argued that cognitive capacities need to be introduced in theories about associative learning. It is thus interesting from a comparative point of view whether the bee with its tiny brain shows the blocking phenomenon. This question cannot yet be definitively answered. Blocking is found in some studies (Smith and Cobey, 1994; Thorn and Smith, 1997), but not in others (Gerber and Ullrich, 1999). Blocking across sensory modalities was also not seen in training free-flying bees (Bitterman, 1996; Funayama, Couvillon, and Bitterman, 1996). Second-order conditioning is another procedure that tests whether associative learning requires contiguity between CS and US. In a positive outcome of second-order conditioning one argues that a CS can acquire the potential of a US. This has been demonstrated for olfactory PER conditioning (Menzel, 1990).

Rules of elementary associative learning assume that in learning a compound stimulus, animals learn the associations between the reinforcer and the compound elements separately (Rescorla and Wagner, 1972). Contrary to this assumption, configural learning theories assume that, in learning a compound, animals build a new entity made from the conjunction of compound elements and that a connection is made between this new configuration and the reinforcer (Rudy and Sutherland, 1992). The different processing strategies underlying elementary and configural olfactory learning were studied by the negative patterning discrimination. In negative patterning two single stimuli are reinforced (A⁺, B⁺), while the compound is not (AB^-). Solving this problem—responding less to the compound than to the single elements—can be explained only by taking configural associations into account. Otherwise, summation of the elementary associative strengths in the compound should result in stronger response to the compound than to the elements. Honeybees can solve negative patterning discrimination in olfactory conditioning of the PER (Deisig, Lachnit, Hellstern, and Giurfa, 2001). The fact that bees solve negative patterning discrimination in olfactory conditioning and in color/odor tasks (Couvillon and Bitterman, 1988) shows that linear associations between single stimuli and the reinforcer are not the only ones underlying associative learning in honeybees (Giurfa et. al, 2001).

Memory Dynamics and Memory Localization

Memory is an animal's capacity to retain acquired information and to use it for future behavior. In the context of association theory, memory is the potential of a conditioned stimulus to activate an established associative link. Some researchers, however, view learning as acquiring information rather than responses, in which case memory would be a dynamic and self-organizing process of information storage. Support for such a cognitive interpretation of memory in the honeybee comes from the fact that olfactory memory formation is not identical to the process of acquisition. Memory needs time to develop and proceeds through phases that differ in their susceptibility to interfering events, their content, and their neural and cellular substrates (Menzel, 1999; Menzel and Müller, 1996; see Figure 1).

The memory trace for olfactory cues is distributed and involves two of the three convergence sites between the olfactory pathway and the reward pathway. The reward pathway was identified by Hammer (1993) and assigned to a single identified neuron, the VUM_mx1 neuron. Two of the three convergence sites—antennal lobes and mushroom bodies—are, respectively, the primary and secondary processing regions in the olfactory pathway, and each of these two neuropils establishes its own memory trace independently of the other (Hammer and Menzel, 1998). The two traces are, however, different at least with respect to their dynamics, and are likely to store different information.

Researchers have made progress in unraveling the neural correlates of memory for the antennal lobe by visualizing the changes in odor coding as a consequence of olfactory conditioning (Faber, Joerges, and Menzel, 1999). The antennal lobe is organized into glomeruli; odors are coded as specific spatial-activation patterns of the glomeruli. These patterns can be imaged using calcium-sensitive fluorescent dyes. As a result of conditioning, the neural representation of a trained odor becomes more pronounced and more distinct from nonrewarded odors, but its general features do not change, indicating that learning at this level intensifies the neural code of the learned signal but does not create a new representation. It is unclear what specifies the odor-induced activity patterns as those of a learned odor, because a stronger stimulus also induces a more pronounced and distinct activity pattern, but bees have no trouble distinguishing between a strong nonlearned odor and a weak learned odor.

A stable, lifelong memory is formed even after only a few learning trials as the result of sequential steps of memory processing. Five memory stages are distinguishuable on the basis of their respective temporal dynamics and their physiological and biochemical properties. The cellular and neural machinery underlying the memory stages is basically similar to those known for other model systems (Aplysia: Botzer, Markovich, and Susswien, 1998; Müller and Carew, 1998; Drosophila: Dubnau and Tully, 1998; Chick: Rose, 1991), although each model system has its own temporal dynamics. This indicates that the cellular and molecular machinery underlying the processing of associative memory follows general rules but is flexible enough to adapt to the particular timing required under natural conditions. Some researchers argue that the timing of memory stages in the honeybee reflects an adaptation to the requirements during foraging at distributed and unreliable food sources (e.g., flowers; Menzel, 1999; Giurfa et al., 2001).

Conclusion

The honeybee provides a model system for the study of neural substrates of low and intermediate levels of cognitive faculties. Neural analysis is supported by robust forms of associative learning that occur even under conditions when intracellular recordings or optophysiological measurements of single or multiple neuron activities are performed. The functional organization of the brain with a considerable number of uniquely identifiable neurons is also advantageous for relating cognitive functions to neural events in circumscribed circuits. Biochemical analyses of the role of protein kinases (e.g., PKA, PKC) and enzymes (e.g., NO synthase) relate directly to behavioral phenomena such as memory stages. Associative processes in the bee brain are not restricted to elementary forms but reflect configural processing and context-dependent associations. Thus the bee brain may serve as a model for the study of cognitive processes at an intermediate level of complexity.

See also:CONDITIONING, CELLULAR AND NETWORK SCHEMES FOR HIGHER-ORDER FEATURES OF; CONDITIONING, CLASSICAL AND INSTRUMENTAL; INSECT LEARNING; KAMIN'S BLOCKING EFFECT: NEURONAL SUBSTRATES; SECOND MESSENGER SYSTEMS

Bibliography

Bitterman, M. E. (1996). Comparative analysis of learning in honeybees. Animal Learning Behavior 24, 123-141.

Botzer, D., Markovich, S., and Susswein, A. J. (1998). Multiple memory processes following training that a food is inedible in Aplysia. Learning and Memory 5, 204-219.

Capaldi, E. A., Smith, A. D., Osborne, J. L., Fahrbach, S. E., Farris, S. M., Reynolds, D. R., Edwards, A. S., Martin, A., Robinson, G. E., Poppy, G. M., and Riley, J. R. (2000). Ontogeny of orientation flight in the honeybee revealed by harmonic radar. Nature 403, 537-540.

Couvillon, P. A., and Bitterman, M. E. (1988). Compound-component and conditional discrimination of colors and odors by honeybees: Further tests of a continuity model. Animal Learning and Behavior 16, 67-74.

Deisig, N., Lachnit, H., Hellstern, F., and Giurfa, M. (2001). Configural olfactory learning in honeybees: Negative and positive patterning discrimination. Learning and Memory 8, 70-78.

Dubnau, J., and Tully, T. (1998). Gene discovery in Drosophila, new insights for learning and memory. Annual Review of Neuroscience 21, 407-444.

Faber, T., Joerges, J., and Menzel, R. (1999). Associative learning modifies neural representations of odors in the insect brain. Nature Neuroscience 2, 74-78.

Funayama, E. S., Couvillon, P. A., and Bitterman, M. E. (1996). Compound conditioning in honeybees, blocking tests of the independence assumption. Animal Learning and Behavior 23, 429-437.

Gerber, B., and Ullrich, J. (1999). No evidence for olfactory blocking in honeybee classical conditioning. Journal of Experimental Biology 202, 1,839-1,854.

Giurfa, M., Eichmann, B., and Menzel, R. (1996). Symmetry perception in an insect. Nature 382, 458-461.

Giurfa, M., Zhang, S., Jenett, A., Menzel, R., and Srinivasan, M. V.(2001). The concepts of "sameness" and "difference" in an insect. Nature 410, 930-933.

Hammer, M. (1993). An identified neuron mediates the unconditioned stimulus in associative olfactory learning in honeybees. Nature 366, 59-63.

Hammer, M., and Menzel, R. (1998). Multiple sites of associative odor learning as revealed by local brain microinjections of octopamine in honeybees. Learning and Memory 5, 146-156.

Kamin, L. J. (1968). Attentionlike processes in classical conditioning. In M. R. Jones, ed., Miami symposium on predictability, behavior and aversive stimulation. pp. 9-32. Miami, FL: University of Miami Press.

Mackintosh, N. J. (1975). A theory of attention: Variations in the associability of stimuli with reinforcement. Psychology Review 82, 276-298.

Menzel, R. (1990). Learning, memory, and "cognition" in honey bees. In R. P. Kesner and D. S. Olton, eds., Neurobiology of comparative cognition. Hillsdale, NJ: Erlbaum.

—— (1999). Memory dynamics in the honeybee. Journal of Comparative Physiology[A] 185, 323-340.

Menzel, R., Brandt, R., Gumbert, A., Komischke. B., and Kunze, J. (2000). Two spatial memories for honeybee navigation. Proceedings of the Royal Society of London B 267, 961-968.

Menzel, R., and Giurfa, M. (2001). Cognitive architecture of a mini-brain: The honeybee. Trends in Cognitive Sciences 5, 62-71.

Menzel, R., and Müller, U. (1996). Learning and memory in honeybees: From behavior to neural substrates. Annual Review of Neuroscience 19, 379-404.

Müller, D., Gerber, B., Hellstern, F., Hammer, M., and Menzel, R.(2000). Sensory preconditioning in honeybees. Journal of Experimental Biology 203, 1,351-1,364.

Müller, U., and Carew, T. J. (1998). Serotonin induces temporally and mechanistically distinct phases of persistent PKA activity in Aplysia sensory neurons. Neuron 21, 1,423-1,434.

Rescorla, R. A., and Wagner, A. R. (1972). A theory of classical conditioning: Variations in the effectiveness of reinforcement and non-reinforcement. In A. H. Black A. H. and W. F. Prokasy, eds., Classical conditioning II: Current research and theory. New York: Appleton-Century-Crofts.

Rose, S. P. R. (1991). How chicks make memories: The cellular cascade from c-fos to dendritic remodelling. Trends in Neurosciences 14, 390-397.

Rudy, J. W., and Sutherland, R. J. (1992). Configural and elemental associations and the memory coherence problem. Journal of Cognitive Neuroscience 4, 208-216.

Smith, B. H., and Cobey, S. (1994). The olfactory memory of the honey bee, Apis mellifera. II: Blocking between odorants in binary mixtures. Journal of Experimental Biology 195, 91-108.

Thorn, R. S., and Smith, B. H. (1997). The olfactory memory of the honeybee Apis mellifera, III. Bilateral sensory input is necessary for induction and expression of olfactory blocking. Journal of Experimental Biology 200, 2,045-2,055.

RandolfMenzel