Effects Of Associative Learning Research Articles

How to solve novel problems: the role of associative learning in problem-solving performance in wild great tits Parus major

Although problem-solving tasks are frequently used to assess innovative ability, the extent to which problem-solving performance reflects variation in cognitive skills has been rarely formally investigated. Using wild breeding great tits facing a new non-food motivated problem-solving task, we investigated the role of associative learning in finding the solution, compared to multiple other non-cognitive factors. We first examined the role of accuracy (the proportion of contacts made with the opening part of a string-pulling task), neophobia, exploration, activity, age, sex, body condition and participation time on the ability to solve the task. To highlight the effect of associative learning, we then compared accuracy between solvers and non-solvers, before and after the first cue to the solution (i.e., the first time they pulled the string opening the door). We finally compared accuracy over consecutive entrances for solvers. Using 884 observations from 788 great tits tested from 2010 to 2015, we showed that, prior to initial successful entrance, solvers were more accurate and more explorative than non-solvers, and that females were more likely to solve the task than males. The accuracy of solvers, but not of non-solvers, increased significantly after they had the opportunity to associate string pulling with the movement of the door, giving them a first cue to the task solution. The accuracy of solvers also increased over successive entrances. Our results demonstrate that variations in problem-solving performance primarily reflect inherent individual differences in associative learning, and are also to a lesser extent shaped by sex and exploratory behaviour.

No pupillometric evidence for effortful proactive control in the proportion-congruent Stroop paradigm.

Cognitive control is the ability to allocate attention away from stimuli that are irrelevant to achieving a goal, towards stimuli that are. When conflict is anticipated, attention is biased in a global, top-down manner called proactive control and this effortful type of cognitive control is engaged before stimulus onset. The list-wise congruency proportion (LWPC) effect, where the Stroop congruency effect is reduced when there are more incongruent than congruent trials compared to vice versa, has been viewed as one of the prime signatures of this type of cognitive control. However, there has been recent debate about the extent to which this effect should be attributed to proactive control instead of alternative explanations such as simpler associative learning or reactive control. Thus, by using pupillometry (i.e., an indicator of cognitive effort), the present study investigated the extent to which LWPC effects result from effortful proactive control. Experiment 1 employed a classic proportion congruency manipulation, while Experiment 2 replaced congruent trials with neutral trials to control for potential effects of associative learning. While in line with past findings, proportion congruency effects were obtained in response times of both experiments and pupillometry showed both proportion congruency and Stroop effects after stimulus onset, no differences in pupil sizes were found during the preparatory phase. Therefore, these results do not support the idea that the observed LWPC effects are due to participants engaging in effortful proactive control.

Combined influence of valence and statistical learning on the control of attention II: Evidence from within-domain additivity.

Attention is biased in favor of stimuli that signal either threat or reward; this experience-dependent attentional bias develops via associative learning and persists into extinction. Physically salient yet task-irrelevant stimuli are also prioritized by the attention system, but the attentional priority of a physically salient distractor can be suppressed when it appears in a location in which it has been frequently encountered in the past. Similar effects of statistical learning on distractor suppression have been observed for distractors appearing in a predictable color. A pair of recent studies demonstrate that statistically learned distractor suppression and valence-based attentional biases combine additively, suggesting independent influences of learning on attentional priority. One limitation of these prior studies, however, is that the effects of statistical learning were defined with respect to spatial attention and the effects of associative learning with respect to feature-based attention. A strong version of the independence account would predict additive influences on attention even when both sources of priority are represented within a single domain of attentional control, which we tested in the present study. The attentional priority of a distractor was elevated when its color was previously associated with electric shock and reduced when its shape was frequently encountered as a distractor in a prior training phase, with these two influences on priority combining additively. Our findings provide strong evidence for the idea that statistical learning and valance-based associative learning exert independent influences on the control of attention, which has implications for contemporary theories of selection history.

Contextual effects in sensorimotor adaptation adhere to associative learning rules.

Traditional associative learning tasks focus on the formation of associations between salient events and arbitrary stimuli that predict those events. This is exemplified in cerebellar-dependent delay eyeblink conditioning, where arbitrary cues such as a tone or light act as conditioned stimuli (CSs) that predict aversive sensations at the cornea (unconditioned stimulus [US]). Here, we ask if a similar framework could be applied to another type of cerebellar-dependent sensorimotor learning - sensorimotor adaptation. Models of sensorimotor adaptation posit that the introduction of an environmental perturbation results in an error signal that is used to update an internal model of a sensorimotor map for motor planning. Here, we take a step toward an integrative account of these two forms of cerebellar-dependent learning, examining the relevance of core concepts from associative learning for sensorimotor adaptation. Using a visuomotor adaptation reaching task, we paired movement-related feedback (US) with neutral auditory or visual contextual cues that served as CSs. Trial-by-trial changes in feedforward movement kinematics exhibited three key signatures of associative learning: differential conditioning, sensitivity to the CS-US interval, and compound conditioning. Moreover, after compound conditioning, a robust negative correlation was observed between responses to the two elemental CSs of the compound (i.e. overshadowing), consistent with the additivity principle posited by theories of associative learning. The existence of associative learning effects in sensorimotor adaptation provides a proof-of-concept for linking cerebellar-dependent learning paradigms within a common theoretical framework.

Rating expectations can slow aversive reversal learning.

The process of learning allows organisms to develop predictions about outcomes in the environment, and learning is sensitive to both simple associations and higher order knowledge. However, it is unknown whether consciously attending to expectations shapes the learning process itself. Here, we directly tested whether rating expectations shapes arousal during classical conditioning. Participants performed an aversive learning paradigm wherein one image (CS+) was paired with shock on 50% of trials, while a second image (CS−) was never paired with shock. Halfway through the task, contingencies reversed. One group of participants rated the probability of upcoming shock on each trial, while the other group made no online ratings. We measured skin conductance response (SCR) evoked in response to the CS and used traditional analyses as well as quantitative models of reinforcement learning to test whether rating expectations influenced arousal and aversive reversal learning. Participants who provided online expectancy ratings displayed slower learning based on a hybrid model of adaptive learning and reduced reversal of SCR relative to those who did not rate expectations. Mediation analysis revealed that the effect of associative learning on SCR could be fully explained through its effects on subjective expectancy within the group who provided ratings. This suggests that the act of rating expectations reduces the speed of learning, likely through changes in attention, and that expectations directly influence arousal. Our findings indicate that higher order expectancy judgments can alter associative learning.

Experimental manipulation of muscularity preferences through visual diet and associative learning.

Body preferences are somewhat flexible and this variability may be the result of one’s visual diet (whereby mere exposure to certain bodies shifts preferences), associative learning mechanisms (whereby cues to health and status within the population are internalised and affect body preferences), or a mixture of both visual diet and associative learning effects. We tested how these factors may drive changes in preferences for muscularity in male bodies across a male and female sample. Three studies were conducted where participants viewed manipulation images of high and/or low muscle mass males which were either aspirational (high status clothing and posture) and/or neutral (no obvious cues to status). Preferences for muscularity were recorded before and after exposure to such manipulation images to assess whether body preferences had changed following manipulation. We found evidence for both the visual diet and associative learning hypotheses. Exposure to non-muscular male bodies decreased preferences for muscular bodies irrespective of image valence. Exposure to aspirational non-muscular male bodies alongside neutral muscular male bodies also led to a decrease in muscularity preferences. Further, when manipulation conditions are less obviously skewed towards a particular body type, preferences still shifted in the direction of the most prevalent body type, suggesting that demand characteristics are unlikely to have confounded results of previous adaptation experiments with more obvious manipulations.

Odor-Taste Interactions in Food Perception: Exposure Protocol Shows No Effects of Associative Learning.

Repeated exposure can change the perceptual and hedonic features of flavor. Associative learning during which a flavor’s odor component is affected by co-exposure with taste is thought to be central in this process. However, changes can also arise due to exposure to the odor in itself. The aim of this study was to dissociate effects of associative learning from effects of exposure without taste by repeatedly presenting one odor together with sucrose and a second odor alone. Sixty individuals attended two testing sessions separated by a 5-day Exposure Phase during which the stimuli were presented as flavorants in chewing gums that were chewed three times daily. Ratings of odor sweetness, odor pleasantness, odor intensity enhancement by taste, and odor referral to the mouth were collected at both sessions. Consistent with the notion that food preferences are modulated by exposure, odor pleasantness increased between the sessions independently of whether the odor (basil or orange flower) had been presented with or without sucrose. However, we found no evidence of associative learning in any of the tasks. In addition, exploratory equivalence tests suggested that these effects were either absent or insignificant in magnitude. Taken together, our results suggest that the hypothesized effects of associative learning are either smaller than previously thought or highly dependent on the experimental setting. Future studies are needed to evaluate the relative support for these explanations and, if experimental conditions can be identified that reliably produce such effects, to identify factors that regulate the formation of new odor–taste associations.

Different brain networks mediate the effects of social and conditioned expectations on pain

Information about others’ experiences can strongly influence our own feelings and decisions. But how does such social information affect the neural generation of affective experience, and are the brain mechanisms involved distinct from those that mediate other types of expectation effects? Here, we used fMRI to dissociate the brain mediators of social influence and associative learning effects on pain. Participants viewed symbolic depictions of other participants’ pain ratings (social information) and classically conditioned pain-predictive cues before experiencing painful heat. Social information and conditioned stimuli each had significant effects on pain ratings, and both effects were mediated by self-reported expectations. Yet, these effects were mediated by largely separable brain activity patterns, involving different large-scale functional networks. These results show that learned versus socially instructed expectations modulate pain via partially different mechanisms—a distinction that should be accounted for by theories of predictive coding and related top-down influences.

Associative Learning of Stimuli Paired and Unpaired With Reinforcement: Evaluating Evidence From Maggots, Flies, Bees, and Rats.

Finding rewards and avoiding punishments are powerful goals of behavior. To maximize reward and minimize punishment, it is beneficial to learn about the stimuli that predict their occurrence, and decades of research have provided insight into the brain processes underlying such associative reinforcement learning. In addition, it is well known in experimental psychology, yet often unacknowledged in neighboring scientific disciplines, that subjects also learn about the stimuli that predict the absence of reinforcement. Here we evaluate evidence for both these learning processes. We focus on two study cases that both provide a baseline level of behavior against which the effects of associative learning can be assessed. Firstly, we report pertinent evidence from Drosophila larvae. A re-analysis of the literature reveals that through paired presentations of an odor A and a sugar reward (A+) the animals learn that the reward can be found where the odor is, and therefore show an above-baseline preference for the odor. In contrast, through unpaired training (A/+) the animals learn that the reward can be found precisely where the odor is not, and accordingly these larvae show a below-baseline preference for it (the same is the case, with inverted signs, for learning through taste punishment). In addition, we present previously unpublished data demonstrating that also during a two-odor, differential conditioning protocol (A+/B) both these learning processes take place in larvae, i.e., learning about both the rewarded stimulus A and the non-rewarded stimulus B (again, this is likewise the case for differential conditioning with taste punishment). Secondly, after briefly discussing published evidence from adult Drosophila, honeybees, and rats, we report an unpublished data set showing that relative to baseline behavior after truly random presentations of a visual stimulus A and punishment, rats exhibit memories of opposite valence upon paired and unpaired training. Collectively, the evidence conforms to classical findings in experimental psychology and suggests that across species animals associatively learn both through paired and through unpaired presentations of stimuli with reinforcement – with opposite valence. While the brain mechanisms of unpaired learning for the most part still need to be uncovered, the immediate implication is that using unpaired procedures as a mnemonically neutral control for associative reinforcement learning may be leading analyses astray.

Constitutive exposure to the volatile methyl salicylate reduces per‐capita foraging efficiency of a generalist predator to learned prey associations

AbstractUnderstanding the factors that influence the ability of predators to find and kill herbivores is central to enhancing their impact on prey populations, but few studies have tested the impact of these factors on predation rates in realistic foraging environments. Using the tri‐trophic system consisting of tomato, Solanum lycopersicum L. (Solanaceae), hornworm caterpillars, Manduca sexta L. (Lepidoptera: Sphingidae), and the predaceous stink bug Podisus maculiventris (Say) (Hemiptera: Pentatomidae), we measured the effects of associative learning and plant volatile camouflage on predator behavior and foraging efficiency in field enclosures. To do so, we compared experienced vs. naive individuals under varying environmental contexts. Experienced predators were those with prior exposure to induced volatiles from the tomato–caterpillar association, whereas naive predators had not experienced tomato, only prey (caterpillars). We varied their environmental foraging matrix using either (1) tomato surrounded by basil (Ocimum basilicum L.; Lamiaceae) or (2) tomato exposed to the synthetic volatile, methyl salicylate (MeSA). We found that (1) experienced predators were more efficient than naive predators, capturing 28% more prey; (2) the tomato–basil combination did not affect predator–prey interactions; and (3) constitutive emission of synthetic MeSA caused a 22% reduction in P. maculiventris predation rate. These differences corresponded with distinct shifts in predator foraging; for example, experienced individuals were less stationary and exhibited unique behaviors such as stylet extension. Taken together, these results suggest that it is possible to improve the function of generalist predators in suppressing prey by coupling odors with food. However, constitutive emission of volatiles to attract natural enemies may ultimately camouflage neighboring plants, reducing the benefits of orientation to learned stimuli such as induced volatiles.

Reward and punishment learning in daily life: A replication study.

Day-to-day experiences are accompanied by feelings of Positive Affect (PA) and Negative Affect (NA). Implicitly, without conscious processing, individuals learn about the reward and punishment value of each context and activity. These associative learning processes, in turn, affect the probability that individuals will re-engage in such activities or seek out that context. So far, implicit learning processes are almost exclusively investigated in controlled laboratory settings and not in daily life. Here we aimed to replicate the first study that investigated implicit learning processes in real life, by means of the Experience Sampling Method (ESM). That is, using an experience-sampling study with 90 time points (three measurements over 30 days), we prospectively measured time spent in social company and amount of physical activity as well as PA and NA in the daily lives of 18-24-year-old young adults (n = 69 with anhedonia, n = 69 without anhedonia). Multilevel analyses showed a punishment learning effect with regard to time spent in company of friends, but not a reward learning effect. Neither reward nor punishment learning effects were found with regard to physical activity. Our study shows promising results for future research on implicit learning processes in daily life, with the proviso of careful consideration of the timescale used. Short-term retrospective ESM design with beeps approximately six hours apart may suffer from mismatch noise that hampers accurate detection of associative learning effects over time.

Conditioned Stimulus Form Does Not Explain Failures to See Pavlovian-Instrumental-Transfer With Ethanol-Paired Conditioned Stimuli.

Pavlovian-Instrumental-Transfer (PIT) examines the effects of associative learning upon instrumental responding. Previous studies examining PIT with ethanol (EtOH)-maintained responding showed increases in responding following presentation of an EtOH-paired conditioned stimulus (CS). Recently, we conducted 2 studies examining PIT with an EtOH-paired CS. One of these found increases in responding, while the other did not. This less robust demonstration of PIT may have resulted from the form of the CS used, as we used a 120-second light stimulus as a CS, while the previous studies used either a 120-second auditory stimulus or a 10-second light stimulus. This study examined whether using conditions similar to our earlier study, but with either a 120-second auditory or a 10-second light stimulus as a CS, resulted in more robust PIT. We also examined the reliability of our previous failure to observe PIT. Three experiments were conducted examining whether PIT was obtained using (i) a 120-second light stimulus, (ii) a 10-second light stimulus, or (iii) a 120-second auditory stimulus as CSs. We found PIT was not obtained using (i) a 120-second light stimulus as a CS, (ii) a 10-second light stimulus as a CS, or (iii) a 120-second auditory stimulus as a CS. These results suggest that CS form does not account for our earlier failure to see PIT. Rather, factors like rat strain or how EtOH drinking is induced may account for when PIT is or is not observed.

Associative learning as higher order cognition: Learning in human and nonhuman animals from the perspective of propositional theories and relational frame theory.

We aim to provide a new perspective on the old debate about whether evidence for higher order cognition in nonhuman animals can be reinterpreted in terms of associative learning. Our starting point is the idea that associative learning is best thought of as an effect (i.e., the impact of paired events on behavior) rather than a specific mental process (e.g., the formation of associations). This idea allows us to consider (a) propositional theories according to which associative learning is mediated by higher order mental processes akin to problem solving and (b) relational frame theory that allows one to think of seemingly simple associative learning effects as instances of a complex phenomenon known as arbitrarily applicable relational responding. Based on these 2 theories, we argue that (a) higher order cognition and associative learning are not necessarily mutually exclusive and (b) a more sophisticated conceptualization of higher order cognition is warranted. (PsycINFO Database Record

Effect of Associative Learning on Memory Spine Formation in Mouse Barrel Cortex.

Associative fear learning, in which stimulation of whiskers is paired with mild electric shock to the tail, modifies the barrel cortex, the functional representation of sensory receptors involved in the conditioning, by inducing formation of new inhibitory synapses on single-synapse spines of the cognate barrel hollows and thus producing double-synapse spines. In the barrel cortex of conditioned, pseudoconditioned, and untreated mice, we analyzed the number and morphological features of dendritic spines at various maturation and stability levels: sER-free spines, spines containing smooth endoplasmic reticulum (sER), and spines containing spine apparatus. Using stereological analysis of serial sections examined by transmission electron microscopy, we found that the density of double-synapse spines containing spine apparatus was significantly increased in the conditioned mice. Learning also induced enhancement of the postsynaptic density area of inhibitory synapses as well as increase in the number of polyribosomes in such spines. In single-synapse spines, the effects of conditioning were less pronounced and included increase in the number of polyribosomes in sER-free spines. The results suggest that fear learning differentially affects single- and double-synapse spines in the barrel cortex: it promotes maturation and stabilization of double-synapse spines, which might possibly contribute to permanent memory formation, and upregulates protein synthesis in single-synapse spines.

Investigating Associative Learning Effects in Patients with Prodromal Alzheimer's Disease Using the Temporal Context Model.

The purpose of this study was to investigate associative learning effects in patients with prodromal Alzheimer's disease (prAD) by referring to the Temporal Context Model (TCM; Howard, Jing, Rao, Provyn, & Datey, 2009), in an attempt to enhance the understanding of their associative memory impairment. TCM explains fundamental effects described in classical free-recall tasks and cued-recall tasks involving overlapping word pairs (e.g., A-B, B-C), namely (1) the contiguity effect, which is the tendency to successively recall nearby items in a list, and (2) the observation of backward (i.e., B-A) and transitive associations (i.e., A-C) between items. In TCM, these effects are hypothesized to rely on contextual representation, binding and retrieval processes, which supposedly depend on hippocampal and parahippocampal regions. As these regions are affected in prAD, the current study investigated whether prAD patients would show reduced proportions of backward and transitive associations in free and cued-recall, coupled to a reduced contiguity effect in free-recall. Seventeen older controls and 17 prAD patients performed a cued-recall task involving overlapping word pairs and a final free-recall task. Proportions of backward and transitive intrusions in cued-recall did not significantly differ between groups. However, in free-recall, prAD patients demonstrated a reduced contiguity effect as well as reduced proportions of backward and transitive associations compared to older controls. These findings are discussed within the hypothesis that the contextual representation, binding and/or retrieval processes are affected in prAD patients compared to healthy older individuals.

Effects of homeostatic constraints on associative memory storage and synaptic connectivity of cortical circuits.

The impact of learning and long-term memory storage on synaptic connectivity is not completely understood. In this study, we examine the effects of associative learning on synaptic connectivity in adult cortical circuits by hypothesizing that these circuits function in a steady-state, in which the memory capacity of a circuit is maximal and learning must be accompanied by forgetting. Steady-state circuits should be characterized by unique connectivity features. To uncover such features we developed a biologically constrained, exactly solvable model of associative memory storage. The model is applicable to networks of multiple excitatory and inhibitory neuron classes and can account for homeostatic constraints on the number and the overall weight of functional connections received by each neuron. The results show that in spite of a large number of neuron classes, functional connections between potentially connected cells are realized with less than 50% probability if the presynaptic cell is excitatory and generally a much greater probability if it is inhibitory. We also find that constraining the overall weight of presynaptic connections leads to Gaussian connection weight distributions that are truncated at zero. In contrast, constraining the total number of functional presynaptic connections leads to non-Gaussian distributions, in which weak connections are absent. These theoretical predictions are compared with a large dataset of published experimental studies reporting amplitudes of unitary postsynaptic potentials and probabilities of connections between various classes of excitatory and inhibitory neurons in the cerebellum, neocortex, and hippocampus.

Behavioral mechanisms underlying nicotine reinforcement.

Cigarette smoking is the leading cause of preventable deaths worldwide, and nicotine, the primary psychoactive constituent in tobacco, drives sustained use. The behavioral actions of nicotine are complex and extend well beyond the actions of the drug as a primary reinforcer. Stimuli that are consistently paired with nicotine can, through associative learning, take on reinforcing properties as conditioned stimuli. These conditioned stimuli can then impact the rate and probability of behavior and even function as conditioning reinforcers that maintain behavior in the absence of nicotine. Nicotine can also act as a conditioned stimulus (CS), predicting the delivery of other reinforcers, which may allow nicotine to acquire value as a conditioned reinforcer. These associative effects, establishing non-nicotine stimuli as conditioned stimuli with discriminative stimulus and conditioned reinforcing properties as well as establishing nicotine as a CS, are predicted by basic conditioning principles. However, nicotine can also act non-associatively. Nicotine directly enhances the reinforcing efficacy of other reinforcing stimuli in the environment, an effect that does not require a temporal or predictive relationship between nicotine and either the stimulus or the behavior. Hence, the reinforcing actions of nicotine stem both from the primary reinforcing actions of the drug (and the subsequent associative learning effects) as well as the reinforcement enhancement action of nicotine which is non-associative in nature. Gaining a better understanding of how nicotine impacts behavior will allow for maximally effective tobacco control efforts aimed at reducing the harm associated with tobacco use by reducing and/or treating its addictiveness.

A Test of Learned Industriousness in the Physical Activity Domain.

The Theory of Learned Industriousness states that durable individual differences in industriousness are due in part to differences in the extent to which individuals were rewarded for high effort at an earlier time. Individuals rewarded for high effort during training are thought to generalize greater persistence to subsequent tasks than those rewarded for low effort. This study tested whether rewarded physical and/or mental effort at different intensities generalized to greater persistence at a subsequent mental task. 80 inactive 18-25 year-olds were randomized into four groups: Low Mental Effort, High Mental Effort, Low Physical Effort, and High Physical Effort. Each completed group-specific effort training and a mental persistence task at baseline and posttest. Factorial analysis of covariance revealed a significant domain x effort interaction on persistence (F[1,75]=4.93, p=.029). High Mental Effort and Low Mental Effort groups demonstrated similar gains in persistence (d=-0.08, p>.05) and points earned (d=0.11, p>.05) following effort training. High Physical Effort and Low Physical Effort diverged on persistence (d=-0.49, p=.004) but not points earned (d =-0.12, p>.05). Findings suggest either that training and test stimuli were too dissimilar to cue effects of associative learning in physical effort groups, or that effects were present but overpowered by the affective and neurocognitive consequences of an acute bout of intense aerobic physical activity. Findings do not support the Theory of Learned Industriousness nor generalization of effort across physical and mental domains.

Conflict-triggered top-down control: Default mode, last resort, or no such thing?

The conflict monitoring account posits that globally high levels of conflict trigger engagement of top-down control; however, recent findings point to the mercurial nature of top-down control in high conflict contexts. The current study examined the potential moderating effect of associative learning on conflict-triggered top-down control engagement by testing the Associations as Antagonists to Top-Down Control (AATC) hypothesis. In 4 experiments, list-wide proportion congruence was manipulated, and conflict-triggered top-down control engagement was examined by comparing interference for frequency-matched, 50% congruent items across mostly congruent (low conflict) and mostly incongruent (high conflict) lists. Despite the fact that global levels of conflict were varied identically across experiments, evidence of conflict-triggered top-down control engagement was selective to those experiments in which responses could not be predicted on the majority of trials via simple associative learning, consistent with the AATC hypothesis. In a 5th experiment, older adults showed no evidence of top-down control engagement under conditions in which young adults did, a finding that refined the interpretation of the patterns observed in the prior experiments. Collectively, these findings suggest that top-down control engagement in high conflict contexts is neither the default mode nor an unused (or nonexistent) strategy. Top-down control is best characterized as a last resort that is engaged when reliance on one's environment, and in particular associative responding, is unproductive for achieving task goals.

Multiple memory traces after associative learning in the honey bee antennal lobe

We investigated the effect of associative learning on early sensory processing, by combining classical conditioning with in vivo calcium-imaging of secondary olfactory neurons, the projection neurons (PNs) in the honey bee antennal lobe (AL). We trained bees in a differential conditioning paradigm in which one odour (A+) was paired with a reward, while another odour (B-) was presented without a reward. Two to five hours after differential conditioning, the two odour-response patterns became more different in bees that learned to discriminate between A and B, but not in bees that did not discriminate. This learning-related change in neural odour representations can be traced back to glomerulus-specific neural plasticity, which depended on the response profile of the glomerulus before training. (i) Glomeruli responding to A but not to B generally increased in response strength. (ii) Glomeruli responding to B but not to A did not change in response strength. (iii) Glomeruli responding to A and B decreased in response strength. (iv) Glomeruli not responding to A or B increased in response strength. The data are consistent with a neural network model of the AL, which we based on two plastic synapse types and two well-known learning rules: associative, reinforcer-dependent Hebbian plasticity at synapses between olfactory receptor neurons (ORNs) and PNs; and reinforcer-independent Hebbian plasticity at synapses between local interneurons and ORNs. The observed changes strengthen the idea that odour learning optimizes odour representations, and facilitates the detection and discrimination of learned odours.