Learning and Memory

In addition to the different perspectives inherent to behavior- ism and cognitivism, further distinctions are often drawn between the researchers who study learning and memory in humans versus animals. Researchers who study human learning and memory tend to adopt cognitivism, whereas researchers who study learning and memory in animals have traditionally adopted behaviorism. Debates are ongoing over which perspec- tive is the most valuable for understanding learning and mem- ory. Both behaviorist and cognitivist approaches have led to crucial insights in the attempt to understand learning and mem- ory. The same can be said for researchers who study learning and memory in animals and those who study humans. The best research model depends on the specific question being addressed. For example, a cognitivist approach in studies using a rodent yields little information because the complex mental components of learning and memory cannot easily be assessed in animals. You cannot easily ask a rodent how it really felt about having to run mazes to get a cheese reward. In contrast, rodents can be experimentally manipulated to test various aspects of learning and memory that would be unacceptable to test in hu- mans. The behaviorist doesn’t really care how the rodent felt about running the maze, nor what its conscious motivation was for failing to do the task, but rather, wants to know if it can still accomplish the task if a specific part of its brain has been re- moved or damaged.

We propose a hypothetical framework for learning-induced associative plasticity that integrates task-dependent effects with the factor of learning strategy ( Fig. S3 ). This framework is built upon three facets of HARP: (1) the factors for its induction, (2) the forms and degrees that HARP can take along the relevant stim- ulus dimension (e.g., tonal frequency) and (3) the functions of its establishment. We suggest that the induction of HARP is deter- mined in a filter-like process that begins with task-specific situa- tional and subjective factors. These can influence the establishment of a learning strategy that, in turn, has direct influ- ence over the induction of HARP, in its presence, form (e.g., thresh- old and bandwidth decreases, area gains, etc.) and degree (e.g., amount of area gain). Once HARP is established, it will have one or many functional consequences (e.g., memory storage, perceptual acuity, initiating goal-directed behavior, etc.). For example, HARP might influence an animal’s learning strategy in future learning sit- uations. Also, if new task-specific variables are introduced with additional training, then HARP could be updated through a new cy- cle. A full discussion of the implications of neurophysiological assessment of plasticity ‘‘online” during learning in an active, behaving animal versus that which is observed after training out- side of the context of learning is beyond the scope of the present discussion and has been reviewed elsewhere ( Weinberger, 2004 ). However, we suggest that if learning strategy changes during task performance (indexed by a shift in the animal’s pattern of behav- ior), then the change would dictate a modification in HARP, and a new form or degree of HARP would be observed after training.

For learning and memory to take place, many different parts of your brain must be working together properly. Did you ever wonder how neuro- scientists find out which parts of the brain are working when you read or speak, or when you try to learn or remember something? One way is through laboratory research. Most laboratory studies on the brain are done with animals other than humans. Many different kinds of animals are used in investigations that produce important information about how the brain works.

Immunosuppressant drugs have produced a cognitive impairment is associated with degeneration of hippocampal neurons histopathologically. Since alteration of immune function affects learning and memory, it was hypothesized that immunostimulant drugs improves learning and memory. The constituents of Saraswatarishta such as Withania somnifera, Tinospora cordifolia, Terminalia chebula, Piper longum, Pueraria tuberose and Asparagus racemosus have been proved for cognitive enhancement. The probable mechanism of cognitive enhancement by Saraswatarishta could be by immunostimulation and increasing the synthesis of acetylcholine which is an important neurotransmitter in learning and memory process (Bisset and Nwai, 1983). This central action could be due to supplementation of choline by Tinospora cordifolia which is an important active constituent of Saraswatarishta (Toes and Mohammed, 1999).

Learning and memory being highly specialized process of human brain involves complex interaction between neurotransmitters and cellular events. Over the years, the understandings of these processes have been evolving from psychological, neurophysiological, and pharmacological perspectives. The most widely appraised model of learning and memory involves attention, acquisition, storage and retrieval. Each of these events involve interplay of neurotransmitters such as dopamine, acetylcholine, norepinephrine, N-methyl- d-aspartic acid, gamma-aminobutyric acid, though preponderance of specific neurotransmitter have been documented. The formation of long-term memory involves cellular events with neuroplasticity. Further, dopamine is documented to play crucial role in the process of forgetting. Understanding of the processes of learning and memory not only facilitates drug discovery, but also helps to understand actions of several existing drugs. In addition, it would also help to enhance psychological interventions in children with learning disabilities. Thus, the review intends to summarize role of neurotransmitters and neuromodulators during different phases of learning and memory.

Learning is acquisition of new knowledge and skills, whereas memory is retention and retrievals of facts composed of experiences (Kumar et al., 2012). Alzheimer’s disease is a progressive neurodegenerative disorder characterized by the gradual onset of dementia. It is characterized by gradually progressive decline in cognitive function, with deficits especially in memory retrieval (Kalaria et al., 2008). The primary cause of Alzheimer’s disease appears to be reduction in cholinergic activity. Acetylcholinesterase enzyme (AChE) plays a key role in the regulation of the cholinergic system and hence, inhibition of AChE is a most promising target for the treatment of Alzheimer’s disease (Lu et al., 2010). A number of cholinesterase inhibitors like donepezil, rivastigmine, tacrine and rivastigmine etc. are in practice for the treatment of various cognitive disorders (Ellis, 2005). However, the adverse effects associated with anti-cholinesterase drugs (rivastigmine, galantamine, donepezil) include anorexia, nausea, vomiting, diarrhoea, and insomnia (Kavirajan and Schneider, 2007). Physostigmine, a cholinesterase inhibitor, improved memory of Alzheimer’s disease patients (Mohs et al., 1985). But this drug has a short half-life and requires complex forms of administration (Filho and Birks, 2001). So there is a need to discover new drugs with better efficacy and having less adverse effects. Plants have been used since ancient times in traditional medicinal systems for the treatment of memory dysfunction. Bacopa monniera , commonly called as Brahmi, has been proven to be clinically effective in treatment of cognitive disorders (Pase et al., 2012). The bioactive compounds isolated from plants such as galantamine (Raskind et al., 2000), huperzine alpha (Zhang et al., 2002), etc. have been reported to be effective for treatment of cognitive deficits in patients of dementia.

Some of these factors are under the direct control of the e-learning designers and developers. Investigating these factors and how they impact on memory is important and can enhance the quality of e-learning. Evidence from cognitive neuroscience suggests that information is stored in a semantically meaningful manner. It follows that e-learning technologies, which either mimic how knowledge is structured in the mind or which allow individuals to organise their own exploration of the information space, should facilitate learning and memory. However, the ability to freely explore an information space is also more taxing on an individual’s cognitive resources. Learners would need to expend cognitive resources remembering where they have been, as well as on deciding where to go next. These additional demands may impede learning, especially for more complex information. Most critical to investigate is the trade-off between the ability to build knowledge according to the learners’ cognitive structures and style (which is also more engaging), on the one hand; and the extra cognitive load associated with giving learners more control and information, on the other hand.

ABSTRACT Data on the effects of serotonin reuptake inhibitors on learning and memory processes are not consistent. In the present study, the effects of citalopram, a very potent and completely selective inhibitor of the serotonin reuptake on spatial discrimination in the T-maze and Morris water maze, were assessed in mice and/or rats. Animals received different doses of citalopram (1, 2, 4, 8 or 16 mg/kg, i.p.) or its vehicle (saline) 30 min before training each day. The results showed no significant effects of citalopram on T-maze discrimination task in mice and rats. However, there was dose-dependent increases in latencies to find the invisible platform and traveled distances in citalopram received groups compared to the control group with the peak effect at doses of 4 and 8 mg/kg in Morris task. Therefore, it appears that citalopram can cause learning deficits in complex spatial tasks. Iran. Biomed. J. 4: 21-29, 2000

Abstract Animals constantly receive and respond to external or internal stimuli, and these experiences are learned and memorized in their brains. In animals, this is a crucial feature for survival, by making it possible for them to adapt their behavioral patterns to the ever-changing environment. For this learning and memory process, nerve cells in the brain undergo enormous molecular and cellular changes, not only in the input-output-related local subcellular compartments but also in the central nucleus. Interestingly, the DNA methylation pattern, which is normally stable in a terminally differentiated cell and defines the cell type identity, is emerging as an important regulatory mechanism of behavioral plasticity. The elucidation of how this covalent modification of DNA, which is known to be the most stable epigenetic mark, contributes to the complex orchestration of animal behavior is a fascinating new research area. We will overview the current understanding of the mechanism of modifying the methyl code on DNA and its impact on learning and memory.

ABSTRACT Cannabinoids are hypothesized to play an important role in modulating learning and memory formation. Here, we identified mRNAs expressed in Lymnaea stagnalis central nervous system that encode two G-protein-coupled receptors (Lymnaea CBr-like 1 and 2) that structurally resemble mammalian cannabinoid receptors (CBrs). We found that injection of a mammalian CBr agonist WIN 55,212-2 (WIN 55) into the snail before operant conditioning obstructed learning and memory formation. This effect of WIN 55 injection persisted for at least 4 days following its injection. A similar obstruction of learning and memory occurred when a severe traumatic stimulus was delivered to L. stagnalis. In contrast, injection of a mammalian CBr antagonist AM 251 enhanced long-term memory formation in snails and reduced the duration of the effects of the severe traumatic stressor on learning and memory. Neither WIN 55 nor AM 251 altered normal homeostatic aerial respiratory behaviour elicited in hypoxic conditions. Our results suggest that putative cannabinoid receptors mediate stressful stimuli that alter learning and memory formation in Lymnaea. This is also the first demonstration that putative CBrs are present in Lymnaea and play a key role in learning and memory formation.

To further examine its role in memory formation, in vivo, we infused Wnt1 into the amygdala either before or after fear conditioning. We found that infusing Wnt1 in the amygdala prior to training did not affect acquisition, but resulted in deficits in the expression of fear as measured 48 h after training. The infusion of Wnt1 after training did not have an effect on learning. Given that acquisition of fear appeared to be normal in the presence of Wnt1, these data suggest that we may be preventing the normal rapid decrease in Wnt signaling that occurs at the end of training and immediately during the early period of consolidation. These data suggest that an immediate decrease in Wnt1 with learning may be required for memory formation; thus, the infusion of Wnt1 prior to training, may allow just enough time to reverse that downregulation of Wnt1 gene expression. Conversely, infusing Wnt1 after training, may not increase expression of Wnt1 until after a critical period of early consolidation has passed. Therefore, our results suggest that inhibiting the decrease in Wnt1 expression which occurs with training produces memory deficits.

Dr. SB Kasture, Sanjivani College of Pharmaceutical Education and Research, Kopargaon, Maharashtra *sangeeta1709@gmail.com Abstract Learning and memory is a complex process mediated by number of receptors and subproteins. Different receptor subtypes plays different role in learning and memory process. An array of mediators like noradrenaline, acetylcholine, dopamine (DA), serotonin (5-HT), GABA, glutamate, nitric oxide and peptides influence cognitive behaviour of the animal. Long term potentiation involving synaptic plasticity in cognition process is mediated by interaction of dopamine and glutamate in different brain regions. Aim of the present review is to highlight the role of different receptors subtypes in modulation of learning and memory process as evidenced by different studies thus providing a source of information for development of new therapeutic strategy for dysfunction in memory through targeting specific receptors subtypes.

The initial focus on bit memory over gated memory is understandable: reinforce- ment learning is about abstract knowledge. Furthermore, as we observed in Gorski & Laird (2011), many partially observable tasks require only a single bit of infor- mation in order to completely capture all aspects of history and make the task fully observable to a reinforcement learning agent; this holds true for tasks used through- out the literature that explored learning to use memory. Given that bit memory was sufficient to represent an optimal policy, exploring alternative memory mechanisms with architectural constraints on behavior and additional representation capacity is an unintuitive direction. Thus only as statistical reinforcement learning, cognitive science and computational neuroscience communities have begun to mix has atten- tion shifted to the capabilities that a perceptually grounded memory (such as gated memory) can afford an agent.

The interaction is the key element used in Collaborative Learning Environments to understand the process of knowledge building and the role played by each student in it. Interaction analysis can provide support for the students' reflection and self-regulation processes as well as for the teachers' activities. But to perform the analysis process, it is important to discover and register the context where each interaction has occurred, in order to understand the meaning of user interactions. However, although there are several approaches for Interaction Analysis in Computer-Supported Collaborative Learning, there is a lack of methods and tools that consider: (1) the context where the interactions have occurred; (2) the different needs of feedback from the point of view of teacher and students; and (3) the necessity of contextualized historical information to produce more complete and semantically rich reports for students and teachers. In this light, this paper presents a Learning Interaction Memory (LIM), used to store the learning interactions occurred in Computer-Supported Collaborative Learning Environments, taking into account the context where they have occurred. The LIM was modelled in a multidimensional structure so that interactions can be viewed from different perspectives and can be presented selectively, according to users' needs. This paper also presents the process of construction and exploration of the LIM and a Context-Based Analytical Environment called SmartChat+: an environment for collaborative discussions

[2 + ] Leonid Peshkin. Reinforcement Learning by Policy Search. PhD thesis. Massachusetts Institute of Technology. 2003. - PhD thesis of the first author, includes VAPS work and investigates other methods of controllers with memory [3 + ] Bram Bakker. Reinforcement learning with long short-term memory. Advances in Neural Information Processing Systems. 2002. - Solves non-Markovian tasks with long-term dependencies between relevant events (using RL-LSTMs)

Yanbei Chen 1 Xiatian Zhu 2 Shaogang Gong 1 Queen Mary University of London 1 Vision Semantics Ltd. 2 {yanbei.chen,s.gong}@qmul.ac.uk eddy@visionsemantics.com Abstract. We consider the semi-supervised multi-class classification prob- lem of learning from sparse labelled and abundant unlabelled training data. To address this problem, existing semi-supervised deep learning methods often rely on the up-to-date “network-in-training” to formu- late the semi-supervised learning objective. This ignores both the dis- criminative feature representation and the model inference uncertainty revealed by the network in the preceding learning iterations, referred to as the memory of model learning. In this work, we propose a novel Memory-Assisted Deep Neural Network (MA-DNN) capable of exploiting the memory of model learning to enable semi-supervised learning. Specif- ically, we introduce a memory mechanism into the network training pro- cess as an assimilation-accommodation interaction between the network and an external memory module. Experiments demonstrate the advan- tages of the proposed MA-DNN model over the state-of-the-art semi- supervised deep learning methods on three image classification bench- mark datasets: SVHN, CIFAR10, and CIFAR100.

pathway was demonstrated to be upstream of PKC-ζ, one of KIBRA’s binding partners. Furthermore, this pathway has been implicated in key neurobiological functions such as neurite outgrowth and other processes that underlie cognitive function. Based upon this knowledge, Huentelman et al. (2009) inhibited the RhoA/ROCK/Rac pathway in a rodent model using hydroxyfasudil to determine if enhanced memory and learning would ensue. Both spatial learning and working memory were improved in aged rats, implicating ROCK activity in these processes. Huentelman et al. (2009) state that one plausible mechanism may be phosphorylation of KIBRA by PKC-ζ after activation of Rac-1. However, this hypothesis must be confirmed by further experimentation. If substantiated, this drug may have clinical relevance as a cognitive enhancer, for the parent drug (Fasudil) is safe and well-tolerated by humans – it is currently used to treat stroke victims. More research must be done, however, in order to determine if only KIBRA is altered and not the molecules it interacts with.

The experiments in this paper show that op- timi~tion of memory use in memory-based learning while preserving generalisation accuracy can only be performed by (i) replacing instan[r]

or reducing explicit memory demands, would be associated with more accu- rate retrieval and execution of that intention than intentions encoded with errors. Crucially, the prediction was that this would be the case despite the intention being adequately retained (in the sense of being accurately reported post-test) in both conditions. A second prediction concerned a differential effect of encoding condition on EBPM and TBPM tasks. Based on the model outlined above, EBPM tasks provide strong environmental cues linked to the stored intention. Improvements in the encoding of the intention after EL should allow the participants to recognise these cues and recall the associated intended action more reliably than after EF learning. TBPM tasks also require adequate retention of the intention but do not provide environmental cues and so require additional monitoring and self-initiated action. If EL enhances encoding of the intention but does not improve moni- toring capacity, a less prominent or even absent EL advantage would be expected.

2 Department of Life Sciences, Graduate School of Arts and Sciences, The University of Tokyo, Tokyo, Japan Email: mingming@10.alumni.u-tokyo.ac.jp Received February 23 rd , 2012; revised April 26 th , 2012; accepted May 29 th , 2012 To investigate the enhancing effect of post-learning stress on memory, we requested 38 Japanese under- graduates to perform a learning task that involved positive, negative, and neutral words with controlled arousal and subsequently assigned them to a stress group (exposed to acute white noise) or a control group. After a 10-min filler task, we administered a delayed free recall test and a recognition test. We found that exposure to acute stress after learning significantly enhanced recognition memory of words, but found no differences in memory scores for stimuli of varying valence. We accordingly propose that post-learning stress, though enhancing memory performance, may not depend on word valence when stimulus arousal is controlled. This is the first study to find that post-learning stress enhances memory af- ter a short delay, and it has several implications with regard to traumatic memories in stress-related dis- orders.