Psychology 325: Psychobiology

First Project

The Neural Basis for Long Term Syntactical Memory

Synaptic Mechanisms of Syntactic Memory: Gary Van Horn

Studies of syntactical memory and processing in the mammalian neocortex have found evidence linking Brodmann's areas 44, 45 (Broca's area), and 46 with the universal human ability to modify the meaning of an utterance based solely on the arrangement of the words (Gaulin, 2001, Vigliocco, 2000). For example without the syntactic processing in these areas one would not be able to appreciate the difference in meaning between the two sentences "Jim is in love with Harriet" and "Harriet is in love with Jim" (Kandel, 2000). Memory occurs in these parts of the neocortex, as it does in most areas of the mammalian brain, via the phenomenon of long term potentiation.

Long term potentiation (LPT) is, in essence, the net effect of changes in the excitability of a neuron in relation to the frequency of exposure to a stimulus (Carlson, 1998). In this case the specific stimulus is syntax, however the type of LPT that has been found in the areas associated with syntactical processing is very much like the LPT found throughout the mammalian brain. It is because of this similarity that I will be talking about LPT in a more general sense rather than pointing to specific instances directly related to syntactic processing or learning.

Although the occurrence of LPT is generally referred to as a singular phenomenon there are many ways of achieving the overall goal of increasing the level of stimulation that the post synaptic cell receives. One well documented way which is regulated by a change in the synaptic mechanisms is an increase in the number of postsynaptic AMPA receptors (Tocco et al. 1992). A synapse that is gated by only NMDA is less likely to cause an excitatory post synaptic potential (EPSP) because the ion channel is blocked by a magnesium ion (Mg2+) at normal resting potentials (~ -70mV) which cannot be removed by the activation of the glutamic acid (glutamate) receptor alone. Essentially a synapse which relied solely on NMDA receptors would not produce an EPSP unless it had been primed by the effects of an EPSP of a nearby synapse, which would eject the magnesium ion, or had been depolarized in some other way, such as input from a second afferent pathway or by a stimulus "of sufficient intensity to activate voltage- dependent calcium channels" (Shors, Matzel 1997). Thus the dendrite of the post synaptic neuron gains sensitivity by including more AMPA receptors, which do not require a change in the polarity of the cell to occur before calcium ions (Ca2+) are allowed to enter the cell.

A schematic illustration of an NMDA receptor, with its binding sites (Carlson, 1998)

Neural Basis of Language: John Meyer

Within the neural structure of the human brain there appear to be specialized areas, represented in many regions, for storage of syntactical /semantical information . In the following, I will first trace a basic path through key language areas and provide the reader with current data in regards to long term syntactic memory; then, I will describe how some of these neural structures develop during the critical language period of early childhood and provide data for this area as well . The information I have gathered is by no means exhaustive; however, it provides a glimpse of how the two complex tasks language and memory develop and interact on a neurological level.

Neural Systems of Syntactical Language: John Meyer

As you read these words, the verbal information is processed through the primary visual cortex to the left hemisphere, extrastriate cortex. A specific region in this cortex , which includes fusiform and lingual gyri, has been shown to be involved in recognizing combinations of letters (Carlson,1998). The information is also routed to the middle and posterior portion of the superior temporal gyrus, known as Wernicke's area. This is, in the classical sense, the area that contains the memories of word sounds. It lies adjacent to the posterior language area, an area that has access to memories of the associative meanings behind the words ; and it is by interaction between Wernicke's area and the posterior language area that symbols such as words can be connected to meaningful associations and categories(Carlson, 1998).

The posterior language area is a junction of the Temporal, occipital/parietal cortices; and in these cortices are where some of our "long term representations of concepts " might be stored (Mazziotta & Toga 2000). In a positron emission tomography (PET) study by Chertkow et al.(1993), presentation of concrete nouns, such as animals or tools,etc., activated the temporooccipital junction. And more recently, a similar study by Lepage, et al (2000) observed, "activation of the medial temporal lobe including the hippocampal region during semantic associative encoding ( relating words to objects)". Left inferior prefrontal cortex activation, beyond Broca's area, was observed as well which implies a participation of the temporal and frontal cortex, though the activity of the frontal area may be "functionally different" than the temporal lobes duties of word/object association (2000) .

Frontal and temporal lobes are connected by a bundle of nerve fibers called the arcuate fascicilus, and it has been hypothesized by Wagner et al. that left prefrontal and temporal regions both work to promote memory formation for words. In a study on long range EEG synchronization during word encoding, Weiss and Rapplesberger (2000) state in their conclusion, "for the first time, an increase interaction of electrical brain activity between cell assemblies of frontal and temporal/parietal brain areas during efficient verbal memory encoding was shown.... This is the first hint that synchronization of electrical activity [coherence] between distant brain regions enables efficient verbal memory encoding" (2000). Whether the frontal lobe is involved in storage or retrieval (or both) of words is unknown. Though we can say from this observation that interaction is occurring during encoding , further investigation is called for before we can understand the exact nature of this interrelationship.

Synaptic Specializations for Syntactical Memory: Nicole Krupnick

The Hebb rule is confirmed by hippocampal formation. (Carlson, 2001) In the hippocampal formation, long-term potentiation can be triggered by electrical stimulation. The part of the hippocampal formation where this takes place is in the axons leading from the entorhinal cortex to the dentate gyrus.

Carlson gives background of anatomy in his text, so the author of this paper will also provide an opportunity to view the components necessary for long-term potentiation. The hippocampal formation is located in the limbic system. The hippocampal formation is made up of the hippocampus, the dentate gyrus, and the subicular complex. Major neocortical inputs and outputs of the hippocampal formation are channeled through the entorhinal cortex. From here, neurons pass incoming information to the granule cells of the dentate gyrus through axons called the perforant path. These neurons send axons into field CA3 in the hippocampus. (Carlson, 2001)

The focus of this section will be to examine dendritic spines and their purpose in long-term potentiation. In the field CA3, the dendritic spines on pyramidal cells make synapse on fibers from the dentate gyrus. The pyramidal cells have an axon that grows downward, and a long thick dendritic trunk grows out of the top. (Carlson, 2001) On the dendrite and branches, there are approximately 30,000 dendritic spines responsible for long-term potentiation. (Carlson, 2001)

The axons from the pyramidal cells in CA3 branch off in two different directions. One branch forms synapses with dendritic spines on other pyramidal cells. The other branch goes into the basal forebrain where the septum and mamillary bodies are, taking a path through the fornix. (Carlson, 2001) According to Carlson, "long- term potentiation can be induced by stimulating the axons in the perforant path with a burst of approximately one hundred pulses of electrical stimulation, delivered within a few seconds."

Researchers at the Max Planck Institute of Neurobiology, in Munich, Germany, determined that long-term potentiation in a hippocampal neuron goes along with morphological changes in dendritic spines. They have shown that spines do change their morphology when the synapses are strengthened. (Bonhoeffer, 1999) They have done this with the aid of two different techniques. One way is the use of a two-photon laser- scanning microscopy. This helps to visualize microscopic objects that are stained with a fluorescent dye. The other method they used is local microsuperfusion. This research helped to confirm earlier studies that had tried to find the changes in the spines, yet had not known exactly where they needed to be looking in the spines.(Bonhoeffer, 1999)

Dendritic spines have been of much interest to neurobiologists over the years. These structures, and the synapses placed on them, and their synaptic plasticity are thought to be a structural basis for learning and memory. (Spacek, Harris, 1997). The research by Spacek and Harris used rats to study dendritic spines. In their research they discuss endoplasmic reticulum in dendritic spines. A smooth endoplasmic reticulum is associated with synapses. In parent dendrites, the amount of endoplasmic reticulum is the same as the number of spines and synapses along their lengths. On the small spines, there was little or no endoplasmic reticulum. Spacek and Harris's study states, "The large spines contain a cisternal spine apparatus derived from reticulum and directed to the extensive 'perforated' synapse."

There are different types synaptic specializations that occur on dendrites. The following list was adapted from Fiala and Harris, 1999. The first pattern is varicosity. It is an enlargement in a thinner dendrite that is associated with synaptic contacts. An example is retinal amacrine cells. The second pattern is Filopodium. It is a long, thin protrusion with a dense actin matrix, and very few internal organelles. It is usually only seen during development. The third is the simple spine known as sessile. They are synaptic protrusions without a neck constriction. It has a stubby spine, and a crook thorn. An example would be the pyramidal cells of the cortex, and the cerebellar dentate nucleus. Fourth is the pedunculated simple spine. It has a bulbous enlargement at the tip. It has a thin spine, a mushroom spine, and a gemmule. Examples are the pyramidal cells of the cortex, and the olfactory bulb granule cell. Fifth is the branched spine. Each of these have a unique presynapted partner, and each has the characteristics of a simple spine. Examples of these are CA1 pyramidal cells, the granule cells of the dentate gyrus, and the cerebellar Purkinje cells. The sixth is the claw ending. It has synaptic protrusions at the tip of the dendrite associated with one or more glomeruli. Examples of these are granule cells of the cerebellar cortex, and the dorsal cochlear nucleus. The seventh is the brush ending. It is a burst of complex dendritic protrusions at the end of the dendrite that extends into glomerulus and contains presynaptic elements. Examples are unipolar brush cells of the cerebellar cortex and the dorsal cochlear nucleus. The eighth is the thorny excrescence. It is densely lobed dendritic protrusion into a glomerulus. Examples of this are proximal dendrites of the CA3 pyramidal cells, and dentate gyrus mossy cells. The ninth is the racemose appendage. They are twig-like branched dendritic appendages that contain synaptic varicosities and bulbous tips. Examples of these are the inferior olive, and the relay cells of the lateral geniculate nucleus. Tenth and last, is the coralline excresecence. These are dendritic varicosity extending numerous thin protrusions, and velamentous expansions and tendrils. Examples of these include, the cerebellar dentate nucleus, and the lateral vestibular nucleus. Again, this was adapted from Fiala and Harris's work from 1999.

Dendrites are not very large. They're usually not more then a millimeter long and many times shorter then that. Out of all the different synaptic specializations, the spiny dendrites are the most common. These are less then a micron. The spiny dendrites are most frequent on the cells in the brain region. For those cells, ninety percent of their excitatory synapses take place on the spines of the dendrites, which is why they are believed to be an important part of learning and memory.

Neural Development Related to Syntactical Memory: John Meyer

In the second year of human development we have what is known as the "language explosion". In this period, the number of words understood and produced quadruples (Herschkowitz, 2000). Increased interconnectedness and efficiency occurs in the network the prefrontal cortex, language areas, hippocampus, cerebellum, and basal ganglia. There is an increase in the number of dendritic spines in the prefrontal cortex. And in Wernicke's area there is "intensive dendritic elongation" and an increase in density , forming a neural base for language (Herschkowitz, 2000). In the hippocampus, which, as we know is heavily involved in learning and memory, there is "intensive differentiation" in the complex dendritic spines (part of Long term potentiation) of the CA-3 region of Ammon's horn( 2000). These processes lead to increased long term potenciation, the neural basis for learning and memory. Studies have shown , when the hippocampus is not functioning well, long term memories (such as, the associations accessed by language area) cannot be consolidated into the neocortical areas (Calvin, 2000).

Concurrently, dedrites of the pyramidal neurons in layer III elongate and reach to layer IV, creating a stronger connection between these layers. GABA ergic inhibitory interneurons extend their dendrites and increased arborization. It is from Layer III that callosal and commissural axons link prefrontal cortices as well as ipsalateral areas within each hemisphere. Layer IV receives axons from the mediodorsal nucleus (MD); the MD relays information to the prefrontal from the associational cortices, reticular and limbic systems (where hippocampus resides), and basal ganglia and cerebellum (Herschkowitz, 2000). Connections like these implicate the relationship of memory areas of the limbic system, such as the hippocampus, with language areas of the prefrontal and associational cortices.

In concluding, knowledge of the neural structure and development of syntactic memory seems to be rapidly accumulating. We now know that areas involved in verbal memory go beyond classic language area and Wernicke's area. Brain mapping for language activity in the frontal lobe is gaining interest . As data accumulates on the prefrontal/temporal connection, Questions arise, such as: if nouns(objects) are primarily stored in the association cortex, is the prefrontal cortex involved in the storage and retrieval of verbs(actions)? And is it by way of the arcuate fasciculus that these retrieved words (and associations) are combined to produce coherent speech? Perhaps a replication of a semantic retrieval study focusing on verbs (condition 1) and/or nouns(condition 2) would be insightful. It would also be interesting to find if there are sites within the prefrontal lobe that are heavily involved in long term potenciation, as well as encoding, thereby implicating such a site for possible long term storage of encoded verbal material.

Intracellular Mechanisms of Syntactical Memory: Gary Van Horn

Long term potentiation (LTP), which can occur due to more overt physical changes of the pre and post synaptic cells, is often carried out on an intracellular level at the same time. In fact many of the more visible changes such as a difference in the number of AMPA receptors, or changes in the shape of the dendrite, can be attributed to the underlying intracellular mechanisms that are active during learning.

One example of such a change being led by intracellular mechanisms is the postsynaptic thickening of neurons after LTP has taken place. According to research it would seem that type II calcium-calmodulin kinase plays an especially important role in the phosphorylation of the cytoskeletal proteins, along with a host of other possible kinases; such as cAMP-dependent protein kinase A (PKA), protein kinase C (PKC), and tyrosine kinases, phosphatases such as protein phosphatase-1 (PP-1), PP-2A, and PP-2B (calcineurin), and phosphatase inhibitors such as inhibitor-1, inhibitor-2, and DARPP-32 (Wagner et al, 1991). Following from that example of intracellular mechanisms providing for long term memory storage is the believed effect of nitrous oxide on soluble guanylyl cyclase. As shown in the diagram the influx of calcium ions into the dendritic spine spurns the synthesis of nitrous oxide via NO synthase, then the NO is released from the cell where it can be absorbed by the terminal bouton and used by soluble guanylyl cyclase "an enzyme found in the cytoplasm that triggers the synthesis of cyclic GMP" (Carlson, 1998). The cyclic GMP then catalyzes chemical reactions which increases the release of glutamate into the synapse, which seems to significantly increase the likelihood that LPT will take place.

References

Gaulin, Steven J.C., and McBurney, Donald H.(2001) Psychology: an evolutionary approach Prentice-Hall, Inc.

Vigliocco, Gabriella (2000) The Anatomy and Meaning of Syntax Current Biology, Vol 10:R78- R80

Kandel, Eric R., Shwartz, James H., and Jessell, Thomas M. (2000) Principles of Neural Science, Fourth Edition McGraw-Hill

Carlson, Neil R. (1998) Physiology of Behavior, Sixth Edition Allyn and Bacon

Tocco, G., Maren, S., Shors, T.J., Baudry, M., and Thompson, R.F. (1992) Long -term Potentiation is Associated with Increased[3H]AMPA Binding in Rat Hippocampus Brain Research, Vol 573, 228-234

Shors, Tracey J. and Matzel, Louis D.(1997) Long-term Potentiation: What's learning got to do with it? Behavioral and Brain Sciences Vol 20(4), 597-655

Wagner, K.R., Mei, L., and Huganir, R.L. (1991) Protein tyrosine kinases and phosphatases in the nervous system Current Opinion in Neurobiology, Vol 1, 65-73

Arthur W. Toga & John C. Mazziotta (2000) . Brain Mapping , The systems(pp.450-458) San Franciso: Academic Press.

Calvin, W.H. (2000) . Lingua Ex Machina(pp 27-940). Cambridge: The MIT Press

Herschkowitz , N. (2000, October 11) Neurological basis of behavior developement in infancy. 411-416

Lepage, M.(2000, december 99) Neural coorelates of semantic associative encoding in episodic memory. Cognitive Brain Research, 271-280.

Wagner, A.D., Semantic repetition priming for verbal and pictorial knowledge: a functional MRI study of the left inferior prefrontal cortex . J. Cogn. Neursci . 9 714-726

Weiss S. , Rapplesberger P.(2000, february, 7). Long range EEG synchronization during word encoding coorelates with successful word encoding coorelates with successful memory performance. Cogitive Brain Research:, 299-312.

Herzog, 3d models Available On-line: http://iesk.et.uni-magdeburg.de/~herzog/ rekspines.html

Fiala, Harris (1999) Some Types of Synaptic Specializations of Dendrites Synapse Web, Available On-line: http://synapses.bu.edu/anatomy/dendrite/tables/table3.stm

Bonheoeffer, Tobias (1999) Do Neurons Learn By Growing Thorns?: Emergence of Dendritic Spines is Associated With Long-Term Synaptic Plasticity Availible On-Line:http://www.eurekalert.org/releases/mpg-dnl043099.html

Spacek, J. and Harris, H.M. (1997) Three-Dimensional Organization of Smooth Endoplasmic Reticulum in Hippocampal CA1 Dendrites and Dendritic Spines of the Immature and Mature Rat Jounal of Neuroscience, 17:190-203