Tomorrow I start the real journey. I will arrive in Iowa in the afternoon, unpack and begin to figure my way around a new city and a new life. I am excited for the opportunity to follow my dreams. However, because of this excitement and moving, I will be lazy and post a sweet paper on memory that I wrote earlier this year.
Does Long Term Potentiation (LTP) Equal Learning/Memory?
Learning is considered the acquisition and development of memories and modification or acquisition of behaviors (Sweatt, 2003). It is the product of experience and ranges from simple forms of learning such as habituation to more complex forms such as play. Learning does not stand alone as having its own mechanism, but instead works within the context of both memory and recall. Memory is the process through which learned information is stored while recall is the conscious or unconscious retrieval process through which this altered behavior is manifest (Sweatt, 2003). Current understanding of neurons and the central nervous system implies that the processes of learning and memory correspond to changes in the relationship between certain neurons in the brain (Martin, Grimwood, & Morris, 2000). This can be defined as synaptic plasticity. As certain neurons are used while learning a behavior, the experience and repetitive use of the neurons will lead to the strengthening or weakening of the connections between them and finally the establishment of the learned behavior. However, does the change in neuronal strength or structure actually equal memory? Could there be other mechanisms that are used in the acquisition of altered behavioral responses?
Long-term potentiation (LTP) is the most studied form of synaptic plasticity in the central nervous system (McEachern and Shaw, 1996). LTP has received extensive attention because it shares a number of characteristics with memory, in that they both have a rapid onset and long duration and are strengthened by repetition. In addition, the duration of LTP being correlated with the time course of forgetting (Diamond, Dunwiddie, & Rose, 1988). However, in the thirty years since the discovery of LTP, the mechanisms have not been resolved. Instead, only many intricacies have been uncovered. Although many mechanisms of LTP have been shown to be involved in different learning and memory paradigms, does LTP equal memory and learning? Investigating these questions requires an explanation of learning, LTP, models of learning outside of LTP and finally a discussion of techniques in future research that should be used to investigated whether LTP is the mechanism for learning.
Learning
Learning, the altered behavioral response due to an environmental stimulus is thought to be stored in the cortex for both explicit (declarative) and implicit (procedural) learning (Sweatt, 2003). There are many forms of learning, from more simple forms of learning including habituation to more complex forms of learning, including associative conditioning. These two forms of learning are some of the most studied and best understood. Habituation is the cessation of a response to a stimulus after repeated presentations of the stimulus (Sweatt, 2003). One example of habituation that is well studied is in the Aplysia. The gill of the Aplysia must be thin and have a large surface area because it is where oxygen exchange occurs and this makes it susceptible to damage, so the animal presents many reflexes in order to protect it (Byrne & Kandel, 1996). Typically a touch applied to the siphon every 3 minutes over a period over 4 hours will result in a gradual cessation of the withdrawal response; a phenomenon called habituation (Byrne & Kandel, 1996). Experimenters showed that depression of the synapse between the sensory and motor neurons was the main underlying cause of habituation. Additional experiments showed that this synaptic depression is caused by a reduction in the amount of neurotransmitter released from the presynaptic sensory neurons (Cohen, Kaplan, Kandel, & Hawkins, 1997). There is inactivation of post-synaptic calcium channels and there is a reduction in the number of presynaptic vesicles that are available to release neurotransmitter. The reduction of the calcium channels is regulated by secondary messenger cascades which causes the down-regulation in the expression of the calcium channels. Two other forms of learning associated with habituation were demonstrated in the Aplysia, including dishabituation and sensitization (Cohen et al., 1997). Dishabituation is the full recovery of the original strength of a habituated response after the presentation of a strong, novel stimulus (Sweatt, 2003). Sensitization is the increase in strength of any reflex to a level above its original strength that is caused by one or more strong stimuli other than the stimulus that usually evokes the stimulus. After an Aplysia has been habituated, if the animal receives a strong touch to the tail, the next touch on the siphon will cause the gill to withdraw. Aplysia can be sensitized to a strong touch to the tail. A tap to the siphon, lighter than that used to habituate the animal, will cause the gill to be strongly withdrawn (Cohen et al., 1997).
In associative conditioning, an animal forms an association between two different stimuli (Dayan et al., 2000). One stimulus, the unconditioned stimulus (UCS), normally evokes a reflex. The other stimulus, the conditioned stimulus (CS) usually evokes no response (Dayan et al., 2000). When the conditioned stimulus is presented shortly before the unconditioned stimulus in a series of trials, the animal eventually responds to the conditioned stimulus alone. After conditioning, the animal forms an association between the conditioned stimulus and the unconditioned stimulus. There are two competing theories about how this form of learning works. In stimulus-response theory, the organism learns to form an association between the CS and the UCS. In the opposing theory, called the stimulus-stimulus theory, it is suggested that a cognitive component is required to understand classical conditioning (Rescorla and Wagner, 1972). In relation to the Aplysia, the presentation of the CS, such as the shock would elicit the siphon withdrawal reaction, whereas in the stimulus-stimulus theory the Aplysia would withdraw its gill because the UCS is associated with the concept of the shock. It is through an understanding of the different types of memory, that one is able to understand the biological underpinnings of these mechanisms. Form is often related to function, by knowing what the form looks or acts like (the behavior of memory) we can investigate its functions (the biological processes) which build the form.
Long-Term Potentiation
Several areas of the brain play a part in the consolidation of memories, but the hippocampus has been recognized as playing a vital role. Many of the following experiments involved an exploration of how synaptic plasticity functions in the hippocampus. One of the least understood, but most interesting problems in neuroscience is the attempt to identify the mechanisms which underlie memory. In 1949, Hebb proposed a well-defined rule of synaptic plasticity. He proposed that coincident activity in two connected neurons leads to strengthening of their connection (Chen & Tonegawa, 1997). This specifically increases the probability that they will fire together in the future. Through experimentation, it has been found that activity-dependent synaptic plasticity plays a vital role in sculpting synaptic connections during development (Lynch, 2000). LTP was first identified by Lomo, who reported that a single, short test shock , following an initial period of conditioning test shocks to the perforant path, elicited a potentiated response in the dentate gyrus (Lynch, 2000). After a train of stimuli, an elevated response of EPSPs (as measured by the membrane potential) in the postsynaptic cell (dentate gyrus) were noted. Later, Bliss and Lomo reported in 1973 that trains of high-frequency stimulation to the rabbit perforant path caused a sustained increase in efficiency of synaptic transmission in the granule cells of the dentate gyrus (Lynch, 2000). Chemical synapses are not static. Postsynaptic potentials (PSPs) wax and wane, depending on the recent and long-term history of presynaptic activity. In some synapses PSPs increase during repetitive stimulation to many times the size of an isolated PSP. A gradual rise of PSP amplitude during stimulation is called potentiation (Zucker, 1989). It is these changes in the frequency of firing and the amount of firing that synapses receive or do not receive that causes the changes in synaptic arboration. Subsequentl, it is the changes in the arboration and the interconnections of the synapses which are hypothesized to represent memory. These discoveries have lead to the formation of the synaptic plasticity model (SPM) of learning and memory (Martin, Grimwood, & Morris, 2000). The formal hypothesis of the SPM states that “activity dependent synaptic plasticity is induced at appropriate synapses during memory formation, and is both necessary and sufficient for information storage underlying the type of memory mediated by the brain area in which that plasticity is observed” (Martin, Grimwood, & Morris, 2000, 650). A thorough evaluation of the SPM hypothesis requires experimentation that addresses both the necessity and sufficiency. However, the current shortfall in the model, lies in the lack of experimentation in establishing the sufficiency of synaptic plasticity to induce memory (Martin, Grimwood, & Morris, 2000).
As demonstrated by the experiments in relating LTP in the hippocampus, a majority of the work on the cellular basis for memory has focused almost exclusively on hippocampal LTP/LTD and the belief that these forms of synaptic plasticity are a substrate for memory (Stevens, 1998). However, LTP has been observed in other brain areas, for example, the amygdala. Fear conditioning has been demonstrated to elicit LTP in the amygdala. Most of the evidence on the issue boils down to the observation that blocking LTP/LTD also interferes with the learning of spatial tasks (Stevens, 1998). Before settling upon an answer to whether LTP is the mechanism of memory in the hippocampus or other brain areas, the understanding of LTP itself, needs to be addressed. As LTP is further investigated, more intricacies are discovered, so to be able to conclude the role of LTP in memory, the meaning/definition of LTP needs to be lain out. However, we may be able to conclude that LTP, as currently defined and understood, is a mechanism that is concurrent with memory and memory associated areas.
Other Experimental Models of Learning in Synaptic Changes
Although LTP has established itself as the leading model learning and memory, there are other models that may explain these phenomena, including kindling (McEachern and Shaw, 1996). In kindling, daily trains of sub-convulsive electrical stimuli delivered to brain cortical or limbic sites eventually culminate in behavioral and electrographic seizure expression. A short burst of action potentials (afterdischarge; AD) follows each stimulus, and with additional stimulations, there is a progression of both electrical and behavioral measures of seizure activity. As AD duration increases, the AD localization spreads from the original focus of activation to downstream synaptic connections (McEachern and Shaw, 1996). The kindling phenomenon is associated with long-lasting facilitation of synaptic transmission. With the spread of activation and the facilitation of transmission, the kindling effect leads to increased proliferation (neurogenesis) of neurons. This model of synaptic plasticity is now best known as the model for epilepsy, but it was also one of the first neuroplasticity phenomenon suggested in learning. While it may appear odd that the abnormal firing behavior of seizures could be similar to learning, it may be that they share similar mechanisms or cascades, just with different outcomes (McEachern & Shaw, 1996).
While kindling may not be a close representative of normal neural firing and activity, other models are more subtle in their attempt to explain plastic changes in the brain, including primed burst potentiation (PBP) and spike-timing dependent plasticity (STDP). Two prominent physiological features of the hippocampus, complex spike discharge and theta rhythm are effective in potentiating synapses (Diamond, Dunwiddie, & Rose, 1988). PBP and LTP have common mechanisms in that both are not additive when both paradigms are used to stimulate a synapse and both are blocked with NMDA antagonists (Diamond, Dunwiddie, & Rose, 1988). PBP consists of a pattern of stimulation, a single priming pulse followed by a high frequency burst of pulses (Diamond, Dunwiddie, & Rose, 1988). Since the mechanisms of PBP are endogenous to the hippocampus during learning, it may activate the mechanisms of memory formation. One aspect of PBP which is very important is the dependence on the temporal components of stimulation, and this effect may be relevant to hippocampal-dependent learning, as a model of associative memory. PBP and its temporal nature are similar in aspect to STDP.
STDP exemplifies the importance of the temporal order of pre/post spiking for synaptic modification in the nervous system (Mu, & Poo, 2006). The synaptic strength can thus be bidirectionally modified by correlated pre/post spiking within a narrow time window, with pre-post spiking leading to LTP like modifications and post-pre spiking leading to LTD like modifications (Mu, & Poo, 2006). However, these temporal windows can change in both the time frame that the spiking has to occur within, and also which modifications occur according to the inputs. With the LTP and LTD consequences associated with the STDP, it may be assumed that STDP is simply these two phenomena on the level of individual neurons, however, STDP may serve to be a more parsimonious explanation as the mechanism for memory. The temporal nature of STDP allows for associations to be established. By pairing the modulation of STDP by inhibitory inputs and complex spike trains, STDP appears to use more complex mechanisms to store the differences in cell communication and thus more accurately represent the messages of the cells.
Examples that Challenge the Definition of Learning
While research often has a very anthropomorphic centered idealism, especially concerning research in the brain, it is ironic that much of what is known about the nervous system and the mechanisms of the brain comes from research with invertebrates. As discussed earlier, much of the initial work with LTP was demonstrated in the Aplysia, however, experiments with LTP use many different animal models today. Higher cognitive processes such as more complex forms of learning are thought to be demonstrated in “higher” species. This bias flows throughout the biological kingdom in varying degrees, and is applied to different organisms. Many have thought that it takes a nervous system to be able to learn. However, if the strictest definitions of learning are used, learning has been demonstrated in two species that many would have denied the ability to learn, the paramecium and the venus flytrap.
Learning in the Paramecium
While certain forms of learning are thought to be processes that requires large neuronal development and higher areas of brain functioning, it can also be seen in the single-celled organism, the paramecium. The paramecium does not have any neurons, so the question becomes, how can it learn? The question of whether paramecia exhibit learning has been the object of many experiments which yielded few answers to it. In a recent experiment by Armus et al. (2006), discrimination learning was demonstrated in the paramecium. Prior investigations into the possibility of classical conditioning in paramecium have shown different results. French (1940) reported that the time it took a paramecium to escape a tube decreased with the number of trials and this result was later confirmed by Huber et al. (1974) cited in Armus et al., 2006. Another approach attempted to condition paramecium to stimuli such as brightness or vibration to electric shock or heat. This included experiments, included in a review by Armus (2006) by Bramstedt (1935) and Soest (1937) who reported successful conditioning, while experiments by Best (1954) and Mirsky and Katz (1958) did not.
In the experiment by Armus et al., 288 paramecia (P. caudatum) were trained in a 22 mm long transparent glass trough which was half light and half dark. There was a stainless steel wire electrode that projected 2mm into the water at each end and the shock was provided by a 6.5 V DC with a duration of 60 ms. The subjects were observed through a microscope under 10X magnification. The paramecium were then collected individually from the colony and subjected to 7 periods of training and 3 periods of test. Three groups of 96 subjects differed in their training procedures, including, the experimental group which received a train of shocks when the paramecium was in the cathode half of the trough, the no shock control, and the paired shock control which received a train of shocks whether it was in the anode or cathode half of the trough (Armus et al., 2006). The paramecium are attracted to electrical shocks so learning would be demonstrated by the paramecium’s ability to discriminate between the side on which it received the shocks and the side in which it did not. The paramecium were supposed to associate the level of brightness of the trough with the shock. The results of the experiment showed that the experimental group spent significantly more time in the portion of the trough which was giving the shocks than the control group. However, there may have been confounding factors in this experiment. The shocks may have changed the liquid itself and made that more attractive to the paramecium. The authors addressed this with a second experiment. The brightness level from the training session was switched for the test session so as to eliminate any possibility that some substance that was hypothetically being produced by the shocks wasn’t the thing attracting the paramecium. The experimental group again spent more time in the shock part of the trough than the control group, demonstrating associative learning by the paramecium (Armus et al., 2006).
The results of this and other associated experiments bring into question our current description and understanding of learning. Is learning at its most basic level the change in behavior in response to time? However, where does the paramecium store its past behaviors, how does it associate the past to the present, is it a simple interaction of the molecules in the paramecium to the environment of the trough or is this organism, learning? This raises important points to consider. Can learning or memory be stored in many different ways? Memory may simply be the correlate of an increase in certain biological substances which in certain concentrations cause different reactions to occur. The paramecium may represent the learning or memory storage of one of our brain cells. A stimulus causes a change in the cells biological chemistry which then causes it to “remember” a situation, which could be viewed as either a positive or negative experience. In the view of a human, with many neurons and neural networks, memory and learning may represent the final outcome of a series of checks and balances representing all of the intricacies of an individual situation or experience through positive and negative gateways. It is interesting to consider the ability of organisms to “remember” and learn when none of the usually proposed manners of memory or learning are present.
Learning in the Venus Flytrap
Using the definition of learning as change in an animal’s behavioral response as a result of a unique environmental stimulus allows various nonassociative forms of learning such as sensitization to be included as learning (Sweatt, 2003). This may raise the question of whether sensitization in the venus flytrap plant is learning. The triggering mechanism for closure of the trap was two lobes with their surfaces facing inward each with three trigger hairs (Sweatt, 2003). To eliminate “false alarms,” the plant has evolved a mechanism whereby stimulation of a single trigger hair is insufficient to cause the trap to close (Sweatt, 2003). Two hairs must be stimulated in succession (or simultaneously) to trigger a trapping response. In one circumstance, stimulating a particular trigger will give no response, whereas stimulating the same trigger in another will, showing an altered response due to particular environmental stimuli (Sweatt, 2003). In a study by Volkov et al. (2008), the kinetics and mechanism of the trap closing were studied. Transmission of a single electric charge (mean 13.63 mC) caused the plant to close. A summation of stimuli is demonstrated through the application of smaller charges, if two or more consecutive charges were applied within a period of less than 50s the trap closed when the total of 14 mC was reached (Volkov, 2008). This cumulative character of the electrical stimuli indicates the existence of electrical memory in the plant. This form of learning may be comparable to PBP or STDP. The small background firing of the cell is interrupted by the sudden firing of one of the trigger hairs and this combination of firing patterns enough times strengthens the connections between the cells. Although the changes may not represent a long lasting change, enough stimulations may lead to a chemical change in the synapse which may make the “memories” longer lasting. This type of memory system may be analogous to planarians memory systems, which appear to be based in chemical changes of the entire organism.
The Future
While much of the work so far has been integral to bringing us to where we are today in our knowledge of memory and learning, there is a lot of work left to do. One of the major obstacles in the future remains identifying and establishing all of the phases of LTP, including the biological functions. However, while it will be interesting to discover all of the components and functions of LTP, is this the mechanism of memory and learning? While it may be integral to memory in the amygdala and the hippocampus, memories, appear to eventually be stored in the cortex. Are memories in the cortex stored by LTP, are they reactivated and reconsolidated through LTP or some other mechanism? While research may be years away from understanding LTP, memory and learning, there are many functions and mechanisms that can be further investigated. Future research should continue the use of genetic manipulation and investigation. By manipulating the expression of genes and other proteins, one by one or in conjunction with a series of other genes or proteins, we can piece together the changing and varying effects have a better understanding of the changes from a biological view-point. The use of a combination of methods will also be very interesting and helpful. Combining, PET, fMRI, genetic manipulation and behavioral paradigms will give researchers information and data from so many different views that they will be able to associate how changes in genes, lead to increases in one part of the brain which cause certain behavioral changes. Researchers should also remember to investigate both cases where there is severely impaired memory, but also where there is superior memory. Cases such as S or other mnemonists should be investigated, looking at genes, their biological makeup and their brain activity during learning and recall. Since memory and learning are so complex, we may never have a full grasp on how it works in all situations, but as technology improves and knowledge increases, memory will be better understood.
Conclusions
While much of the research in memory and learning points to the importance and necessity of LTP for memory and learning, other evidence points to the possibility of other mechanisms. Much of the current research points to the hippocampus as the site of memory consolidation, however, it has also been demonstrated that long-term memories are not stored in the hippocampus. If LTP/LTP were the only mechanism for the storage of memory, then memory and LTP should follow a linear relationship that when plotted against each other, shows memory increasing as LTP increases. However, as Sweatt (2003) demonstrates, there are many studies that show that this is not the case. While LTP may appear to be a very functionally sound mechanism for storing memories and also one that appears to fit with many people’s preconceptions of memory, in some cases it may not serve the best function. If LTP were to be the only mechanism for memory and the locations of memories were pinpointed, they could be easily erased with depotentiation of the synapse. However, while these discussions may last for a long time, it may come down to a person’s definition of the related components of the argument. To decide whether LTP is memory or learning or if LTP is just one of many mechanisms, it will take a better understanding of LTP itself, reconsolidation properties and the mechanisms which store memory in the cortex.
