A neural model of hippocampal–striatal interactions in associative learning and transfer generalization in various neurological and psychiatric patients

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.     

Neurogenesis in adulthood: a possible role in learning

The role of the hippocampal formation in learning and memory has long been recognized. However, despite decades of intensive research, the neurobiological basis of this process in the hippocampus remains enigmatic. Over 30 years ago, the production of new neurons was found to occur in the brains of adult rodents. More recently, the documentation of adult neurogenesis in the hippocampal formation of a variety of mammals, including humans, has suggested a novel approach towards understanding the biological bases of hippocampal function. Contemporary theories of hippocampal function include an important role for this brain region in associative learning. The addition of new neurons and consequently, their novel contribution to hippocampal circuitry could conceivably be a mechanism for relating spatially or temporally disparate events. In this review, we examine several lines of evidence suggesting that adult-generated neurons are involved in hippocampal-dependent learning. In particular, we examine the variables that modulate hippocampal neurogenesis in adulthood and their relation to learning and memory.     

NMDA and muscarinic receptors of the nucleus accumbens have differential effects on taste memory formation

Animals recognize a taste cue as aversive when it has been associated with post-ingestive malaise; this associative learning is known as conditioned taste aversion (CTA). When an animal consumes a new taste and no negative consequences follow, it becomes recognized as a safe signal, leading to an increase in its consumption in subsequent presentations (attenuation of neophobia, AN). It has been shown that the nucleus accumbens (NAcc) has an important role in taste learning. To elucidate the involvement of N-methyl-D-aspartate (NMDA) and muscarinic receptors in the NAcc during safe and aversive taste memory formation, we administrated bilateral infusions of DL-2-amino-5-phosphonopentanoic acid (APV) or scopolamine in the NAcc shell or core respectively. Our results showed that pre-training injections of APV in the NAcc core and shell disrupted aversive but not safe taste memory formation, whereas pre-training injections of scopolamine in the NAcc shell, but not core, disrupted both CTA and AN. These results suggest that muscarinic receptors seem to be necessary for processing taste stimuli for either safe or aversive taste memory, whereas NMDA receptors are only involved in the aversive taste memory trace formation.     

Simple and associative recognition memory in the hippocampal region

Although it is well established that the hippocampal region is involved in the formation of declarative memory, the exact nature of its involvement is unclear. One view is that the hippocampal region is involved only in tasks that require the formation or use of associations. According to this view, the hippocampal region is not involved in traditional tests of recognition memory. An alternative view is that the hippocampal region combines and extends the processing carried out by structures in the parahippocampal gyrus and that it is involved in all forms of declarative memory, including recognition memory. Using event-related functional magnetic resonance imaging (fMRI), we observed hippocampal activity during both traditional and associative recognition memory tasks. Critically, the hippocampal region was no more active in the associative recognition task than in the traditional recognition task.     

Hippocampal inactivation enhances taste learning

Learning tasks are typically thought to be either hippocampal-dependent (impaired by hippocampal lesions) or hippocampal-independent (indifferent to hippocampal lesions). Here, we show that conditioned taste aversion (CTA) learning fits into neither of these categories. Rats were trained to avoid two taste stimuli, one novel and one familiar. Muscimol infused through surgically implanted intracranial cannulae temporarily inactivated the dorsal hippocampus during familiarization, subsequent CTA training, or both. As shown previously, hippocampal inactivation during familiarization enhanced the effect of that familiarization on learning (i.e., hippocampal inactivation enhanced latent inhibition of CTA); more novel and surprising, however, was the finding that hippocampal inactivation during training sessions strongly enhanced CTA learning itself. These phenomena were not caused by specific aspects of our infusion technique--muscimol infusions into the hippocampus during familiarization sessions did not cause CTAs, muscimol infusions into gustatory cortex caused the expected attenuation of CTA, and hippocampal inactivation caused the expected attenuation of spatial learning. Thus, we suggest that hippocampal memory processes interfere with the specific learning mechanisms underlying CTA, and more generally that multiple memory systems do not operate independently.     

Not all risk taking behavior is bad: Associative sensitivity predicts learning during risk taking among high sensation seekers

Risk taking behavior can be both adaptive and maladaptive depending on context. The majority of studies on risk taking, however, focus on clinical populations and dangerous or harmful risk taking. Individual differences in learning during risk taking are rarely examined in relation to task performance. The present study examined risk taking and associated outcomes in an exploration-based instrumental learning task (Balloon Emotional Learning Task; BELT), which presented a series of balloons in which participants pump up for points. Consistent with prior work, sensation seeking predicted increased risk taking behavior. Importantly, however, a significant interaction between sensation seeking and associative sensitivity, an attentional construct defined as the frequency and remoteness of automatic cognitive activity, was found. Specifically, among individuals high in sensation seeking, associative sensitivity predicted fewer balloon explosions and an increase in points earned on the balloon condition with the most potential for feedback driven learning. Thus, these findings suggest that sensation seekers are a heterogeneous group, and secondary traits such as associative sensitivity moderate risk taking and learning according to context.     

Trace eyeblink conditioning requires the hippocampus but not autophosphorylation of alphaCaMKII in mice

Little is known about signaling mechanisms underlying temporal associative learning. Here, we show that mice with a targeted point mutation that prevents autophosphorylation of alphaCaMKII (alphaCaMKII(T286A)) learn trace eyeblink conditioning normally. This forms a sharp contrast to the severely impaired spatial learning in the water maze and contextual fear conditioning observed in alphaCaMKII(T286A) mutants. Importantly, hippocampal lesions impaired trace eyeblink conditioning in alphaCaMKII(T286A) mice, suggesting a potential role of hippocampal alphaCaMKII-independent mechanisms. These results indicate that hippocampal signaling mechanisms that underlie temporal associative learning as assessed by trace eyeblink conditioning may differ from those of spatial and contextual learning.     

Structural-functional correlations between hippocampal volume and cortico-limbic emotional responses in depressed children

Although hippocampal atrophy and altered functional brain responses to emotional stimuli have been found in major depressive disorder (MDD), the relationship between the two is not yet well understood. The present study focused on children with and without a history of preschool onset MDD (PO-MDD) and directly examined the relations between hippocampal volume and functional brain activation to affect-eliciting stimuli. Children completed annual diagnostic assessments starting at preschool. When children were school-aged, high-resolution structural MRI and task-related functional MRI data were acquired from N = 64 nonmedicated children. During fMRI, subjects were shown emotional faces. Results from the total sample indicated that smaller bilateral hippocampal volumes were associated with greater cortico-limbic (e.g., amygdala, hippocampus, dorsolateral prefrontal cortex) activation to sad or negative faces versus neutral faces. Left hippocampal volume was negatively associated with the cortico-limbic activation in both the PO-MDD and healthy children. Right hippocampal volume was negatively correlated with amygdala responses in the PO-MDD group, but not in the healthy comparison group. These findings suggest that there may be important interrelationships between reduced hippocampal volume and hyperactivation of brain responses in children, both those with and those without a history of PO-MDD.     

Neurotransmitter release during delay eyeblink classical conditioning: role of norepinephrine in consolidation and effect of age

Delay classical eyeblink conditioning (EBC) is an important model of associative, cerebellar-dependent learning. Norepinephrine (NE) plays a significant modulatory role in the acquisition of learning; however, other neurotransmitters are also involved. The goal was to determine whether NE, gamma-aminobutyric acid (GABA) and glutamate (GLU) release are observed in cerebellar cortex during EBC, and whether such release was selectively associated with training. Further studies examined the role of the beta-noradrenergic receptor in consolidation of the learned response by local infusion of propranolol at 5-120 min following training into the cerebellar cortex. In vivo microdialysis coupled to EBC was performed to examine neurotransmitter release. An increase in the extracellular level of NE was observed during EBC and was maximal on day 1 and diminished in amplitude with subsequent days of training. No changes in baseline NE release were observed in pseudoconditioning indicating that NE release is directly related to the associative learning process. The extracellular levels of GABA were also increased selectively during paired training however, the magnitude of GABA release increased over days of training. GLU release was observed to increase during both paired and unpaired training, suggesting that learning does not occur prior to the information arriving in the cerebellum. When propranolol was administered at either 5-, 60-, or 120-min post-training, there was an inhibition of conditioned responses, these data support the hypothesis that NE is important for consolidation of learning. In another set of experiments we demonstrate that the timing of release of NE, GABA and glutamate are significantly delayed in onset and lengthened in duration in the 22-month-old F344 rats. Over days of training the timing of release becomes closer to the timing of training and this is associated with increased learning of conditioned responses in the aged rats.     

A role for ERK MAP kinase in physiologic temporal integration in hippocampal area CA1

Recent studies demonstrate a requirement for the Extracellular signal Regulated Kinase (ERK) mitogen-activated protein kinase (MAPK) cascade in both the induction of long-lasting forms of hippocampal synaptic plasticity and in hippocampus-dependent associative and spatial learning. In the present studies, we investigated mechanisms by which ERK might contribute to synaptic plasticity at Schaffer collateral synapses in hippocampal slices. We found that long-term potentiation (LTP) induced with a pair of 100-Hz tetani does not require ERK activation in mice whereas it does in rats. However, in mice, inhibition of ERK activation blocked LTP induced by two LTP induction paradigms that mimicked the endogenous θ rhythm. In an additional series of studies, we found that mice specifically deficient in the ERK1 isoform of MAPK showed no impairments in tests of hippocampal physiology. To investigate ERK-dependent mechanisms operating during LTP-inducing stimulation paradigms, we monitored spike production in the cell body layer of the hippocampus during the period of θ-like LTP-inducing stimulation. θ-burst stimulation (TBS) produced a significant amount of postsynaptic spiking, and the likelihood of spike production increased progressively over the course of the three trains of TBS independent of any apparent increase in Excitatory Post-Synaptic Potential (EPSP) magnitude. Inhibition of ERK activation dampened this TBS-associated increase in spiking. These data indicate that, for specific patterns of stimulation, ERK may function in the regulation of neuronal excitability in hippocampal area CA1. Overall, our data indicate that the progressive increase in spiking observed during TBS represents a form of physiologic temporal integration that is dependent on ERK MAPK activity.     

Ovariectomy increases the age-induced hyperphosphorylation of Tau at hippocampal CA1

One of the main hallmarks of Alzheimer’s disease includes the neurofibrillary tangles formation produced by hyperphosphorylation of the Tau protein, whose expression is putatively regulated by the ovarian hormones estradiol and progesterone. Hippocampus is a brain region that participates in many functions related to learning and memory; in addition, it is abundant in both estradiol and progesterone receptors. In this study, we explore the expression of Tau hyperphosphorylation at hippocampus and the performance of rats in an autoshaping learning task at 5, 10 and 15 months after the ovaries removal. In these animals, ovariectomy was performed at 3 months of age. These data were compared with those derived from intact rats at 8, 13 and 18 months old. A clear decrease in the number of conditioned responses of both intact and ovariectomized rats in the autoshaping learning task was observed. The interaction of both factors confirms that, in this test, learning varies depending on aging and the presence or absence of ovaries. A progressive increase in hippocampal Tau phosphorylation at Ser-396 was observed in either intact or ovariectomized rats. Interestingly, an interaction between the analyzed factors shows that such hyperphosphorylation was potentiated by the absence of ovaries. These results emphasize the importance of aging and the lack of ovarian hormones for an associative learning test and for the expression of one of the most important hallmarks of Alzheimer’s disease.     

Effects of concurrent load on feature- and rule-based generalization in human contingency learning

The effect of concurrent load on generalization performance in human contingency learning was examined in 2 experiments that employed the combined positive and negative patterning procedure of Shanks and Darby (1998). In Experiment 1, we tested 32 undergraduates and found that participants who were trained and tested under full attention showed generalization consistent with the application of an opposites rule (i.e., single cues signal the opposite outcome to their compound), whereas participants trained and tested under a concurrent cognitive load showed generalization consistent with surface similarity. In Experiment 2, we replicated the effect with 148 undergraduates and provided evidence that it was the presence of concurrent load during training, rather than during testing, that was critical. Implications for associative, inferential, and dual-process accounts of human learning are discussed.     

Linking major depression and the neural substrates of associative processing

It has been proposed that mood correlates with the breadth of associative thinking. Here we set this hypothesis to the test in healthy and depressed individuals. Generating contextual associations engages a network of cortical regions including the parahippocampal cortex (PHC), retrosplenial complex, and medial prefrontal cortex. The link between mood, associative processing, and its underlying cortical infrastructure provides a promising avenue for elucidating the mechanisms underlying the cognitive impairments in major depressive disorder (MDD). The participants included 15 nonmedicated individuals with acute major depressive episodes and 15 healthy matched controls. In an fMRI experiment, participants viewed images of objects that were either strongly or weakly associated with a specific context (e.g., a beach chair vs. a water bottle) while rating the commonality of each object. Analyses were performed to examine the brain activation and structural differences between the groups. Consistent with our hypothesis, controls showed greater activation of the contextual associations network than did depressed participants. In addition, PHC structural volume was correlated with ruminative tendencies, and the volumes of the hippocampal subfields were significantly smaller in depressed participants. Surprisingly, depressed participants showed increased activity in the entorhinal cortex (ERC), as compared with controls. We integrated these findings within a mechanistic account linking mood and associative thinking and suggest directions for the future.     

Event-related fMRI of category learning: differences in classification and feedback networks

Eighteen healthy young adults underwent event-related (ER) functional magnetic resonance imaging (fMRI) of the brain while performing a visual category learning task. The specific category learning task required subjects to extract the rules that guide classification of quasi-random patterns of dots into categories. Following each classification choice, visual feedback was presented. The average hemodynamic response was calculated across the eighteen subjects to identify the separate networks associated with both classification and feedback. Random-effects analyses identified the different networks implicated during the classification and feedback phases of each trial. The regions included prefrontal cortex, frontal eye fields, supplementary motor and eye fields, thalamus, caudate, superior and inferior parietal lobules, and areas within visual cortex. The differences between classification and feedback were identified as (i) overall higher volumes and signal intensities during classification as compared to feedback, (ii) involvement of the thalamus and superior parietal regions during the classification phase of each trial, and (iii) differential involvement of the caudate head during feedback. The effects of learning were then evaluated for both classification and feedback. Early in learning, subjects showed increased activation in the hippocampal regions during classification and activation in the heads of the caudate nuclei during the corresponding feedback phases. The findings suggest that early stages of prototype-distortion learning are characterized by networks previously associated with strategies of explicit memory and hypothesis testing. However as learning progresses the networks change. This finding suggests that the cognitive strategies also change during prototype-distortion learning.     

Mutation at the TrkB PLC{gamma}-docking site affects hippocampal LTP and associative learning in conscious mice

Previous in vitro studies have characterized the electrophysiological properties and molecular events associated with long-term potentiation (LTP), but as yet there are no in vivo data from molecular-level dissection that directly identify LTP as the biological substrate for learning and memory. Understanding whether the molecular pathways required for learning are also those generating LTP when measured directly on the relevant circuit of a learning animal is clearly important, although so far has proved technically difficult. Here, for the first time, we combine highly defined genetic mouse models with behavior and in vivo recordings. We recorded the activity-dependent changes taking place at the CA3-CA1 synapses during the acquisition and extinction of a simple form of an associative learning task in mice carrying point mutations on specific docking sites of TrkB receptors (trkB(SHC), trkB(PLC)). The learning task consisted of a classical eyeblink conditioning using a trace paradigm. The conditioned stimulus (CS) consisted of a tone and was followed by a periorbital electrical shock as an unconditioned stimulus (US). The US started 500 msec after the end of the CS. We show that a single pulse presented to the Schaffer collateral-commissural pathway during the CS-US interval evoked a monosynaptic field excitatory postsynaptic potential (fEPSP) at the CA1 pyramidal cells, with a slope linearly related to learning evolution in controls and trkB(SHC) mutants, but the relationship was impaired in trkB(PLC) mice. These data support a link between the PLCgamma-docking site downstream of TrkB and the activity-dependent synaptic changes evoked at the CA3-CA1 synapses during associative learning in conscious mice, and indicate that TrkB PLCgamma-site-activated molecular pathway(s) underlie both associative learning and LTP triggered at the CA3-CA1 synapse.     

Impaired associative taste learning and abnormal brain activation in kinase-defective eEF2K mice

Memory consolidation is defined temporally based on pharmacological interventions such as inhibitors of mRNA translation (molecular consolidation) or post-acquisition deactivation of specific brain regions (systems level consolidation). However, the relationship between molecular and systems consolidation are poorly understood. Molecular consolidation mechanisms involved in translation initiation and elongation have previously been studied in the cortex using taste-learning paradigms. For example, the levels of phosphorylation of eukaryotic elongation factor 2 (eEF2) were found to be correlated with taste learning in the gustatory cortex (GC), minutes following learning. In order to isolate the role of the eEF2 phosphorylation state at Thr-56 in both molecular and system consolidation, we analyzed cortical-dependent taste learning in eEF2K (the only known kinase for eEF2) ki mice, which exhibit reduced levels of eEF2 phosphorylation but normal levels of eEF2 and eEF2K. These mice exhibit clear attenuation of cortical-dependent associative, but not of incidental, taste learning. In order to gain a better understanding of the underlying mechanisms, we compared brain activity as measured by MEMRI (manganese-enhanced magnetic resonance imaging) between eEF2K ki mice and WT mice during conditioned taste aversion (CTA) learning and observed clear differences between the two but saw no differences under basal conditions. Our results demonstrate that adequate levels of phosphorylation of eEF2 are essential for cortical-dependent associative learning and suggest that malfunction of memory processing at the systems level underlies this associative memory impairment.     

Associative retrieval processes in the human medial temporal lobe: hippocampal retrieval success and CA1 mismatch detection

Hippocampal subfields CA(3) and CA(1) are hypothesized to differentially support the generation of associative predictions and the detection of associative mismatches, respectively. Using high-resolution functional MRI, we examined hippocampal subfield activation during associative retrieval and during subsequent comparisons of memory to matching or mismatching decision probes. Activity in the dentate gyrus/CA(2/3), CA(1), and other medial temporal lobe subregions tracked associative retrieval success, whereas activity in CA(1) and the perirhinal cortex tracked the presence of associative mismatches. These data support the hypothesis that CA(1) acts as a "comparator," detecting when memory for the past and sensory input in the present diverge.     

Diminished motor skill development in elderly: Indications for limited motor chunk use

The present study examined whether elderly use motor chunks after practicing discrete keying sequences, just like young adults, or whether they perhaps learn these movement patterns in a different way. To that end, elderly (75-88) and young adults (18-28) practiced as part of the discrete sequence production (DSP) task two fixed series of three and six key presses. The results demonstrate that elderly did improve with practice but this improvement was largely sequence-unspecific. Detailed analyses showed that, in contrast to young adults, most elderly did not use motor chunks, had little explicit sequence knowledge, and remained highly dependent on external stimuli. Still, elderly did show sequence-specific learning with a 6-key sequence that can be explained by an associative learning mechanism.     

A simple neural network model of the hippocampus suggesting its pathfinding role in episodic memory retrieval

The goal of this work is to extend the theoretical understanding of the relationship between hippocampal spatial and memory functions to the level of neurophysiological mechanisms underlying spatial navigation and episodic memory retrieval. The proposed unifying theory describes both phenomena within a unique framework, as based on one and the same pathfinding function of the hippocampus. We propose a mechanism of reconstruction of the context of experience involving a search for a nearly shortest path in the space of remembered contexts. To analyze this concept in detail, we define a simple connectionist model consistent with available rodent and human neurophysiological data. Numerical study of the model begins with the spatial domain as a simple analogy for more complex phenomena. It is demonstrated how a nearly shortest path is quickly found in a familiar environment. We prove numerically that associative learning during sharp waves can account for the necessary properties of hippocampal place cells. Computational study of the model is extended to other cognitive paradigms, with the main focus on episodic memory retrieval. We show that the ability to find a correct path may be vital for successful retrieval. The model robustly exhibits the pathfinding capacity within a wide range of several factors, including its memory load (up to 30,000 abstract contexts), the number of episodes that become associated with potential target contexts, and the level of dynamical noise. We offer several testable critical predictions in both spatial and memory domains to validate the theory. Our results suggest that (1) the pathfinding function of the hippocampus, in addition to its associative and memory indexing functions, may be vital for retrieval of certain episodic memories, and (2) the hippocampal spatial navigation function could be a precursor of its memory function.     

Prefrontal and hippocampal contributions to visual associative recognition: interactions between cognitive control and episodic retrieval

The ability to recover episodic associations is thought to depend on medial-temporal lobe mnemonic mechanisms and frontal lobe cognitive control processes. The present study examined the neural circuitry underlying non-verbal associative retrieval, and considered the consequences of successful retrieval on cognitive control demands. Event-related fMRI data were acquired while subjects retrieved strongly or weakly associated pairs of novel visual patterns in a two-alternative forced choice associative recognition paradigm. Behaviorally, successful retrieval of strongly associated relative to weakly associated pairs was more likely to be accompanied by conscious recollection of the pair's prior co-occurrence. At the neural level, right ventrolateral prefrontal cortex (VLPFC) and hippocampus were more active during successful retrieval of Strong than of Weak associations, consistent with a role in visual associative recollection. By contrast, Weak trials elicited greater activation in right anterior cingulate cortex (ACC), which may detect conflict between the similarly familiar target and foil stimuli in the absence of recollection. Consistent with this interpretation, stronger ACC activity was associated with weaker hippocampal and stronger right dorsolateral PFC (DLPFC) responses. Thus, recollection of relevant visual associations (hippocampus and VLPFC) results in lower levels of mnemonic conflict (ACC) and decreased familiarity-based monitoring demands (DLPFC). These findings highlight the interplay between cognitive control and episodic retrieval.