Neural modification by paired sensory stimuli

With repetitive stimulation of two sensory pathways which are intact within the isolated nervous system of Hermissenda, features of a cellular conditioning paradigm were identified. Type A photoreceptors, unlike type B photoreceptors, produce fewer impulses in response to light following temporally specific pairing of light stimuli with rotation stimuli. Type A photoreceptor impulse wave-forms are also specifically changed by such stimulus regimens. These findings can be explained, at least in part, by increased inhibition of type A cells by type B cells after stimulus pairing.     

Paired Turbulence and Light do not Produce a Supralinear Calcium Increase in Hermissenda

The sea slug Hermissenda learns to associate light and hair cell stimulation, but not when the stimuli are temporally uncorrelated. Memory storage, which requires an elevation in calcium, occurs in the photoreceptors, which receive monosynaptic input from hair cells that sense acceleration stimuli such as turbulence. Both light and hair cell activity increase calcium concentration in the photoreceptor, but it is unknown whether paired calcium signals combine supralinearly to initiate memory storage. A correlate of memory storage is an enhancement of the long lasting depolarization (LLD) after light offset, which is attributed to a reduction in voltage dependent potassium currents; however, it is unclear what causes the LLD in the untrained animal.These issues were addressed using a multi-compartmental computer model of phototransduction, calcium dynamics, and ionic currents of the Hermissenda photoreceptor. Simulations of the interaction between light and hair cell activity show that paired stimuli do not produce a greater calcium increase than unpaired stimuli. This suggests that hair cell activity is acting via some other pathway to initiate memory storage. In addition, simulations show that a potassium leak channel, which closes with an increase in calcium, is required to produce both the untrained LLD and the enhanced LLD due to the decrease in voltage dependent potassium currents. Thus, the expression of this correlate of classical conditioning may depend on a leak potassium current.     

Biochemical mechanisms of memory storage

A series of studies on Hermissenda classical conditioning has lead to a discovery that the biophysical events (accumulation of Ca2+ and depolarization in B cell) found during memory acquisition are clearly distinct from those (suppression of K-currents, IA and ICa2+K+) detected in the retention phase of memory. Biochemical analysis of eyes isolated shortly after (a few hours) training revealed increased phosphorylation of a 20,000 M.W. protein which is very likely one of the substrates for both Ca/CaM-dependent protein kinase and C-kinase and possibly a locus of convergence for conditioned stimulus and unconditioned stimulus pathways. Furthermore, conditioning-specific changes in the two K+ currents have been reproduced by simultaneous activation of the CaM-kinase pathway (via iontophoretic injection of CaM-kinase II plus Ca2+-load or IP3 injection) and the C-kinase pathway (via bath application of phorbol-ester or diacylglycerol analog plus Ca2+-load). Therefore, synergistic interaction between the two Ca2+-dependent phosphorylation systems in the identified B cell is considered to be critically important for acquisition of associative memory. Evidence also has been obtained for similar biophysical changes and molecular mechanisms during retention of classical conditioning in the mammalian brain. Further work will be needed to uncover the biochemical mechanism(s) responsible for transforming short-term into long-lasting memory.     

Effects of α2-Adrenergic Agonists and Antagonists on Photoreceptor Membrane Currents

Type B photoreceptors of the nudibranch mollusc Hermissenda crassicornis receive excitatory synaptic potentials (EPSPs) whose frequency is controlled by potential changes of a neighboring cell known as the S optic ganglion cell which is thought to be electrically coupled to the presynaptic source of these EPSPs, the E optic ganglion cell. The frequency of the EPSPs increases when a conditioned stimulus (light) is paired with an unconditioned stimulus (rotation) during acquisition of a Pavlovian conditioned response. The results of the present study are consistent with an adrenergic origin for these EPSPs. Noradrenergic agonists (greater than 100 microM), norepinephrine and clonidine, only slightly depolarize the type B cell but clearly prolong its depolarizing response to light. Serotonin, by contrast, causes hyperpolarization of the type B cell's resting potential as well as after a light step. Clonidine reduces voltage-dependent outward K+ currents (IA, an early current, ICa2+-K+, a late Ca2+-dependent current) that control the type B cell's excitability (and thus its light response and membrane potential). These effects of clonidine are reduced or blocked by the alpha 2-receptor antagonist, yohimbine (0.5 microM), but not the alpha 1-blocker, prazosin. The same yohimbine concentration also blocked depolarizing synaptic excitation of the type B cell in response to depolarization of a simultaneously impaled S optic ganglion cell. Histochemical techniques (both the glyoxylic acid method of de la Torre and Surgeon and the formaldehyde-induced fluorescence or Falck-Hillarp method) demonstrated the presence of a biogenic amine(s) within a single neuron in each optic ganglion as well as three or four cells within the vicinity of previously identified visual interneurons. No serotonergic neurons were found within the optic ganglion or in proximity to visual interneurons. A clonidine-like synaptic effect on type B cells, therefore, could amplify conditioning-specific changes of membrane currents by increasing type B depolarization and possibly, as well, by elevating intracellular second messengers.     

Trace Fear Conditioning in Mice

In this experiment we present a technique to measure learning and memory. In the trace fear conditioning protocol presented here there are five pairings between a neutral stimulus and an unconditioned stimulus. There is a 20 sec trace period that separates each conditioning trial. On the following day freezing is measured during presentation of the conditioned stimulus (CS) and trace period. On the third day there is an 8 min test to measure contextual memory. The representative results are from mice that were presented with the aversive unconditioned stimulus (shock) compared to mice that received the tone presentations without the unconditioned stimulus. Trace fear conditioning has been successfully used to detect subtle learning and memory deficits and enhancements in mice that are not found with other fear conditioning methods. This type of fear conditioning is believed to be dependent upon connections between the medial prefrontal cortex and the hippocampus. One current controversy is whether this method is believed to be amygdala-independent. Therefore, other fear conditioning testing is needed to examine amygdala-dependent learning and memory effects, such as through the delay fear conditioning.     

Does conditioned taste aversion learning in the pond snail Lymnaea stagnalis produce conditioned fear?

In conditioned taste aversion (CTA) training performed on the pond snail Lymnaea stagnalis, a stimulus (the conditional stimulus, CS; e.g., sucrose) that elicits a feeding response is paired with an aversive stimulus (the unconditional stimulus, US) that elicits the whole-body withdrawal response and inhibits feeding. After CTA training and memory formation, the CS no longer elicits feeding. We hypothesize that one reason for this result is that after CTA training the CS now elicits a fear response. Consistent with this hypothesis, we predict the CS will cause (1) the heart to skip a beat and (2) a significant change in the heart rate. Such changes are seen in mammalian preparations exposed to fearful stimuli. We found that in snails exhibiting long-term memory for one-trial CTA (i.e., good learners) the CS significantly increased the probability of a skipped heartbeat, but did not significantly change the heart rate. The probability of a skipped heartbeat was unaltered in control snails given backward conditioning (US followed by CS) or in snails that did not acquire associative learning (i.e., poor learners) after the one-trial CTA training. These results suggest that as a consequence of acquiring CTA, the CS evokes conditioned fear in the conditioned snails, as evidenced by a change in the nervous system control of cardiac activity.     

Potassium currents in presynaptic hair cells of Hermissenda.

Apart from their primary function as balance sensors, Hermissenda hair cells are presynaptic neurons involved in the Ca(2+)-dependent neuronal plasticity in postsynaptic B photoreceptors that accompanies classical conditioning. With a view to beginning to understand presynaptic mechanisms of plasticity in the vestibulo-visual system, a locus for conditioning-induced neuronal plasticity, outward currents that may govern the excitability of hair cells were recorded by means of a whole-cell patch-clamp technique. Three K+ currents were characterized: a 4-aminopyridine-sensitive transient outward K+ current (IA), a tetraethyl ammonium-sensitive delayed rectifier K+ current (IK,V), and a Ca(2+)-activated K+ current (IK,Ca). IA activates and decays rapidly; the steady-state activation and inactivation curves of the current reveal a window current close to the apparent resting voltage of the hair cells, suggesting that the current is partially activated at rest. By modulating firing frequency and perhaps damping membrane oscillations, IA may regulate synaptic release at baseline. In contrast, IK,V and IK,Ca have slow onset and exhibit little or no inactivation. These two K+ currents may determine the duration of the repolarization phase of hair-cell action potentials and hence synaptic release via Ca2+ influx through voltage-gated Ca2+ channels. In addition, IK,Ca may be responsible for the afterhyperpolarization of hair cell membrane voltage following prolonged stimulation.     

Interaction of chemosensory, visual, and statocyst pathways in Hermissenda crassicornis

Neurons in the cerebropleural ganglia (CPG), photoreceptors in the eye, optic ganglion cells, and statocyst hair cells of the nudibranch mollusk Hermissenda crassicornis responded in specific ways, as recorded intracellularly, to stimulation of the chemosensory pathway originating at the tentacular chemoreceptors as well as to stimulation of the visual pathway originating at the photoreceptors. Synaptic inhibition of photoreceptors occurs via the chemosensory pathway. The possible significance of such intersensory interaction is discussed with reference to preliminary investigation of the animal's gustatory behavior and possible neural mechanisms of behavioral choice.     

Dynamical aspects of in vitro conditioning in Hermissenda type B photoreceptor

Dynamics of changes in physiology and morphology were studied in Hermissenda photoreceptors after in vitro conditioning with paired light and vibration. An increase in input resistance of the type B photoreceptor was observed following 5 paired presentations of light and vibration. It peaked at 10 min after in vitro conditioning, then decreased to a level twice the pre-conditioning level for more than 60 min. Contraction of the terminal branches along centro-lateral direction was initiated 5 min after conditioning and reached its final state at 10 min after conditioning. The pairing specific contraction of the axon terminal was not observed in ASW containing anisomycin. The dynamics in physiology and morphology were completely parallel 30 min after conditioning. These findings suggested that in vitro conditioning induced contraction was dependent on protein synthesis dependent process initiated within 5 min after training trials and that the change of cell morphology is a form of short-term synaptic plasticity that involves changes in macromolecular synthesis. Present findings that functional remodeling at the terminal branch of the type B photoreceptor occurred within 10 min after conditioning was the fastest modification process reported so far.     

The role of calcium in prolonged modification of a GABAergic synapse.

Caudal hair cell impulses cause postsynaptic inhibition of ipsilateral type B photoreceptors in the snail Hermissenda. This inhibition is shown to be GABAergic according to a number of criteria. HPLC, mass spectrophotometric, and immunocytochemical techniques demonstrated the presence of GABA in the hair cells and their axons. GABA agonists and antagonists mimic and block the synaptic effect in a manner consistent with endogenous GABAergic transmission. Other properties, including I-V relations, conductance changes and reversal potentials, are comparable for exogenous GABA responses and endogenous effects of the hair cell impulses. This inhibitory synapse has been found to undergo a long-lasting transformation into an excitatory synapse if GABA release is paired with post-synaptic depolarization. GABA, via GABAA and GABAB receptors in the B cell, causes the opening of calcium sensitive chloride and potassium channels that leads to the post-synaptic hyperpolarization. GABA also induces a long-lasting intracellular calcium elevation at the terminal branches of the B cell that greatly outlasts the voltage responses. Synaptic transformation induced by pairings is caused by a decrease in both GABA induced chloride and potassium conductances in the post-synaptic B cell, as well as a significant prolongation of the intracellular calcium accumulation in the B cell's terminal axonal branches.     

Learning-induced activation of protein kinase C

PKC activation has been shown to mimic the biophysical consequences of classical conditioning in both rabbit hippocampus and Hermissenda type B cells. Furthermore, conditioning in rabbits results in the 24 h translocation of PKC from cytosol to membrane, which is probably responsible for mediating the biophysical consequences of conditioning. A model has been presented that suggests that long-term translocation of PKC occurs via the synergistic activation of a DG dependent pathway that activates PKC and a calcium dependent pathway that activates CaM kinase. Translocation of PKC to the plasma membrane, by altering ion channel properties, could subserve memory lasting for days, whereas translocation to the nuclear membrane could induce cellular change, by genomic regulation, lasting beyond days. We are, therefore, suggesting that protein kinase C may play a critical role in the formation of short, intermediate, and long-term associative memory.     

Conditioning and time representation in long short-term memory networks

Dopaminergic models based on the temporal-difference learning algorithm usually do not differentiate trace from delay conditioning. Instead, they use a fixed temporal representation of elapsed time since conditioned stimulus onset. Recently, a new model was proposed in which timing is learned within a long short-term memory (LSTM) artificial neural network representing the cerebral cortex (Rivest et al. in J Comput Neurosci 28(1):107–130, 2010). In this paper, that model’s ability to reproduce and explain relevant data, as well as its ability to make interesting new predictions, are evaluated. The model reveals a strikingly different temporal representation between trace and delay conditioning since trace conditioning requires working memory to remember the past conditioned stimulus while delay conditioning does not. On the other hand, the model predicts no important difference in DA responses between those two conditions when trained on one conditioning paradigm and tested on the other. The model predicts that in trace conditioning, animal timing starts with the conditioned stimulus offset as opposed to its onset. In classical conditioning, it predicts that if the conditioned stimulus does not disappear after the reward, the animal may expect a second reward. Finally, the last simulation reveals that the buildup of activity of some units in the networks can adapt to new delays by adjusting their rate of integration. Most importantly, the paper shows that it is possible, with the proposed architecture, to acquire discharge patterns similar to those observed in dopaminergic neurons and in the cerebral cortex on those tasks simply by minimizing a predictive cost function.     

Conditioning of active type avoidance in the rat : effect of the interval between conditioned stimulus and unconditioned stimulus

Lengthening the time interval between the conditioned stimulus and the unconditioned stimulus increases the number of active avoidance conditioned responses in subjects that have been trained to a stable level of performance in many previous conditioning sessions. In the present research, rats chosen from a population specially selected for low rates of avoidance conditioning have been used. In addition to this characteristic, subjects were chosen for the exhibition of an apparent absence of retention from one day to another. The dependency of the number of conditioned responses on the time interval between conditioned stimulus and unconditioned stimulus may lead to wrong evaluation of the subjects' conditioning level. In fact, the level of conditioning may be attributed to either learning or memory processes when in many cases it is determined only by the latency time of the conditioned response. The conditioned response has no possibility of manifesting itself when its latency time exceeds in length the time interval between conditioned stimulus and unconditioned stimulus.     

Effects of an extinguished CS on competition with another CS

Three experiments were conducted using a conditioned taste aversion procedure with rats to examine the effect of nonreinforced presentations of a conditioned stimulus (CS) on its ability to compete with a target stimulus for manifest conditioned responding. Two CSs (A and B) were presented in a serial compound and then paired with the unconditioned stimulus. CS A was first paired with the US and then presented without the US (i.e., extinction) prior to reinforced presentation of the AB compound. Experiment 1 showed that A was poor at competing with B for conditioned responding when given conditioning and extinction prior to reinforcement of AB relative to a group that received both A and B for the first time during compound conditioning. That is, an extinguished A stimulus allowed greater manifest acquisition to B. Experiment 2 found that extinction treatment produced a poor CR to the pretrained and extinguished CS itself following compound conditioning. Experiment 3 found that interposing a retention interval after extinction of A and prior to compound conditioning enhanced A's ability to compete with B. The results of these experiments are discussed with regard to different theories of extinction and associative competition.     

Neural Organization of a Molluscan Visual System

Intracellular recording was used to study the response to light of second order visual cells within the optic ganglion of Hermissenda crassicornis. Simultaneous recordings revealed that type B but not type A photoreceptors inhibit the second order cells. Additional details of the neural organization of the visual system were obtained. Possible functional implications of this neural organization are discussed.     

Associative learning in a network model of Hermissenda crassicornis

A companion paper in a previous issue of this journal presented a resistance-capacitance circuit computer model of the four-neuron visual-vestibular network of the invertebrate marine mollusk Hermissenda crassicornis. In the present paper, we demonstrate that changes in the model's output in response to simulated associative training is quantitatively similar to behavioral and electrophysiological changes in response to associative training of Hermissenda crassicornis. Specifically, the model demonstrates many characteristics of conditioning: sensitivity to stimulus contingency, stimulus specificity, extinction, and savings. The model's learning features also are shown to be devoid of non-associative components. Thus, this computational model is an excellent tool for examining the information flow and dynamics of biological associative learning and for uncovering insights concerning associative learning, memory, and recall that can be applied to the development of artificial neural networks.     

Reversing Stimulus Timing in Visual Conditioning Leads to Memories with Opposite Valence in Drosophila

Animals need to associate different environmental stimuli with each other regardless of whether they temporally overlap or not. Drosophila melanogaster displays olfactory trace conditioning, where an odor is followed by electric shock reinforcement after a temporal gap, leading to conditioned odor avoidance. Reversing the stimulus timing in olfactory conditioning results in the reversal of memory valence such that an odor that follows shock is later on approached (i.e. relief conditioning). Here, we explored the effects of stimulus timing on memory in another sensory modality, using a visual conditioning paradigm. We found that flies form visual memories of opposite valence depending on stimulus timing and can associate a visual stimulus with reinforcement despite being presented with a temporal gap. These results suggest that associative memories with non-overlapping stimuli and the effect of stimulus timing on memory valence are shared across sensory modalities.     

Associative learning in ants: Conditioning of the maxilla-labium extension response in Camponotus aethiops

Associative learning has been studied in many vertebrates and invertebrates. In social insects, the proboscis extension response conditioning of honey bees has been widely used for several decades. However, a similar paradigm has not been developed for ants, which are advanced social insects showing different morphological castes and a plethora of life histories. Here we present a novel conditioning protocol using Camponotus aethiops. When the antennae of a harnessed ant are stimulated with sucrose solution, the ant extends its maxilla-labium to absorb the sucrose. We term this the “maxilla-labium extension response” (MaLER). MaLER could be conditioned by forward pairing an odour (conditioned stimulus) with sucrose (unconditioned stimulus) in the course of six conditioning trials (absolute conditioning). In non-rewarded tests following conditioning, ants gave significantly higher specific responses to the conditioned stimulus than to a novel odour. When trained for differential conditioning, ants discriminated between the odour forward-paired with sucrose and an odour forward-paired with quinine (a putative aversive stimulus). In both absolute and differential conditioning, memory lasted for at least 1 h. MaLER conditioning allows full control of the stimulation sequence, inter-stimulus and inter-trial intervals and satiety, which is crucial for any further study on associative learning in ants.     

Bryostatin enhancement of memory in Hermissenda

Bryostatin, a potent agonist of protein kinase C (PKC), when administered to Hermissenda was found to affect acquisition of an associative learning paradigm. Low bryostatin concentrations (0.1 to 0.5 ng/ml) enhanced memory acquisition, while concentrations higher than 1.0 ng/ml down-regulated the pathway and no recall of the associative training was exhibited. The extent of enhancement depended upon the conditioning regime used and the memory stage normally fostered by that regime. The effects of two training events (TEs) with paired conditioned and unconditioned stimuli, which standardly evoked only short-term memory (STM) lasting 7 min, were--when bryostatin was added concurrently--enhanced to a long-term memory (LTM) that lasted about 20 h. The effects of both 4- and 6-paired TEs (which by themselves did not generate LTM), were also enhanced by bryostatin to induce a consolidated memory (CM) that lasted at least 5 days. The standard positive 9-TE regime typically produced a CM lasting at least 6 days. Low concentrations of bryostatin (<0.5 ng/ml) elicited no demonstrable enhancement of CM from 9-TEs. However, animals exposed to bryostatin concentrations higher than 1.0 ng/ml exhibited no behavioral learning. Sharp-electrode intracellular recordings of type-B photoreceptors in the eyes from animals conditioned in vivo with bryostatin revealed changes in input resistance and an enhanced long-lasting depolarization (LLD) in response to light. Likewise, quantitative immunocytochemical measurements using an antibody specific for the PKC-activated Ca2+/GTP-binding protein calexcitin showed enhanced antibody labeling with bryostatin. Animals exposed to the PKC inhibitor bisindolylmaleimide-XI (Ro-32-0432) administered by immersion prior to 9-TE conditioning showed no training-induced changes with or without bryostatin exposure. However, if animals received bryostatin before Ro-32, the enhanced acquisition and demonstrated recall still occurred. Therefore, pathways responsible for the enhancement effects induced by bryostatin were putatively mediated by PKC. Overall, the data indicated that PKC activation occurred and calexcitin levels were raised during the acquisition phases of associative conditioning and memory initiation, and subsequently returned to baseline levels within 24 and 48 h, respectively. Therefore, the protracted recall measured by the testing regime used was probably due to bryostatin-induced changes during the acquisition and facilitated storage of memory, and not necessarily to enhanced recall of the stored memory when tested many days after training.     

Pavlovian conditioning-specific increases of the Ca2+- and GTP-binding protein,calexcitin in identified Hermissenda visual cells

Hermissenda CNS, immunolabeled for the memory protein calexcitin showed significant immunostaining over background in the B-photoreceptor cells of the eye. The degree of staining correlated positively with the number of Pavlovian training events experienced by the animals and the degree of Pavlovian conditioning induced. The training regime consisted of exposing animals to light (conditioned stimulus, CS) paired with orbital rotation (unconditioned stimulus, US). In animals that exhibited the conditioned response, calexcitin immunolabeling was more intense than was found for naive (unconditioned) animals or animals given the CS and US in random sequence. Animals exposed to lead (maintained in 1.2 ppm lead acetate) at a dosage known to impair learning in children, showed reduced learning and less intense calexcitin staining whether the CS and US were paired or given randomly. However, the levels were still higher than that of naive animals. Immuno-electron microscopy indicated that the labeling was predominantly within calcium sequestering organelles such as the endoplasmic reticulum, and to lesser extent within mitochondria, and photopigments. The calexcitin density after a short-term memory (STM) regime was the same whether measured 5 minutes after conditioning (when STM was evidenced by foot contraction) or 90 minutes later when no recall was detected. The staining density was also similar to the levels found 5 minutes after long-term memory (LTM) conditioning. However, the LTM regime produced a greater calexcitin intensity at 90 minutes when the memory had been consolidated. This learning-specific increase in calexcitin is consistent with the previously implicated sequence of molecular events that are associated with progressively longer time domains of memory storage.