So, how exactly is our brain "storing" and "retrieving" information? The study of the molecular basis for human memory continues to be a fascinating topic in neurobiology. And still, much of it is not understood. However, in just the past three decades, science has seen intriguing advances in this area. The primary concept of interest here is that of neural synaptic plasticity. In other words, the ability of the nervous system to change, in structure or function. These "changes" of the synapse can include any number of things, such as increased number of receptors, decreased neurotransmitter release, appearance of different types of receptors, etc. Such observed changes are shown to vary based on past stimulus or "previous experience." And these changes have been shown to exist anywhere from milliseconds to years, thus creating some semblance of short and long term memory.
We will first consider the case of short term plasticity, and then use that to better understand long term plasticity, more commonly understood as an interplay between Long Term Potentiation (LTP) and Long Term Depression (LTD).
Before going further, if you'd like to review neuronal signaling in the human nervous system click here.
The complexity of the human or mammalian CNS makes it more appealing for scientists to first study synaptic plasticity in much simpler nervous systems. One of the first and most often cited experiments is that of Eric Kandel and colleagues at Columbia University using the marine mollusk Aplysia Californica. Having much fewer (few tens of thousands) and much larger (up to 1mm) neurons, the aplysia becomes an attractive candidate for neuro-electrophysiological study.
Kandel studied a form of behavioral plasticity called sensitization. This is a process where the pairing of a weak and strong stimulus makes the animal respond strongly to both. For instance, when the aplysia's protruding siphon is touched, it will contract its gill and withdraw the siphon into its body. When touched repeatedly, the aplysia's response habituates (it is able to "remember" the stimulus is a benign one) and it responds only slightly to this now weak stimulus. However, if touching the siphon is paired with a simultaneous strong electrical stimulus to the tail, then the siphon stimulus again elicits strong withdrawal (because it can "associate" the idea that a newly introduced stimulus can have an effect on another stimulus that used to be a weak one). So it is said that the weak and non-noxious stimulus (touching of the siphon) has become sensitized by the strong and noxious stimulus of the tail. The schematic below can explain how only a few different types of neurons can account for the simple behavior of gill withdrawal and plasticity during sensitization.
4. PKA phosphorylates several proteins, probably including K+ channels. The net effect of PKA is to reduce the probability that the K+ channels open during a presynaptic action potential. 5. This prolonged presynaptic action potential opens more presynatpic Ca 2+ channels. 6. Finally, increased influx of Ca 2+ into presynatpic terminals increases the amount of transmitter released into motor neurons during a sensory neuron action potential.
In summary, a cascade of signal transduction events that include neurotransmitters, second messengers, and ion channels work toward enhancing synaptic transmission between sensory and motor neurons in the gill, to facilitate the short-term sensitization of gill withdrawal.
Same mechanisms can mediate long-term sensitization as well, which involves changes in gene expression and protein synthesis. With repeated stimulus to the tail, the serotonin-activated cAMP-dependent PKA now phosphorylates and activates a transcriptional activator called CREB. CREB increases rate of transcription of downstream genes. Gene activation can increase the number of protein receptors at the synapse, or even the number of synapses between sensory and motor neurons.
Major Conclusions:
1. Behavioral plasticity can clearly arise from changes in efficacy in synaptic transmission.
2. These changes can be either short-term (reliant on post-transcriptional modifications of synapses or synaptic proteins), or long-term (requiring changes in gene-expression, new protein synthesis, and elimination/creation of synapses.)
This experiment and mechanism evidences that neural plasticity is indeed a reality, as synapses are seen as dynamic entities that strengthen and weaken according to changing circumstances. Work on smaller invertebrates such as the aplysia allows us to extend and generalize some of the underlying principles in neural plasticity to mammalian nervous systems.
Having gained optimism after observing plasticity in simple invertebrates, scientists were eager to test their findings on mammalian systems. By now, we have accumulated ample evidence to support the existence of synaptic plasticity in mammals. Two types of changes are well studied: Synaptic facilitation and synaptic depression.
1. Synaptic Facilitation - Allows the synapse to transiently strengthen as a response to two or more action potentials invading the pre-synaptic terminal in close succession. When so many action potentials come in rapidly, the concentration of calcium (which is released in a quantized amount with each incoming action potential) builds up. Increase in calcium concentrations allow more neurotransmitter to be released into the synapse. This process strengthens the overall synaptic connection between the pre- and post- synaptic terminals. Strengthening of a synapse simply refers to the closer association of two adjacent neurons; when one becomes active, the other is more likely to become active.
2. Synaptic Depression - This is a process in which repeated synaptic activity can lead to diminished synaptic transmission. After a series of repeated stimuli leading to synaptic facilitation, a point is reached where too much calcium has accumulated within the pre-synaptic terminal, to the point where the neurotransmitter release has reached its maximum efficiency and cannot be furthered. The neurotransmitter vesicles that fuse into the synapse and release the neurotransmitters become depleted and the strength of the synapse therefore declines until this pool of vesicles can be replenished. Synaptic depression could be the process by which the aplysia was able to habituate its response to the siphon touch.
We see therefore, that the efficacy of chemical synaptic transmission changes dynamically as a consequence of the recent history of synaptic activity. These two mechanisms alone, however, cannot explain long -term memory in humans or other manifestations of behavioral plasticity that exist for months or years. Therefore, researchers are en-route to unraveling the mystery of long-term memory by focusing on such hypotheses as Long-Term Potentiation (LTP) and Long-Term Depression (LTD).
Concepts of LTP and LDP are popularly accepted as the closest approximations to the underlying molecular events dictating long-term memory.
Long Term Potentiation - Roughly stated, LTP refers to a process whereby if two neurons are usually active together, the connection between them will be strengthened; over time, this means that activity in one neuron will tend to produce activity in the other neuron. LTP has been most thoroughly studied in the hippocampus of the mammalian brain, an area known to be critical in formation and/or retrieval of memory. A lot of work has been done at the junction where schaffer collaterals (axons of pyramidal cells in the anterior end of the hippocampus) synapse onto another set of pyramidal cells near the posterior end of the hippocampus. Brief and high frequency stimulus of the schaffer collaterals cause long-lasting increase in the post-synaptic depolarization levels, and further can lead to a strengthened synapse. LTP of the schaffer collateral synapse exhibits three properties that make it an attractive and suitable mechanism for information storage.
1. LTP is state-dependent: The degree of depolarization of the postsynaptic cell determines whether LTP occurs. If a single weak stimulus is paired with one that yields strong depolarization of the postsynaptic cell, and if the two stimuli are coincident, we observe LTP.
2. LTP has input specificity: LTP only occurs at activated synapses rather than at all synapses of a given cell. Without this condition, it would be difficult to selectively enhance particular inputs, as is the case in memory.
3. LTP is associative: Pairing of stimulus pathways can yield a strengthened synapse.
The best known mechanism to fit LTP properties involves NMDA type glutamate receptors. NMDA receptor channels open to induce LTP only when 1. glutamate is bound to NMDA, and 2. the post-synaptic cell is depolarized enough to expel the magnesium block in the NMDA channel. This "and" condition accounts for two properties of LTP, specificity and associativity. It is specific in the sense that only synaptic sites containing NMDA receptors will be activated upon strong stimulus. It is associative in the sense that a weakly stimulated input may only be able to release and bind glutamate to the NMDA receptor, but not be able to alleviate its magnesium block because of insufficient depolarization. But if neighboring inputs are strongly stimulated, then this "associative" depolarization is enough to relieve the magnesium block and open the NMDA channels for LTP.
After NMDA channels open, several signaling pathways can lead to LTP. What must happen is this: we first muse see a rise in calcium levels in the postsynaptic cell (as it comes in through NMDA channels). Afterward, several theories have been proposed, that explains how these calcium ions activate one or more calcium activated protein kinases in the postsynaptic neuron. At least two protein kinases have been implicated in LTP induction: Calcium/calmodulin-dependent protein kinase (CaMKII) and protein kinase C (PKC). CaMKII is the most abundant postsynaptic protein at schaffer collateral synapses, and seems to to have downstream targets in transcriptional modification of postsynaptic receptors. Other good candidates for LTP induction include mitogen-activated protein kinases (MAPK). These activated protein kinases could lead to a variety of changes, such as, a rapid addition of new receptors to the potentiated synapse or an increased current flow through receptors already present. Other theories suggest that LTP causes sustained increase in transmitter release at the presynaptic terminal. The ultimate avenue leading to sustained changes in the synapse has not yet been resolved, but it could surely be one or a combination of some of the proposals mentioned above.Long Term Depression - LTD is a weakening of synaptic transmission. The mechanism of LTD is seen to be different from one place to the next. LTD at the schaffer collaterals of the hippocampus is seen to also require activation of NMDA channels, occurring when calcium levels rise only minutely, triggering activation of calcium dependent phosphatases (instead of kinases). LTD at the purkinje neurons of the cerebellum act differently, using metabotropic glutamate receptors (instead of NMDA) to activate second messengers such as DAG and inositol trisphosphate (IP3). These eventually activate voltage gated calcium channels, which then alter AMPA receptors (via an unknown mechanism) to weaken the overall synapse.
In summary, studies of synaptic plasticity show that modifications at the molecular level, both short and long term, come as a consequence and reaction to previous events at the synapse. This retention of recent events and ability to distinguish between them is what we hope can explain our more complex behavioral manifestations relating to human memory. While significant advancements have been made in understanding the cellular and molecular bases of some forms of plasticity, we simply do not know how selectively changing synaptic strength encodes memories or other complex behavioral modifications in the mammalian brain.