Search

Classical conditioning has been explained by two main types of theories that postulate different learning mechanisms. Rescorla and Wagner (1972) put forth a theory in which conditioning is based on the ability of the US to drive learning through error correction. Alternatively, Mackintosh (1973) put forth a theory in which the ability of the CS to be associated with the unconditioned stimulus is modulated. We have proposed a reconciliation of these two mechanisms as working in parallel within different neural systems: a cerebellar system for US modulation and a hippocampal system for CS modulation. We developed a computational model of cerebellar function in eyeblink conditioning based on the error correction mechanism of the Rescorla-Wagner rule in which learningrelated activity from the cerebellum inhibits the inferior olive, which is the US input pathway to the cerebellum (Gluck et al., 1994). We developed a computational model of the hippocampal region that forms altered representations of conditioned stimuli based on their behavioral outcomes (Gluck & Myers, 1993; Myers et al., 1995). Overall, computational modeling and empirical findings support the idea that, at least in the case of eyeblink conditioning, there may be two different neural systems: the cerebellum which mediates US-based error correction and hippocampus which alters representations of CSs.

Hypoxic brain injury can cause severe anterograde amnesia with intact intelligence and attention in select patients. This neuropsychological profile is similar among a broad class of etiologies resulting in damage to the hippocampus and associated medial temporal structures. Several recent studies have demonstrated that subjects with anterograde amnesia can nonetheless acquire simple associative tasks. In the present study, the performance of subjects with anterograde memory impairments resulting from hypoxic brain injury on a conditional spatial discrimination was examined. Quantitative magnetic resonance imaging revealed significant hippocampal atrophy in the hypoxic subjects. The hypoxic subjects were able to acquire and reverse the discrimination, although they were slower relative to matched control subjects. Hypoxic subjects also showed a tendency to perseverate after reward contingencies were reversed.

Recent evidence suggests that (a) auditory cortical neurons are tuned to complex timevarying acoustic features, (b) auditory cortex consists of several "elds that decompose sounds in parallel, (c) the metric for such decomposition varies across species, and (d) auditory cortical representations can be rapidly modulated. Past computational models of auditory cortical processing cannot capture such representational complexity. This paper proposes a novel framework in which auditory signal processing is characterized as an adaptive transformation from a one-dimensional space into an n-dimensional auditory parameter space. This transformation can be modeled as a chirplet transform implemented via a self-organizing neural network. ( 2000 Elsevier Science B.V. All rights reserved.

Gluck and Myers (Hippocampus 3 (1993) 491–516) modeled the hippocampus as an auto-encoder; Myers et al. (Neurobiol. Learning Memory 66 (1996) 51–66) argued that the cholinergic input from medial septum modulates learning rate in this auto-encoder. Neurophysiological evidence suggests the hippocampus self-regulates septal acetylcholine release in response to novel stimuli (Hasselmo and Schnell, J. Neurosci. 14 (1994) 3898–3914). We have extended our earlier model of septohippocampal modulation to include such a feedback loop. The resulting dynamic model learns faster and better than the earlier version on phenomena such as blocking and shift reversal. It can also be applied to data regarding the effects of the anticholinergic drug scopolamine.

▪ Abstract We review current computational models of hippocampal function in learning and memory, concentrating on those that make strongest contact with psychological issues and behavioral data. Some models build upon Marr's early theories for modeling hippocampal field CA3's putative role in the fast, temporary storage of episodic memories. Other models focus on hippocampal involvement in incrementally learned associations, such as classical conditioning. More recent efforts have attempted to bring functional interpretations of the hippocampal region in closer contact with underlying anatomy and physiology. In reviewing these psychobiological models, three major themes emerge. First, computational models provide the conceptual glue to bind together data from multiple levels of analysis. Second, models serve as important tools to integrate data from both animal and human studies. Third, previous psychological models that capture important behavioral principles of memory provide an important top-down constraint for developing computational models of the neural bases of these behaviors.

In recent modeling of hippocampal function, we have at- tempted to integrate formal behavioral analyses of classical conditioning with psychobiological data on brain lesions (Gluck and Myers (1993) Hippocampus 3:491-516; Myers and Gluck (I9941 Behav Neurosci 108(5):835-847). Based on comparative behavioral analyses, we have ar- gued that animals with hippocampal region damage are unable to alter stimulus similarity based on experience. While hippocampal-damaged an- imals can still learn whether to respond to an individual stimulus, they are notably impaired at many tasks involving learning relationships be- tween stimuli-especially in the absence of explicit reinforcement. These analyses lead to a computational theory which identifies two represen- tational recoding processes-predictive differentiation and redundancy compression-which alter stimulus similarity relationships in intact ani- mals but are dependent on intact hippocampal region processing. More recent, and ongoing, modeling aims to broaden this model of hippocam- pal region function in classical conditioning, with an emphasis on phys- iological and anatomical constraints, including the role of the fornix and subcortical modulation, preprocessing in sensory cortices, and localiza- tion of the proposed representational functions within more precisely identified hippocampal region substrates (Myers et al. (1995) Psychology 23(2):116-138; Myers and Gluck (19961 Behav Neurosci; Myers et al. (1996) Neurobiol learning Memory). Working to bridge between behav- ioral and physiological levels of analysis, we ultimately hope to develop a more complete understanding of hippocampal region function in mem- ory across a wider range of behavioral paradigms, elucidating how this functionality emerges from underlying physiological and anatomical sub- strates. 0 1997 Wiley-Liss, Inc.

Behavioral inhibition (BI) is an anxiety vulnerability factor associated with hypervigilance to novel stimuli, threat, and ambiguous cues. The progression from anxiety risk to a clinical disorder is unknown, although the acquisition of defensive learning and avoidance may be a critical feature. As the expression of avoidance is also central to anxiety development, the present study examined avoidance acquisition as a function of inhibited temperament using classical eyeblink conditioning. Individuals were classified as behaviorally inhibited (BI) or non-inhibited (NI) based on combined scores from the Adult and Retrospective Measures of Behavioural Inhibition (AMBI and RMBI, respectively). Acquisition was assessed using delay, omission, or yoked conditioning schedules of reinforcement. Omission training was identical to delay, except that the emission of an eyeblink conditioned response (CR) resulted in omission of the unconditioned airpuff stimulus (US) on that trial. Each subject in the yoked group was matched on total BI score to a subject in the omission group, and received the same schedule of CS and US delivery, resulting in a partial reinforcement training schedule. Delay conditioning elicited significantly more CRs compared to the omission and yoked contingencies, the latter two of which did not differ from each other. Thus, acquisition of an avoidance response was not apparent. BI individuals demonstrated enhanced acquisition overall, while partial reinforcement training significantly distinguished between BI and NI groups. Enhanced learning in BI may be a function of an increased defensive learning capacity, or sensitivity to uncertainty. Further work examining the influence of BI on learning acquisition is important for understanding individual differences in disorder etiology in anxiety vulnerable cohorts.

Behavioral inhibition (BI) is a temperamental tendency to avoid or withdraw from novel social and nonsocial situations, and has been shown to predispose individuals to anxiety disorders. However, adequate means to assess individual differences in avoidance learning in humans are presently limited. Here, we tested whether individuals with high self-reported BI show faster associative learning on a purely cognitive task and whether such inhibited individuals are more prone to avoid aversive outcomes. In Experiment 1, we tested 74 healthy undergraduate students (mean age 19.5 years; 55.4% female) on a computer-based probabilistic classification task, where participants were asked to classify four distinct visual stimuli into two categories. Two stimuli were associated with reward (point gain) and two were associated with punishment (point loss). In Experiment 2, 79 participants from the same population (mean age 19.8 years; 62% female) were tested on a novel modification of the same task, where they also had the option to opt out of responding on each trial, thus avoiding any chance of being punished (or rewarded) on that trial. Results show that inhibited participants demonstrated better associative learning in Experiment 1, while exhibiting a greater tendency to opt out in Experiment 2 (repeated-measures analysis of variance, main effects of BI, both p 0.05). These results indicate that the facilitated classically conditioned learning previously observed in inhibited individuals can be extended to a cognitive task, and also highlight a specific preference in inhibited individuals for withdrawal ("opting out") as a response strategy, when multiple strategies are available to avoid punishment.

One barrier to interpreting past studies of cognition and major depressive disorder (MDD) has been the failure in many studies to adequately dissociate the effects of MDD from the potential cognitive side effects of selective serotonin reuptake inhibitors (SSRIs) use. To better understand how remediation of depressive symptoms affects cognitive function in MDD, we evaluated three groups of subjects: medication-naïve patients with MDD, medicated patients with MDD receiving the SSRI paroxetine, and healthy control (HC) subjects. All were administered a category-learning task that allows for dissociation between learning from positive feedback (reward) vs. learning from negative feedback (punishment). Healthy subjects learned significantly better from positive feedback than medication-naïve and medicated MDD groups, whose learning accuracy did not differ significantly. In contrast, medicated patients with MDD learned significantly less from negative feedback than medication-naïve patients with MDD and healthy subjects, whose learning accuracy was comparable. A comparison of subject's relative sensitivity to positive vs. negative feedback showed that both the medicated MDD and HC groups conform to Kahneman and Tversky's (1979) Prospect Theory, which expects losses (negative feedback) to loom psychologically slightly larger than gains (positive feedback). However, medicated MDD and HC profiles are not similar, which indicates that the state of medicated MDD is not "normal" when compared to HC, but rather balanced with less learning from both positive and negative feedback. On the other hand, medication-naïve patients with MDD violate Prospect Theory by having significantly exaggerated learning from negative feedback. This suggests that SSRI antidepressants impair learning from negative feedback, while having negligible effect on learning from positive feedback. Overall, these findings shed light on the importance of dissociating the cognitive consequences of MDD from those of SSRI treatment, and from cognitive evaluation of MDD subjects in a medication-naïve state before the administration of antidepressants. Future research is needed to correlate the mood-elevating effects and the cognitive balance between reward- and punishment-based learning related to SSRIs.

Unsupervised stimulus-stimulus redundancy compression, one component of Gluck and Myers’s (1993) representational theory of the hippocampal-region function, could emerge from the anatomy and physiology of the entorhinal cortex. This hypothesis is suggested by a physiologically and anatomically realistic model of the entorhinal cortex derived from a similar model of the olfactory cortex previously proposed by Ambros-Ingerson, Granger, and Lynch (1990). To the extent that entorhinal function can survive damage strictly limited to the hippocampal formation (the H lesion), this has implications for interpreting the behavioral consequences of lesions which either do or do not spare overlying cortical areas. In particular, we expect that the H lesion should not interrupt stimulus-stimulus redundancy compression, thereby sparing conditioning behaviors, such as latent inhibition, which are eliminated by broader (H++) lesions to the hippocampal region. However, such other behaviors as the context sensitivity of latent inhibition and of learned responses are expected to be affected by the H lesion. These predictions are consistent with empirical data. The theory also leads to several novel predictions for behavioral comparisons of intact, H-lesioned, and H++-lesioned animals on tasks such as sensory preconditioning, compound preconditioning, and easy-hard transfer. A major theme of this paper is to illustrate how a bottom-up model of cortical processing can be integrated with a top-down model of hippocampal-region function to yield a more complete mapping from physiology to behavior.

A recurrent-network model provides a unified account of the hippocampal region in mediating the representation of temporal information in classical eyeblink conditioning. Much empirical research is consistent with a general conclusion that delay conditioning (in which the conditioned stimulus CS and unconditioned stimulus US overlap and co-terminate) is independent of the hippocampal system, while trace conditioning (in which the CS terminates before US onset) depends on the hippocampus. However, recent studies show that, under some circumstances, delay conditioning can be hippocampal-dependent and trace conditioning can be spared following hippocampal lesion. Here, we present an extension of our prior trial-level models of hippocampal function and stimulus representation that can explain these findings within a unified framework. Specifically, the current model includes adaptive recurrent collateral connections that aid in the representation of intra-trial temporal information. With this model, as in our prior models, we argue that the hippocampus is not specialized for conditioned response timing, but rather is a general-purpose system that learns to predict the next state of all stimuli given the current state of variables encoded by activity in recurrent collaterals. As such, the model correctly predicts that hippocampal involvement in classical conditioning should be critical not only when there is an intervening trace interval, but also when there is a long delay between CS onset and US onset. Our model simulates empirical data from many variants of classical conditioning, including delay and trace paradigms in which the length of the CS, the inter-stimulus interval, or the trace interval is varied. Finally, we discuss model limitations, future directions, and several novel empirical predictions of this temporal processing model of hippocampal function and learning.

To test a prediction of our previous computational model of cortico‐hippocampal interaction (Gluck and Myers [1993, 2001]) for characterizing individual differences in category learning, we studied young healthy subjects using an fMRI‐adapted category‐learning task that has two phases, an initial phase in which associations are learned through trial‐and‐error feedback followed by a generalization phase in which previously learned rules can be applied to novel associations (Myers et al. [2003]). As expected by our model, we found a negative correlation between learning‐related hippocampal responses and accuracy during transfer, demonstrating that hippocampal adaptation during learning is associated with better behavioral scores during transfer generalization. In addition, we found an inverse relationship between Blood Oxygenation Level Dependent (BOLD) activity in the striatum and that in the hippocampal formation and the orbitofrontal cortex during the initial learning phase. Conversely, activity in the dorsolateral prefrontal cortex, orbitofrontal cortex and parietal lobes dominated over that of the hippocampal formation during the generalization phase. These findings provide evidence in support of theories of the neural substrates of category learning which argue that the hippocampal region plays a critical role during learning for appropriately encoding and representing newly learned information so that that this learning can be successfully applied and generalized to subsequent novel task demands. Hum Brain Mapp 35:3122–3131, 2014. © 2013 Wiley Periodicals, Inc.

The circuitry of the dentate gyrus of the hippocampus is unique compared to other hippocampal subfields because there are two glutamatergic principal cells instead of one: granule cells, which are the vast majority of the cells in the dentate gyrus, and the so-called ‘mossy cells.’ The distinctive appearance of mossy cells, the extensive divergence of their axons, and their vulnerability to excitotoxicity relative to granule cells has led to a great deal of interest in mossy cells. Nevertheless, there is no consensus about the normal functions of mossy cells and the implications of their vulnerability. There even seems to be some ambiguity about exactly what mossy cells are. Here we review initial studies of mossy cells, characteristics that define them, and suggest a practical definition to allow investigators to distinguish mossy cells from other hilar neurons even if all morphological and physiological information is unavailable due to technical limitations of their experiments. In addition, hypotheses are discussed about the role of mossy cells in the dentate gyrus network, reasons for their vulnerability and their implications for disease.

OBJECTIVE Drawing on two different populations, Israeli police and Hungarian civilians, the present study assessed the ability of individuals with posttraumatic stress disorder (PTSD) to generalize previous learning to novel situations. Past neuroimaging studies have demonstrated diminished medial temporal lobe (MTL) activation and/or reduced hippocampal volume in individuals with PTSD. Our earlier computational models of cortico-hippocampal function and subsequent experimental tests of these models in MTL-impaired clinical populations argue that even mild hippocampal dysfunction may result in subtle impairments in generalization. Therefore, we predicted that individuals with PTSD would show impaired generalization. METHOD We compared the performance of five groups from two countries, including 19 Israeli police with PTSD and 22 trauma-exposed police without PTSD, and 22 Hungarian civilians with PTSD, 25 trauma-exposed civilians without PTSD, and 25 individuals without PTSD unexposed to the same trauma. Participants were tested on a two-phase learning paradigm, the Acquired Equivalence Task, which measures the ability to generalize past learning to novel situations. RESULTS We found that both PTSD and non-PTSD participants were capable of learning the initial stimulus-outcome associations, F(4, 108) = 1.79, p = .14. However, as predicted, only individuals with PTSD showed a selective deficit in generalization of this learning to novel situations (F(4, 108) = 8.35, p < .001, Partial η2 = 0.26). CONCLUSIONS Individuals with PTSD show a selective impairment in generalization of past learning similar to other clinical populations with MTL/hippocampal dysfunction. This is consistent with an emerging view of PTSD as being not only an anxiety disorder but also a learning disorder.

Damage to the hippocampal and frontostriatal systems can occur across the adult life span. As these 2 systems are involved in learning processes, mild impairments of learning and generalization might be observed even in healthy aging. In this study, we examined both learning and generalization performance in 3 groups of older adults: young-older (ages 45 to 60 y), middle-older (ages 61 to 75 y), and oldest-older (ages 76 to 90 y). We used a simple computerized concurrent discrimination task in which the learning phase has shown sensitivity to frontostriatal dysfunction, and the generalization phase to hippocampal damage. We found that age significantly affected initial learning performance, but generalization was spared in all but the oldest group, with some individuals still generalizing very well. This finding suggests that (a) learning abilities are affected in healthy aging (consistent with earlier reports of frontostriatal dysfunction in healthy aging) and (b) generalization deficit does not necessarily occur in early older age. We hypothesize that generalization deficits in some in the oldest group may be related to hippocampal pathology. Our data shed light on possible neural system dysfunction in healthy aging and Alzheimer disease.

Previous studies have shown that high total homocysteine levels are associated with Alzheimer's disease (AD) and mild cognitive impairment (MCI). In this study, we test the relationship between cognitive function and total homocysteine levels in healthy subjects (Global Dementia Rating, CDR = 0) and individuals with MCI (CDR = 0.5). We have used a cognitive task that tests learning and generalization of rules, processes that have been previously shown to rely on the integrity of the striatal and hippocampal regions, respectively. We found that total homocysteine levels are higher in MCI individuals than in healthy controls. Unlike what we expected, we found no difference between MCI subjects and healthy controls in learning and generalization. We conducted further analysis after diving MCI subjects in two groups, depending on their Global Deterioration Scale (GDS) scores: individuals with very mild cognitive decline (vMCD, GDS = 2) and mild cognitive decline (MCD, GDS = 3). There was no difference among the two MCI and healthy control groups in learning performance. However, we found that individuals with MCD make more generalization errors than healthy controls and individuals with vMCD. We found no difference in the number of generalization errors between healthy controls and MCI individuals with vMCD. In addition, interestingly, we found that total homocysteine levels correlate positively with generalization errors, but not with learning errors. Our results are in agreement with prior results showing a link between hippocampal function, generalization performance, and total homocysteine levels. Importantly, our study is perhaps among the first to test the relationship between learning (and generalization) of rules and homocysteine levels in healthy controls and individuals with MCI.

In classical conditioning, proactive interference may arise from experience with the conditioned stimulus (CS), the unconditional stimulus (US), or both, prior to their paired presentations. Interest in the application of proactive interference has extended to clinical populations as either a risk factor for disorders or as a secondary sign. Although the current literature is dense with comparisons of stimulus pre-exposure effects in animals, such comparisons are lacking in human subjects. As such, interpretation of proactive interference over studies as well as its generalization and utility in clinical research is limited. The present study was designed to assess eyeblink response acquisition after equal numbers of CS, US, and explicitly unpaired CS and US pre-exposures, as well as to evaluate how anxiety vulnerability might modulate proactive interference. In the current study, anxiety vulnerability was assessed using the State/Trait Anxiety Inventories as well as the adult and retrospective measures of behavioral inhibition (AMBI and RMBI, respectively). Participants were exposed to 1 of 4 possible pre-exposure contingencies: 30 CS, 30 US, 30 CS, and 30 US explicitly unpaired pre-exposures, or Context pre-exposure, immediately prior to standard delay training. Robust proactive interference was evident in all pre-exposure groups relative to Context pre-exposure, independent of anxiety classification, with CR acquisition attenuated at similar rates. In addition, trait anxious individuals were found to have enhanced overall acquisition as well as greater proactive interference relative to non-vulnerable individuals. The findings suggest that anxiety vulnerable individuals learn implicit associations faster, an effect which persists after the introduction of new stimulus contingencies. This effect is not due to enhanced sensitivity to the US. Such differences would have implications for the development of anxiety psychopathology within a learning framework.

Behavioral inhibition (BI), a temperamental bias to respond to novel stimuli with avoidance behaviors, is a risk factor for posttraumatic stress disorder (PTSD). It is unclear whether BI accounts for additional variance in PTSD symptom severity beyond that accounted for by general anxiety. Here, 109 veterans (mean age 50.4 years, 9.2% female) provided self-assessment of PTSD symptoms, state and trait anxiety, combat exposure, and current (adult) and retrospective (childhood) BI. Adult BI was correlated with anxiety and PTSD symptom severity, especially cluster C (avoidance) symptoms, but not with combat exposure. A regression model including adult BI, state and trait anxiety, and combat exposure was able to correctly classify over 80% of participants according to presence or absence of severe PTSD symptoms. Because avoidance behaviors are a core component of PTSD, self-assessments of BI may be an important tool in understanding PTSD and potentially assessing vulnerability to the disorder.

α-Synuclein (SNCA) plays an important role in the regulation of dopaminergic neurotransmission and neurodegeneration in Parkinson disease. We investigated reward and punishment learning in asymptomatic carriers of a rare SNCA gene duplication who were healthy siblings of patients with Parkinson disease. Results revealed that healthy SNCA duplication carriers displayed impaired reward and intact punishment learning compared with noncarriers. These results demonstrate that a copy number variation of the SNCA gene is associated with selective impairments on reinforcement learning in asymptomatic carriers without the motor symptoms of Parkinson disease.

Building on our previous neurocomputational models of basal ganglia and hippocampal region function (and their modulation by dopamine and acetylcholine, respectively), we show here how an integration of these models can inform our understanding of the interaction between the basal ganglia and hippocampal region in associative learning and transfer generalization across various patient populations. As a common test bed for exploring interactions between these brain regions and neuromodulators, we focus on the acquired equivalence task, an associative learning paradigm in which stimuli that have been associated with the same outcome acquire a functional similarity such that subsequent generalization between these stimuli increases. This task has been used to test cognitive dysfunction in various patient populations with damages to the hippocampal region and basal ganglia, including studies of patients with Parkinson’s disease (PD), schizophrenia, basal forebrain amnesia, and hippocampal atrophy. Simulation results show that damage to the hippocampal region—as in patients with hippocampal atrophy (HA), hypoxia, mild Alzheimer’s (AD), or schizophrenia—leads to intact associative learning but impaired transfer generalization performance. Moreover, the model demonstrates how PD and anterior communicating artery (ACoA) aneurysm—two very different brain disorders that affect different neural mechanisms—can have similar effects on acquired equivalence performance. In particular, the model shows that simulating a loss of dopamine function in the basal ganglia module (as in PD) leads to slow acquisition learning but intact transfer generalization. Similarly, the model shows that simulating the loss of acetylcholine in the hippocampal region (as in ACoA aneurysm) also results in slower acquisition learning. We argue from this that changes in associative learning of stimulus–action pathways (in the basal ganglia) or changes in the learning of stimulus representations (in the hippocampal region) can have similar functional effects.

In a study of acquired equivalence in Parkinson disease (PD), in which patients were tested on normal dopaminergic medication, we found that comorbid clinical depression impairs initial acquisition, whereas the use of anticholinergic therapy impairs subsequent transfer generalization. In addition, this study provides a replication of the basic finding of Myers et al (2003) that patients with PD on dopaminergic therapy are impaired at initial acquisition, but normal at subsequent transfer generalization, generalizing these results to an Arabic-speaking population including many participants with no formal education. These results are consistent with our past computational modeling, which argues that acquisition of incrementally acquired, feedback-based learning tasks is dependent on cortico-striatal circuits, whereas transfer generalization is dependent on medial temporal (MT) structures. They are also consistent with prior computational modeling, and with empiric work in humans and animals, suggesting that anticholinergic drugs may particularly impair cognitive abilities that depend on the MT lobe.

Recent human fMRI studies have shown that the hippocampal region is essential for probabilistic category learning, memory formation-retrieval and context based performance. We present an artificial neural network model that can qualitatively simulate the BOLD signal for these tasks. The model offers ideas on the functional architecture and the relationship between the hippocampus and other brain structures. We also show that symptoms of neurobiological diseases like Parkinson's disease (PD) and Schizophrenia can be simulated and studied using the model.

Making appropriate choices often requires the ability to learn the value of available options from experience. Parkinson's disease is characterized by a loss of dopamine neurons in the substantia nigra, neurons hypothesized to play a role in reinforcement learning. Although previous studies have shown that Parkinson's patients are impaired in tasks involving learning from feedback, they have not directly tested the widely held hypothesis that dopamine neuron activity specifically encodes the reward prediction error signal used in reinforcement learning models. To test a key prediction of this hypothesis, we fit choice behavior from a dynamic foraging task with reinforcement learning models and show that treatment with dopaminergic drugs alters choice behavior in a manner consistent with the theory. More specifically, we found that dopaminergic drugs selectively modulate learning from positive outcomes. We observed no effect of dopaminergic drugs on learning from negative outcomes. We also found a novel dopamine-dependent effect on decision making that is not accounted for by reinforcement learning models: perseveration in choice, independent of reward history, increases with Parkinson's disease and decreases with dopamine therapy.

Computational modeling predicts that the hippocampus plays an important role in the ability to apply previously learned information to novel problems and situations (referred to as the ability to generalize information or simply as ‘transfer learning’). These predictions have been tested in humans using a computer-based task on which individuals with hippocampal damage are able to learn a series of complex discriminations with two stimulus features (shape and color), but are impaired in their ability to transfer this information to newly configured problems in which one of the features is altered. This deficit occurs despite the fact that the feature predictive of the reward (the relevant information) is not changed. The goal of the current study was to develop a mouse analog of transfer learning and to determine if this new task was sensitive to pathological changes in a mouse model of AD. We describe a task in which mice were able to learn a series of concurrent discriminations that contained two stimulus features (odor and digging media) and could transfer this learned information to new problems in which the irrelevant feature in each discrimination pair was altered. Moreover, we report age-dependent deficits specific to transfer learning in APP + PS1 mice relative to non-transgenic littermates. The robust impairment in transfer learning may be more sensitive to AD-like pathology than traditional cognitive assessments in that no deficits were observed in the APP + PS1 mice on the widely used Morris water maze task. These data describe a novel and sensitive paradigm to evaluate mnemonic decline in AD mouse models that has unique translational advantages over standard species-specific cognitive assessments (e.g., water maze for rodent and delayed paragraph recall for humans).

Computational modeling predicts that the hippocampus plays an important role in the ability to apply previously learned information to novel problems and situations (referred to as the ability to generalize information or simply as ‘transfer learning’). These predictions have been tested in humans using a computer-based task on which individuals with hippocampal damage are able to learn a series of complex discriminations with two stimulus features (shape and color), but are impaired in their ability to transfer this information to newly configured problems in which one of the features is altered. This deficit occurs despite the fact that the feature predictive of the reward (the relevant information) is not changed. The goal of the current study was to develop a mouse analog of transfer learning and to determine if this new task was sensitive to pathological changes in a mouse model of AD. We describe a task in which mice were able to learn a series of concurrent discriminations that contained two stimulus features (odor and digging media) and could transfer this learned information to new problems in which the irrelevant feature in each discrimination pair was altered. Moreover, we report age-dependent deficits specific to transfer learning in APP + PS1 mice relative to non-transgenic littermates. The robust impairment in transfer learning may be more sensitive to AD-like pathology than traditional cognitive assessments in that no deficits were observed in the APP + PS1 mice on the widely used Morris water maze task. These data describe a novel and sensitive paradigm to evaluate mnemonic decline in AD mouse models that has unique translational advantages over standard species-specific cognitive assessments (e.g., water maze for rodent and delayed paragraph recall for humans).

Parkinson's disease is characterized by the degeneration of dopaminergic pathways projecting to the striatum. These pathways are implicated in reward prediction. In this study, we investigated reward and punishment processing in young, never-medicated Parkinson's disease patients, recently medicated patients receiving the dopamine receptor agonists pramipexole and ropinirole and healthy controls. The never-medicated patients were also re-evaluated after 12 weeks of treatment with dopamine agonists. Reward and punishment processing was assessed by a feedback-based probabilistic classification task. Personality characteristics were measured by the temperament and character inventory. Results revealed that never-medicated patients with Parkinson's disease showed selective deficits on reward processing and novelty seeking, which were remediated by dopamine agonists. These medications disrupted punishment processing. In addition, dopamine agonists increased the correlation between reward processing and novelty seeking, whereas these drugs decreased the correlation between punishment processing and harm avoidance. Our finding that dopamine agonist administration in young patients with Parkinson's disease resulted in increased novelty seeking, enhanced reward processing, and decreased punishment processing may shed light on the cognitive and personality bases of the impulse control disorders, which arise as side-effects of dopamine agonist therapy in some Parkinson's disease patients.

Learning with delayed reinforcement refers to situations where the reinforcement to a learning system occurs only at the end of a string of actions or outputs, and it must then be assigned back to the relevant actions. A method for accomplishing this is presented which buffers a small number of past actions based on the unpredictability of or attention to each as it occurs. This approach allows for the buffer size to be small, and yet learning can reach indefinitely far back into the past; it also allows the system to learn when reinforcement is not only delayed but also reinforcements from other unrelated actions may arrive during this delay. An example of a simulated food-finding creature is used to show the system at work in a predictive application where reinforcements show this interleaving behaviour.

Patients with depression show cognitive impairment, including executive function deficit, impairments in attention, declarative memory and psychomotor performance. In addition to classic, widely studied cognitive functions, in depression implicit learning and the interpretation of feedback and its impact on performance can also be impaired compared to healthy individuals. While cognitive functions have been widely studied, much less is known about implicit learning in depression.The two-phased Kilroy sequence association test examines the basal ganglia-mediated and the temporal lobe and hippocampus-mediated learning processes within one test. We compared the performance of 22 depressed patients (according to DSM-IV) and 20 healthy control subjects using the Kilroy test. In the depressed group, we also compared the performance on each step of the test with the symptom severity measured by the Hamilton D symptom scale.Depressed patients showed impaired performance compared to healthy subjects on the first, learning phase of the test. The degree of deficit on the learning phase correlated with symptom severity. We found no difference between the two groups on the second, context-dependent phase of the test.Our results confirm the presence of a striatal deficit in depressed patients. Results indicate that parallel memory systems are not equally affected in depression, and the character of deficit in depression may be specific to the illness.

In probabilistic categorization tasks, various cues are probabilistically (but not perfectly) predictive of class membership. This means that a given combination of cues sometimes belongs to one class and sometimes to another. It is not yet clear how categorizers approach such tasks. Here, we review evidence in favor of two alternative conceptualizations of learning in probabilistic categorization: as rule-based learning, or as incremental learning. Each conceptualization forms the basis of a way of analyzing performance: strategy analysis assumes rule-based learning, while rolling regression analysis assumes incremental learning. Here, we contrasted the ability of each to predict performance of normal categorizers. Both turned out to predict responses about equally well. We then reviewed performance of patients with damage to regions deemed important for either rule-based or incremental learning. Evidence was again about equally compatible with either alternative conceptualization of learning, although neither predicted an involvement of the medial temporal lobe. We suggest that a new way of conceptualizing probabilistic categorization might be fruitful, in which the medial temporal lobe help set up representations that are then used by other regions to assign patterns to categories. © 2007 Elsevier Ltd. All rights reserved.

Objective The purpose of this study was to examine stimulus–response (S–R) learning in active cocaine users. Participants and methods Twenty-two cocaine-dependent participants (20 males and 2 females) and 21 non-drug using control participants (19 males and 2 females) who were similar in age and education were administered two computerized learning tasks. The Acquired Equivalence task initially requires learning of simple antecedent–consequent discriminations, but later requires generalization of this learning when the stimuli are presented in novel recombinations. The Weather Prediction task requires the prediction of a dichotomous outcome based on different stimuli combinations when the stimuli predict the outcome only probabilistically. Results On the Acquired Equivalence task, cocaine users made significantly more errors than control participants when required to learn new discriminations while maintaining previously learned discriminations, but performed similarly to controls when required to generalize this learning. No group differences were seen on the Weather Prediction task. Conclusions Cocaine users’ learning of stimulus discriminations under conflicting response demands was impaired, but their ability to generalize this learning once they achieved criterion was intact. This performance pattern is consistent with other laboratory studies of long-term cocaine users that demonstrated that established learning interfered with new learning on incremental learning tasks, relative to healthy controls, and may reflect altered dopamine transmission in the basal ganglia of long-term cocaine users.

Alpha-synuclein (SNCA) is a key factor in the regulation of dopaminergic transmission and is related to Parkinson's disease. In this study, we investigated the effects of risk and protective SNCA haplotypes associated with Parkinson's disease on cognitive sequence learning in 204 healthy volunteers. We found that the 3'-block risk SNCA haplotypes are associated with less effective stimulus-reward learning of sequences and with superior context representation of sequences. In contrast, participants with protective haplotypes exhibit better stimulus-reward learning and worse context representation, which suggest that these functions are inversely affected by risk and protective haplotypes. The Rep1 promoter polymorphism does not influence cognitive sequence learning. Because stimulus-reward learning may be mediated by the basal ganglia and context learning may be related to the medial temporal lobe, our data raise the possibility that dopaminergic signals regulated by SNCA inversely affect these memory systems.

Several lines of evidence implicate the prefrontal cortex in learning but there is little evidence from studies of human lesion patients to demonstrate the critical role of this structure. To this end, we tested patients with lesions of the frontal lobe (n=36) and healthy controls (n=35) on two learning tasks: the weather prediction task (WPT), and an eight-pair concurrent visual discrimination task ('Choose'). Performance of both tasks was previously shown to be disrupted in patients with Parkinson's disease; the Choose deficit was only present when patients were medicated. Patients with damage to the orbitofrontal cortex (OFC) were significantly impaired on Choose, compared to both healthy controls and non-OFC lesion patients. The OFC lesion patients showed a mild deficit on the first 50 trials of the WPT, compared to the control subjects but not non-OFC lesion patients. The selective deficit in the OFC patients on Choose performance could not be attributed to the larger lesion size in this group, and the deficit was not correlated with the volume of damage to adjacent prefrontal subregions (e.g. anterior cingulate cortex). These data support the notion that the OFC play a role in normal discrimination learning, and suggest qualitative similarities in learning performance of patients with OFC damage and medicated PD patients.

Introduction and Objectives Previous studies have shown that high total homocysteine levels are associated with Alzheimer’s disease (AD) and mild cognitive impairment (MCI). Aims In this study, we test the relationship between cognitive function and total homocysteine levels in healthy subjects (Global Dementia Rating, CDR = 0) and individuals with MCI (CDR = 0.5). Methods We have used a cognitive task that tests learning and generalization of rules, processes that have been previously shown to rely on the integrity of the striatal and hippocampal regions, respectively. Results We found that total homocysteine levels are higher in MCI individuals than in healthy controls. We conducted further analysis after diving MCI subjects in two groups, depending on their Global Deterioration Scale (GDS) scores: individuals with very mild cognitive decline (vMCD) and mild cognitive decline. There was no difference among the two MCI and healthy control groups in learning performance. However, we found that individuals with MCD make more generalization errors than healthy controls and individuals with vMCD. In addition, interestingly, we found that total homocysteine levels correlate positively with generalization errors, but not with learning errors. Conclusions Our results are in agreement with prior results showing a link between hippocampal function, generalization performance, and total homocysteine levels. Importantly, our study is perhaps among the first to test the relationship between learning (and generalization) of rules and homocysteine levels in healthy controls and individuals with MCI.

Evidence suggests that dopaminergic mechanisms in the basal ganglia are important in feedback-guided habit learning. To test hypothesis, we assessed cognitive sequence learning in 120 healthy volunteers and measured plasma levels of homovanillic acid [HVA] (a metabolite of dopamine), 5-hydroxyindoleacetic acid [5-HIAA] (a metabolite of serotonin), and 3-methoxy-4-hydroxypheylglycol [MHPG] (a metabolite of norepinephrine). Results revealed a significant negative relationship between errors in the feedback-guided training phase of the sequence learning task and the plasma HVA level. The HVA level accounted for 10.5% of variance of performance. Participant who had lower HVA level than the median value of the whole sample committed more errors during the training phase compared with participants who had higher HVA plasma level than the median value. A similar phenomenon was not observed for the context-dependent phase of the task and for 5-HIAA and MHPG. These results suggest that dopamine plays a special role in feedback-guided cognitive sequence learning.

Evidence suggests that dopaminergic mechanisms in the basal ganglia (BG) are important in the learning of sequential associations. To test the specificity of this hypothesis, we assessed never-medicated patients with Parkinson's disease (PD) and amnestic mild cognitive impairment (aMCI) using a chaining task. In the training phase of the chaining task, each link in a sequence of stimuli leading to reward is trained step-by-step using feedback after each decision, until the complete sequence is learned. In the probe phase of the chaining task, the context of stimulus-response associations must be used (the position of the associations in the sequence). Results revealed that patients with PD showed impaired learning during the training phase of the chaining task, but their performance was spared in the probe phase. In contrast, patients with aMCI with prominent medial temporal lobe (MTL) dysfunctions showed intact learning during the training phase of the chaining task, but their performance was impaired in the probe phase of the chaining task. These results indicate that when dopaminergic mechanisms in the BG are dysfunctional, series of stimulus-response associations are less efficiently acquired, but their sequential manner is maintained. In contrast, MTL dysfunctions may result in a non-sequential learning of associations, which may indicate a loss of contextual information.

The purpose of the present study was to investigate basal ganglia (BG) and medial temporal lobe (MTL) dependent learning in patients with schizophrenia. Acquired equivalence is a phenomenon in which prior training to treat two stimuli as equivalent (if two stimuli are associated with the same response) increases generalization between them. The learning of stimulus-response pairs is related to the BG, whereas the MTL system participates in stimulus generalization. Forty-three patients with DSM-IV schizophrenia and 28 matched healthy controls participated in the study. Volunteers received the acquired equivalence associative learning task (FISHES) during which associations between faces and pet fishes are learned. Results revealed that patients with schizophrenia showed a selective deficit on stimulus generalization, whereas stimulus-response learning was unaffected. The number of errors during stimulus-response learning correlated with the daily chlorpromazine-equivalent dose of antipsychotics. In conclusion, patients with schizophrenia showed deficits during MTL-dependent learning, but not during BG-dependent learning. High-dose first generation antipsychotics may disrupt BG-dependent learning by blocking dopaminergic neurotransmission in the nigro-striatal system.

Human anterograde amnesia can develop following bilateral damage to the hippocampus and medial temporal lobes, as in hypoxic brain injury, or following damage to the basal forebrain, as following anterior communicating artery (ACoA) aneurysm rupture. In both cases, the mnestic deficit may be similar when assessed by standard neuropsychological measures. However, animal and computational models suggest that there are qualitative differences in the pattern of impaired and spared memory abilities following damage to hippocampus versus basal forebrain. Here, we show such a dissociation in human amnesia using a single two-stage task, involving conditional discrimination and reversal. Consistent with a prior study, 10 individuals with anterograde amnesia subsequent to hypoxic brain injury were spared on acquisition but impaired at reversal. However, 10 individuals with amnesia subsequent to ACoA aneurysm showed the opposite pattern of impaired acquisition but spared reversal. The differences between groups cannot be easily ascribed to severity of mnestic or cognitive deficit, since the two amnesic groups performed similarly on neuropsychological tests of memory, intelligence and attention. The results illustrate qualitative differences in memory impairments in hypoxic and ACoA amnesics and highlight the importance of considering etiology in evaluating mnemonic deficits in amnesic populations.

We have described our efforts to further the understanding of the neural substrates of rabbit eyeblink conditioning (in particular the cerebellum and hippocampus) through the use of neural networks. Over the years, the initial theoretical assumptions of the computational models have been maintained while more anatomically and physiologically detailed instantiations of these models have been developed. These neural networks have increased our understanding of how classical conditioning occurs in the cerebellum and hippocampus while making novel predictions which are guiding ongoing and future empirical studies. Neural networks result in simulations of various lesion conditions which offer subtle distinctions which may not be evident in a purely qualitative theory.

In this study we examined the effect of dopaminergic modulation on learning and memory. Parkinson's patients were tested 'on' versus 'off' dopaminergic medication, using a two-phase learning and transfer task. We found that dopaminergic medication was associated with impaired learning of an incrementally acquired concurrent discrimination task, while patients withdrawn from dopaminergic medication performed as well as controls. In addition, we found a dissociation of the effect of medication within a single two-phase task: patients tested 'on' medication were not impaired at the ability to generalize based on learned information. The deficit among medicated patients appeared to be related specifically to the concurrent, incremental, feedback-based nature of the task: such a deficit was not found in a version of the task in which demands for concurrent error-processing learning were reduced. Taken together with a growing body of evidence emphasizing a role for midbrain dopamine in error-correcting, feedback-based learning processes, the present results suggest a framework for understanding previously conflicting results regarding the effect of medication on learning and memory in Parkinson's disease.

Previous research has suggested that a probabilistic category learning task (e.g. weather prediction task) can be used to elucidate brain substrates of learning. We tested amnesic subjects with bilateral hippocampal damage due to hypoxia and matched controls on the weather prediction task and a variant, the "ice cream" task, which maintains a similar category structure. The hypoxic subjects were impaired relative to controls on both tasks; in the ice cream task, this difference was evident even early in training (first 50 trials). This finding is similar to functional neuroimaging (fMRI) studies in healthy subjects, which show medial temporal involvement even in early learning on this task. Additionally, strategy analysis of response patterns during learning suggest that the hypoxic group relied more heavily on simple, degraded learning strategies than did the control group. These results may suggest a qualification of the generally held conclusion that amnesic patients are not impaired at probabilistic category learning: at least under some circumstances, amnesic patients show an early and lasting deficit.

The medial temporal (MT) lobes and basal ganglia have both been implicated as brain substrates of associative learning. Here, we show a dissociation between medial temporal and basal ganglia damage using a latent learning task, in which prior exposure to cues, uncorrelated with each other, slows subsequent learning of an association between them. Consistent with prior work, we found a robust exposure effect in healthy controls, with exposed controls learning more slowly than non-exposed controls. This effect was abolished in medial temporal amnesia: both exposed and non-exposed amnesic patients learned at the same speed. A group of patients with basal ganglia damage due to Parkinson’s disease showed a reversal of the effect: exposed subjects learned faster than non-exposed subjects. Our findings point to distinct and dissociable contributions of medial temporal lobe and basal ganglia structures to learning and memory.

Cortical representations of sound can be modified by repeatedly pairing presentation of a pure tone with electrical stimulation of neuromodulatory neurons located in the basal forebrain (Bakin & Weinberger, 1996; Kilgard & Merzenich, 1998a). We developed a computational model to investigate the possible effects of basal forebrain modulation on map reorganization in the auditory cortex. The model is a self-organizing map with acoustic response characteristics mimicking those observed in the mammalian auditory cortex. We simulated the effects of basal forebrain modulation, using parameters intrinsic to the self-organizing map, such as the learning rate (controlling the adaptability of map nodes) and the neighborhood function (controlling the excitability of map nodes). Previous research has suggested that both parameters can be useful for characterizing the effects of neuromodulation on plasticity (Kohonen, 1993; Myers et al., 1996; Myers, Ermita, Hasselmo, & Gluck, 1998). The model successfully accounts for experimentally observed effects of pairing basal forebrain stimulation with the presentation of a single tone, but not of two tones, suggesting that auditory cortical plasticity is constrained in ways not accounted for by current theories. Despite this limitation, the model provides a useful framework for describing experience-induced changes in auditory representations and for relating such changes to variations in the excitability and adaptability of cortical neurons produced by neuromodulation.