ALG Colloquium series

The ALG periodically invites colloquia from various visiting and local academics. Talks are usually held at the University of NSW or University of Sydney and are now arranged on an ad hoc basis (the regular fortnightly meetings that were held up until 2011 have been replaced with the winter workshop). If you have a visitor who would like to present their research to the ALG, please contact Ben Colagiuri (b.colagiuri AT

Schedule of Talks at Recent ALG Meetings

November 22, 2011

Mike Le Pelley, University of NSW.

The amazing alien fruit machine: Incentive value and attention in human associative learning.

In these experiments we look at whether the rate of learning about a cue, and attention to that cue, is influenced by the incentive value of the outcome that the cue predicts, using a novel contingency learning paradigm in humans; the alien fruit machine. We go on to demonstrate that the influence of incentive value on attention depends on whether the relevant relationships are trained or instructed, suggesting that learned changes in attention can be influenced by both habitual (automatic) and goal-directed (strategic) processes. This encourages the view that learned attentional responses can treated analogously to other examples of conditioned instrumental action, and has implications for clinical conditions involving maladaptive attentional biases.

May 3, 2011

Chris Mitchell, University of NSW.

Human and Animal Perceptual Learning.

March 22, 2011

Geoff Hall, University of York.

Latent Inhibition in Flavour-Preference Conditioning

I will describe three simple experiments looking at the effects of preexposure to the to-be-conditioned flavour on the acquisition (or expression) of a sucrose-based preference for that flavour. They demonstrate that the latent inhibition effect is critically dependent on the motivational state of the animal at time of test (no effect if the animal is not hungry). Could it really be that when the animal is not hungry it goes in for a special form of learning that does not obey the standard rules? I hope that our discussion will come up with some answer to this question.

2nd Semester, 2010

November 30, 2010

Natalie Tronson, Northwestern University.

Stress enhancement of fear conditioning is mediated by metabotropic glutamate receptors

Glutamatergic transmission is one of the main components of the stress response; nevertheless, its role in the emotional stress sequelae is not known. Here we investigated the role of type I metabotropic glutamate receptors (metabotropic glutamate receptors 1 and 5 [mGluR1 and mGluR5]) in the delayed and persistent enhancement of fear induced by acute stress. Whereas glutamate is the primary endogenous ligand, type I mGluRs can also be activated via ligand-independent dissociation from their "long" Homer scaffold proteins (Homer1-3) by the inducible, dominant negative "short" Homer1a. Antagonists and inverse agonists of mGluR1 and mGluR5 were injected into the hippocampus after immobilization stress and before contextual fear conditioning. Metabotropic glutamate receptor 5 was displaced from constitutive Homer scaffolds by viral transfection of Homer1a or injection of Tat decoy peptides. The effects of these manipulations on stress-enhanced fear were determined. We show that stress induces interactions between hippocampal mGluR5 and Homer1a; causes a sustained, ligand-independent mGluR5 activity; and enhances contextual fear. Consistent with this mechanism, enhancement of fear was abolished by delayed poststress application of inverse agonists, but not antagonists, of mGluR5. The effect of stress was mimicked by virally transfected Homer1a or injection of Tat-metabotropic glutamate receptor C-tail decoy peptides into the hippocampus. Constitutive activation of mGluR5 is thus a principal hippocampal mechanism underlying the delayed stress effects on emotion and memory.

November 23, 2010

Philip Corlett, Yale University.

Beliefs in the brain: Using fMRI to explore delusions

Delusions, the false and often tenacious beliefs that attend mental illness, are a challenge to cognitive neuroscience. I will discuss an attempt to meet that challenge by conceptualizing delusions as incorrect causal inferences. We have explored the causal inference process using functional neuroimaging. Using pharmacological tools to model psychosis as well as studying patients with delusions it has been possible to identify a key pathological process; prediction error. I will go on to discuss how this process has been implicated in the perceptual dysfunction associated with psychosis; perceiving is believing; as well as in the maintenance of delusions in the face of incontrovertible contradictory evidence.

November 2, 2010

Aurore San-Galli, University of Bordeaux.

Retrospective revaluation in rats: behavioural and neurobiological studies.

Retrospective revaluation (RR) refers to the ability of humans or animals to make inferences about an absent event, on the basis of events that were associated with it: After conditioning of a compound stimulus, altering the predictive value of one of the elements of the compound (competitor) affects the predictive value of the other, absent, element (target). RR involves a competitive mechanism, so that reinforced presentation (inflation) or non-reinforced presentation (extinction) of the competitor stimulus respectively decrease (backward blocking) or increase (unovershadowing) the predictive value of the target stimulus. This phenomenon is a challenge to traditional associationist models, which all assume that a cue must be present to allow learning about this cue. Studying RR in rats, and its neurobiological bases, was the aim of my thesis. The main results will be described in that talk. I first sought to identify an experimental situation allowing the observation of the RR phenomenon, still rarely described in animals. A conditioned taste aversion procedure revealed non-competitive mediation processes that were opposite to those expected in RR. Finally, a conditioned magazine approach was used to demonstrate RR with the inflation procedure, highlighting an important role of the relative salience of competitor and target stimuli. These results, confronted with current theoretical models of RR, support retrieval-based models such as the comparator hypothesis (Miller & Matzel, 1988). In a final part, I addressed the neuroanatomical bases of RR. A c-Fos immunohistochemistry study revealed an activation of a prefronto-striatal circuit (infralimbic and orbitofrontal cortices as well as nucleus accumbens) at the time of testing, consistent with the notion that important RR processes occur during retrieval of the competing associations. Although preliminary, this work opens the possibility to investigate in animals the neurobiological processes involved in a complex task closely related to human contingency learning.

October 12, 2010

Hiu Leung, Simon Killcross, and Fred Westbrook, UNSW.

Compound presentation of preexposed stimuli deepens latent inhibition.

A series of experiments examined the effect of compound stimulus preexposure on latent inhibition. Multiple stimuli were first preexposed individually and then compounded for further exposures. Stimulus given further exposures in compound with another preexposed stimulus was more latent inhibitory than one preexposed individually (Experiments 1-4). Latent inhibition was also deepened when a stimulus was given further preexposures in compound with a novel stimulus lacking prior exposures (Experiment 5), but this deepening was greater when both elements in the compound were preexposed (Experiments 6 and 7). These results were interpreted to mean that compound presentation of preexposed stimuli promotes attentional loss by reducing prediction error with nonreinforcement.

September 14, 2010

Sindy Cole, Rick Richardson and Gavan P. McNally, UNSW.

Kappa opioid receptors mediate where fear is expressed following extinction.

A neural circuit consisting of the hippocampus, medial prefrontal cortex and amygdala has been proposed to underlie the context-dependent expression of extinguished fear, however the exact pathways and pharmacological mechanisms through which this circuit function are poorly understood. These studies used a within-subjects ABA renewal design to examine the involvement of kappa opioid receptors (KORs) in the renewal of extinguished fear. Rats were trained to fear a tone conditioned stimulus (CS) via pairings with foot shock in a distinctive context (A), followed by extinction training of the tone in a second context (B). Finally all rats were tested for fear to the tone in both the extinction (ABB) and training (ABA) contexts. A pre-test i.c.v infusion of the KOR antagonist nor-BNI dose-dependently prevented renewal (Experiments 1 and 2) but had no effect on the expression of extinction (Experiment 3). Pre-test infusion of the KOR agonist U50,488H selectively facilitated the expression of renewal (Experiment 4). Finally pre-test infusion of nor-BNI had no effect on the expression of context-specific latent inhibition (Experiment 5). Collectively, these results suggest that KORs gate the expression of fear following extinction training and may comprise a critical pharmacological component of the neural circuit underlying the context-dependent expression of extinguished fear.

August 31, 2010

Ben Colagiuri, University of Sydney.

The Placebo Effect.

The placebo effect is an interesting phenomenon whereby real changes occur in response to an otherwise inert intervention. Much of the early work on placebo effects sought to determine whether they were caused by expectancy or classical conditioning. More recently, integrative models have been developed in which classical conditioning is considered as one source of expectancies that may produce placebo effects. However, these integrative models maintain that classical conditioning can produce placebo effects that are not consciously mediated. In the first half of the presentation I will discuss evidence for the role of classical conditioning in producing placebo effects and relevant theories of the placebo effect drawing attention to some potential inconsistencies with the broader literature on classical conditioning (e.g. the role of awareness and conditioned compensatory responses). In the second half, I will present the results of my research on placebo effects in double-blind RCTs, placebo-induced side effects, and placebo effects on implicit learning.

1st Semester, 2010

June 1, 2010

Bronwyn Graham, University of New South Wales.

How does fibroblast-growth factor-2 enhance fear extinction?

In recent years several behavioural and pharmacological manipulations that enhance extinction of conditioned fear have emerged, several of which have led to advancements in the treatment of anxiety disorders in humans. However, in many cases, the mechanisms by which these manipulations enhance extinction have remained unclear. Fibroblast growth factor-2 (FGF2) is a potent mitogen for several cell types and it has recently been shown that acute systemic administration of FGF2 enhances extinction of conditioned fear in rats (Graham & Richardson, Neuropsychopharmacology, 2009). This presentation will describe several experiments that have attempted to determine the mechanisms by which FGF2 enhances extinction. Specifically, the effect of FGF2 on reinstatement and renewal of extinguished fear is examined. Furthermore, the effects of MK-801 (an n-methyl-d-aspartate, or NMDA, antagonist) on the re-acquisition and re-extinction of a CS that was initially extinguished with FGF2 is examined. Together, these experiments indicate that FGF2 significantly reduces susceptibility to relapse following extinction. These experiments also show that extinction combined with FGF2 leads to NMDA dependent re-acquisition and re-extinction of original CS, processes that are NMDA independent in rats that do not receive FGF2 with initial extinction training. On the basis of these results it is suggested that FGF2 may render the original fear memory inaccessible, or considerably weakens the original memory trace.

May 18, 2010

Nathan Marchant, University of New South Wales.

Hypothalamus mediates both renewal and extinction of instrumental responses

These experiments studied the role of the hypothalamus, its afferents, and efferents in response recovery (ABA renewal) and extinction (ABB) of alcoholic beer seeking. The experiments employed a dual approach. The first was use of microinjections of inhibitory drugs into discrete hypothalamic nuclei to identify a causal role in the behaviours observed. The second was to combine neural tract-tracing with c-Fos (a marker of activity) to identify activation in specific neural circuits associated with both renewal and extinction. In the first set of experiments, functional inactivation of lateral hypothalamus (LH) blocked renewal of both sucrose and alcoholic beer seeking. Retrograde tracing experiments showed projections from nucleus accumbens shell (AcbSh) were differentially activated: dorsomedial AcbSh projections to LH showed increased c-Fos expression in rats tested for extinction (ABB), whereas ventral AcbSh projections to LH showed increased c-Fos expression in rats tested for renewal (ABA). The second set of experiments identified a causal role for the medial hypothalamus (MH) in mediating extinction. Infusions of the inhibitory neuropeptide cocaine- and amphetamine-regulated transcript (CART) reinstated responding in the extinction context. This effect was specific to the CART fragment, anatomically restricted to MH, specific to the expression of extinction, and dose-dependent. The anatomical experiments identified MH projections to paraventricular thalamus are activated when rats are tested for extinction (ABB), the majority of these neurons express the opioid peptide dynorphin. Finally, infralimbic prefrontal cortex neurons projecting to MH also showed increased activity in rats tested for extinction.

May 4, 2010

David Guez, University of Canberra.

From Cognition to Consciousness: a discussion about learning, reality representation and decision making

The scientific understanding of cognition and consciousness is currently hampered by the lack of rigorous and universally accepted definitions that permit comparative studies. This paper proposes new functional and unambiguous definitions for cognition and consciousness in order to provide clearly defined boundaries within which cognition and consciousness experiments may be conducted. The proposed definitions are built upon the construction and manipulation of reality representation, decision making and learning and are scoped in terms of an underlying logical structure. It is argued that the presentation of reality also necessitates the concept of absence and the capacity to perform transitive inference.

April 20, 2010

Roger Fulton, Brain and Mind Research Institute, University of Sydney

Functional Imaging of the Awake Rat Brain with microPET

Small laboratory animals must usually be anaesthetised or forcibly immobilised during microPET imaging to keep them within the scanner field of view and prevent motion artifacts in the reconstructed images. However in brain imaging these interventions can severely confound the interpretation of the acquired data by altering the very physiological parameters that we are trying to measure. The ability to image animals relatively unrestrained and without stress in their awake state could provide new information about the brain by enabling entirely new functional and behavioural experimental paradigms. In this talk we report the development and successful application of a motion correction technique that enables microPET brain imaging to be performed in awake rats using a conventional preclinical PET scanner.

March 30, 2010

Richard Morris, Prince of Wales Medical Research Institute.

Altered Bold Activation During Positive and Negative Prediction-Error Signals in Schizophrenia

Prediction-errors occur when there is a difference between received and expected rewards, and phasic firing of dopamine neurons directly encodes the sign of these errors. Errors may be “positive” (after unexpected reward) or “negative” (after a reward is omitted) and BOLD signals in fMRI experiments have been associated with positive and negative prediction-errors. However brain regions modulated by both positive and negative errors have rarely been examined in schizophrenia. We tested healthy adults and people with schizophrenia with a prediction-error task during fMRI to determine whether the BOLD signal can distinguish positive and negative prediction errors within the same dopamine-related brain regions. Among controls, positive and negative prediction-error signals were found in the midbrain, striatum and medial prefrontal cortex (FDR < 0.05), consistent with the prediction-error function of dopamine neurons. In both groups, prediction-error signals occurred in the midbrain and prefrontal cortex. Relative to controls, patients had attenuated responses to positive and negative prediction errors in the left ventral striatum and midbrain, but exaggerated responses to predicted outcomes in the right caudate. These results identify dysfunctional dopamine-related regions in people with schizophrenia, which may be related to the attentional and motivational deficits exhibited in this disease.

March 16, 2010

Gavan McNally, University of New South Wales.

Extinction, reconsolidation, and memory updating

A defining characteristic of Pavlovian extinction is that CR loss is not permanent. Instead, extinguished responding can be recovered under a number of circumstances. Two recent papers (Monfils et al., 2009, Science, 324, 951-955; Schiller et al., 2010, Nature, 463, 49-53) using aversive conditioning in rats and humans have reported that a single, recent non-reinforced exposure to the CS (‘retrieval’) prior to extinction training alters the nature of extinction learning and results in fear loss which is resistant to renewal, reinstatement, rapid re-acquisition, and spontaneous recovery. We will describe the results of six experiments from our laboratory that have studied this ‘retrieval – extinction’ effect in rats using both Pavlovian aversive and appetitive conditioning. Across the six experiments we document evidence for the acquisition, extinction, and recovery (renewal, reinstatement, re-acquisition) of responding regardless of whether extinction training was preceded or not by a single, non-reinforced exposure to the CS. We will also present some evidence that such recent exposures may, in fact, augment the recovery of extinguished responding. Reasons for the discrepancies between these results and those reported by Monfils et al. and Schiller et al. will be considered.

March 2, 2010

Joanna Fardell, University of Sydney.

Chemotherapy causes cognitive impairments in laboratory rodents

A significant subset of cancer patients who receive adjuvant chemotherapy report problems in working memory, executive function and processing speed that persist for 5-10 years after treatment has finished. These reports have been corroborated in studies using standardised neuropsychological testing. The work presented here shows that several chemotherapeutic agents cause a range of subtle impairments in laboratory animals free from cancer.

2nd Semester, 2009

Aug 11, 2009

Moriel Zelikowsky, University of California.

Contexts, fear and safety: insights from the hippocampus and infralimbic cortex

Animals will invariably learn to fear a place in which they have experienced an aversive event, so long as they have a good representation of that place (e.g. "context conditioning"). Animals may also use contexts to help reduce the ambiguity of a discrete cue with multiple meanings - such as a cue that predicts both danger and safety (e.g. "fear renewal after extinction"). A number of brain regions have been implicated in context learning, most notably the hippocampus. I will present some data here suggesting unique but also compensatory roles for the hippocampus and infralimbic cortex in various forms of context-sensitive fear. In particular, I am interested in the degree to which the roles of these structures may change according to the type of context learning that is occuring.

August 25, 2009

Matthew Bell, Santa Clara University

Signals, Resistance to Change, and Value

Behavioral Momentum Theory is ambiguous about the role of signals in controlling responding in chain schedules (A-B-C-food).The formal models suggest that the stimuli should acquire conditioned value through Pavlovian conditioning. Thus, one possibility is that the signals function as conditioned reinforcers. A second approach championed by the timing literature suggests that the signals merely function as cues for responding, lacking any conditioned value. My work focuses on the theoretical issue of whether or not stimuli hold "value" for organisms.

Sept 8, 2009

Kate Stevens, University of Western Sydney

Unspoken knowledge: implicit learning of structured human dance movement

The sequencing of dance movements may be thought of as a grammar. We investigate implicit learning of regularities that govern sequences of unfamiliar, discrete dance movements. It was hypothesized that observers without prior experience with contemporary dance are able to learn regularities that underpin structured human movement. Thirty-one adults were assigned to either an exposure or control group. Exposure consisted of 22 grammatical three-, four- and five-movement sequences presented twice in random order; sequence duration ranged from 9-19 s. In a test phase, exposure and control groups identified previously-unseen sequences as grammatical or ungrammatical, and rated confidence of judgment. The exposure group selected significantly more new grammatical sequences in the test phase than the control group. In addition, for the exposure group, the zero correlation criterion, wherein no relation between confidence and accuracy indicates unconscious knowledge, was satisfied. Through exposure, novice observers can learn a grammar that governs the sequencing of dance movements. This has implications for implicit learning of long sequences, working memory, and the development of expectations through exposure to contemporary dance.

Sept 22, 2009

Evan Livesey, University of Sydney.

Disturbing trends: Further tests of Perruchet's dissociation between response performance and event expectancy

Perruchet, Cleeremans, & Destrebecqz (2006) reported a striking dissociation between trends in the conscious expectancy of an event and the speed of a response which is cued by that event. They argue that this indicates the operation of independent processes in human associative learning. However, there remains a strong possibility that this dissociation is not a consequence of associative learning, and is instead caused by changes in vigilance or sensitivity based on the recency of events on previous trials. We tested this possibility using a similar task in which trends in performance cannot be explained by these non-associative factors. In each experiment, similar trends in performance were evident, suggesting a genuine automatic influence of associative learning on the execution of a cognitively controlled response. I will also discuss some other tasks requiring controlled attention that appear to yield similar trends in performance.

October 6, 2009

Part I: Pascal Carrive, University of New South Wales.
Part II: Nick Olsen, University of New South Wales.

Emotional stress: I. The role of orexin. II. Amygdala output: excitatory or inhibitory?

I. Orexin/hypocretin plays a major role in the control of arousal and wakefulness. Functional/anatomical and pharmacological experiments indicate that it also contributes to the expression of some forms of emotional stress but not others (eg, conditioned fear but not restraint). Interaction with the environment might be the critical factor.

II. Output neurons of the amygdala to the brainstem, which are mainly located in the medial part of the central nucleus, are GABAergic. Are these neurons tonically active and inhibited during conditioned emotional response or are they phasically active and excited during conditioned responses? Electrophysiological and Fos data suggest the first option is more likely.

Oct 20, 2009

Stephen Provost, Southern Cross University.

Latent inhibition in an unmasked stimulus equivalence task.

Latent inhibition refers to the decrease in learning obtained when a stimulus is pre-exposed before conditioning. The phenomenon has been widely investigated in animal preparations, and is central to some theoretical treatments of learning. The degree to which latent inhibition is obtained in human learning preparations is a matter of some doubt, however, and Lubow has argued that this will only be the case when the learning task is masked. I began to doubt this when an honours student (Kerrie Dennis) found evidence for latent inhibition in a stimulus equivalence task, in both the final equivalence tests and in the initial matching-to-sample task. This experiment was not specifically designed for this purpose, however, but a second student (Todd Lattimer) replicated the essential matching-to-sample component with appropriate controls. Closer examination of Todd's data has led to an unexpected and (I hope) interesting possibility: at least some of the data can be adequately explained in terms of a simple associative model (Rescorla & Wagner, 1972) if one makes some (plausible) assumptions about context-cue learning during the pre-exposure phase. I will present some simulations outlining this idea for discussion and comments.

1st Semester 2009

June 3, 2009

Andros Hoan, Macquarie University.

Generalisation and peak shift in remembering: Hens versus humans, and the dilemma of how to measure memory.

In behavioural terms, to remember a stimulus is to respond to that stimulus at a later time, i.e. remembering can be seen as discriminative behaviour along the stimulus dimension of time. This implies that learning phenomena such as stimulus generalization, which have been shown in discrimination learning along many other stimulus dimensions, may occur also in remembering. Data from operant studies with pigeons and hens have supported this, for short delays. An extrapolation from this view is that peak shift should also occur in remembering. Evidence for this was obtained from a DMTS study with domestic hens. The hens were poorly reinforced for correct remembering after a short delay, but strongly reinforced for correct remembering after a longer delay. As a result, they remembered best after an even longer delay. This radically counter-intuitive effect was not predicted by any traditional theory of memory. It cannot be explained by a trace decay process, and suggests that the "decay" curve typically found, is actually a result of or adaptation to the temporal structure and function of the animal's environment. However, attempts to show the same effect with humans have had much more variable results. Hen and human data will be presented and discussed in terms of the themes of remembering as active behaviour, the viability of basic operant (or respondent) research with humans, and audience feedback invited.

May 20, 2009

Jee Kim, University of New South Wales

More Findings on Extinction of Conditioned Fear across Development

Anxiety disorders is the most common of all psychological disorders, and cause a substantial personal, societal, and financial burden. The analysis of anxiety disorders in the lab employs the paradigm of extinction to improve the treatment of such disorders (e.g., exposure therapy). Research on adult animals strongly suggests that extinction depends, at least partly, on new inhibitory learning that is specific to the context in which it is learned. However, several recent studies show that extinction processes are dissociated across development. To summarize, various behavioural and neural studies show that extinction is context-, NMDA-, GABA-, and mPFC-independent in pre-weanling aged rats (i.e., 17-day-olds), whereas post-weanling aged rats (i.e., 24-day-olds) display adult-like extinction. Interestingly, our recent findings suggest that extinction is also different across adolescence in rats (i.e., 35-day-olds). The existing models of extinction cannot account for these developmental differences. These findings have significant clinical implications because they suggest that exposure therapy across different ages may make them less or more susceptible to relapse compared to the same treatments in adults.

May 6, 2009

Simon Killcross, University of New South Wales.

Fronto-striatal systems in fear conditioning.

Although aversively motivated Pavlovian learning centred on amygdala circuitry is regarded as the prototypical system subserving mammalian fear, instrumental learning also plays a key role in the adaptive response of animals to anxiety provoking cues and situations. A series of experiments examining the neural systems underpinning instrumental fear learning in the form of conditioned punishment will be presented, and will attempt to make the case for a critical role of cortical and subcortical systems in guiding animals' responses to fear-related cues.

April 22, 2009

Bob Boakes, University of Sydney.

The missing calories effect: Learned avoidance of flavours signalling reduction in a nutrient.

Food-deprived rats learned to avoid a flavour negatively correlated with access to a rich nutrient, 20% maltodextrin (20M) solution. This avoidance in 2-bottle choice tests was produced by training consisting of either an unpaired condition where sessions of unflavoured 20M were intermixed with sessions of 2 or 3% maltodextrin (2M or 3M) flavoured with salt (Experiment 1) or almond (Experiments 3 and 4) or a differential conditioning procedure where one flavour was mixed with 20M and another with 2M (Experiment 2). Avoidance was counterconditioned by mixing the target flavour with 20M (Experiment 1), generalized to a neutral context (Experiment 3) and displayed strong resistance to extinction (Experiment 4). The results demonstrated that food avoidance learning can occur in the absence of an aversive unconditioned stimulus and indicated that unpaired control groups and differential conditioning procedures may be misleading in flavour preference learning research when further control conditions are absent.

April 8, 2009

Hannah Salvin, University of Sydney.

Learning and Memory in Dogs - Testing Without Training.

Canine cognitive dysfunction (CCD) is a neurodegenerative disease which affects aged dogs. It has many pathological, physiological and behavioural similarities to human Alzheimer's disease. In developing the dog as a model, testing procedures for memory and learning have played a vital role. Current established protocols involve extensive training, often in excess of several months, before data can be collected. In addition to the time requirements, training based protocols can introduce handler/trainer bias, be influenced by motivation levels and run the risk of interference from the "clever Hans" effect.

March 25, 2009

Mike Le Pelley, Cardiff University

Attentional processes in associative learning

Animal conditioning studies indicate that the influence of the previously-experienced predictiveness of cues on the rate of subsequent learning about those cues depends on the manner in which predictiveness is established. Pretraining with multiple, simultaneously-presented cues, some of which are more predictive than others, tends to promote learning about more predictive cues, while pretraining with individual cues which are the best available predictors of reinforcement tends to promote learning about less predictive cues. A "hybrid" model incorporating two opposing attentional processes, both of which influence learning, is introduced to account for this discrepancy. The applicability of this model to attentional processes in human associative learning is then assessed in a series of experiments.

March 11, 2009

Bernard Balleine, University of California, Los Angeles

Dissociable role of endogenous opioids in reward detection and incentive learning

Common neuronal circuits are posited to mediate reward detection and the encoding and retrieval of incentive value used to direct instrumental actions. However, reward "liking" and "wanting" are argued to constitute dissociable processes in popular constructs of addiction. Furthermore, whereas opioid circuitry in the nucleus accumbens shell (NAs) and ventral pallidum (VP) has been implicated in affective responses to rewards, the opioid receptor rich basolateral amygdala (BLA) has been argued to encode the incentive value of instrumental rewards. Here we investigated whether these structures form a network controlling the detection of reward palatability and the encoding and retrieval of incentive value. Naloxone infused into the VP or NAs blocked food deprivation-induced increases in sucrose palatability, but not the increases in reward-seeking responses produced by incentive learning. Conversely, intra-BLA naloxone blocked the effects of incentive learning without affecting sucrose palatability, an effect specific to the encoding of incentive value; intra-BLA naloxone did not affect retrieval of previously updated incentive value information. This anatomical double-dissociation of opioid mediated reward detection and reward encoding determinants of goal value suggests that differential targeting of these processes may be possible in addiction treatments.