2.1 Synaptic Modifications
Non-associative plasticity
Habituation & Sensitization
Non-associative learning is when a behaviour towards the same stimulus changes over time, without any link to outcome or any other stimuli. In non-associative learning, the properties of a single stimulus are learned. This type of learning is based on the frequency of the occurrence of the stimulus. There are two main types of non-associative learning: habituation and sensitization. Habituation occurs when there is a decreased behavioral response to a harmless stimulus that is repeatedly presented (Fig 1). Sensitization, on the other hand, occurs when there is an increased response to various stimuli following the presentation of a noxious stimulus.
An everyday example of this is when watching fireworks on a holiday. When you hear the first firework you may get startled, but as you hear more and more fireworks, you are no longer startled as you get habituated to the sound. After your startle response has weakened, if your friend were to pinch you, the next time you hear a firework, your startle response would be restored and you would show the same level of startle as you did to the first firework you heard. The pinch has caused you to become sensitized to an unrelated stimuli, the sound of the firework. (Kandel textbook)
Habituation & Dishabituation
In a neural circuit, a stimulus can act on sensory neurons in an organism and subsequently elicit a response in the form of behavioural output (Fig 2). Habituation occurs when repeated presentations of the same stimulus cause an output response that becomes weaker each presentation (Fig 3). The first response’s magnitude can be considered the baseline level of response, and each response thereafter is weaker than the baseline response. Dishabition occurs when the output response returns back to the previous baseline level, usually after a stimulus free time interval (Fig 3). Both habituation and dishabituation have been extensively studied in the sea slug Aplysia californica.
Habituation experiments
The gill withdrawal reflex (GWR) in the aplysia is a well-studied example of habituation. This organism was chosen due to its large neurons and relatively simple nervous system. From an evolutionary perspective, the gill withdrawal reflex is a defensive reflex that causes the delicate siphon and gill to become retracted when the animal is disturbed, thereby protecting it from potential threats (Fig 4).
The circuit involved is relatively simple. There is a siphon, which is used to expel seawater, that when a tactile stimulus comes into contact with, elicits the gill-withdrawal reflex. Mechanoreceptors in the abdominal ganglion innervate the siphon skin. These excitatory sensory neurons form synapses with the motor neurons that innervate the gill as well as on both inhibitory and excitatory interneurons that also synapse on the motor neurons. Repeated tactile stimulation of the siphon leads to a reduced response in the GWR (Fig 5).
This phenomenon has also been studied in rodents using the odour habituation test. In this task, a mouse is presented with one specific odour, where the amount of sniffing time directed at the odour is measured (Fig 6). The first time a mouse smells a particular odour, it will investigate it for a long period of time because it is curious about this novel smell. However, after repeatedly presenting the mouse with the same odour, the mouse will begin to sniff at the odour less and less, as it starts to become habituated to that odour. When the mouse is presented with a new odour, it will now become dishabituated and will begin sniffing longer towards this odor. Again, after multiple presentations, the mouse will become habituated to the new odour as well. This has also been studied in humans, where the human salivary response becomes habituated when participants are given the same flavour of juice, but becomes dishbautated when a different flavour of juice is given.
Aplysia Recordings
Since this is a relatively simple circuit from input to output, there are only a few potential synapses where this change could occur. The change could occur in the sensory neuron, in the motor neuron or at the gill itself (Fig 7). One potential way to find out where the changes happen is to record electrical activity from the potential areas of interest, and compare the electrical activity to the behavioural response. The magnitude of the behavioural response, in this case the GWR, should parallel the electrical activity of where the change is happening.
Stimulating the siphon while recording electrical activity of the sensory neuron and motor neuron and observing the gill response helps shed light on the mechanism. Repeated stimulation of the siphon elicits an action potential in the sensory neuron, where the action potential magnitude is the same each time the siphon is stimulated, even though the gill is moving less each time. This suggests the change underlying habituation of the GWR doesn’t seem to be related to the sensory neuron’s input changing. Recording electrical activity in the motor neuron shows that with repeated stimulation of the siphon, there is less electrical activity in the motor neuron with each stimulus. This reduction in EPSP amplitude seems to correlate quite well with the gill withdrawal. When the EPSP is high in the motor neuron, the gill withdraws a lot and when the motor neuron’s EPSP is smaller, the gill withdraws less. This suggests that the change is happening somewhere after the sensory neuron fires but before the motor neuron does, since the change occurs in EPSP magnitude. The potential synapse involved therefore seems to be the sensory neuron to motor neuron.
Quantal analysis has shown that the mechanism behind this habituation is due to presynaptic depression, where less neurotransmitter is released from the presynaptic neuron for a given action potential (Fig 8a). Over the course of habituation, calcium channels are modified to allow less influx of calcium in response to an action potential. This results in less neurotransmitter being released from the presynaptic sensory neuron and therefore less depolarization of the postsynaptic membrane in the motor neuron, thereby leading to the weakened GWR (Fig 8b).
Short-term & Long-term Habituation
Habituation can be either short-lived or long-lasting, and can depend on both the frequency and intensity of the stimulus. Short-term habituation that lasts from seconds to minutes seems to be due to presynaptic depression. Long-term habituation, which generally lasts longer than 10 hours, cannot be accounted for by only presynaptic depression (Fig 9). Long-term habituation is both RNA and protein synthesis dependent and occurs by changes in both the presynaptic and postsynaptic neurons. These changes can include: a reduction in the number of presynaptic terminal varicosities, the number and area of active zones and number of presynaptic vesicles, as well as changes in postsynaptic AMPA/NMDA-type receptors and calcium signaling (Potential Figure x1).
SENSITIZATION
Sens/Desens
Sensitization is another form of non-associative learning involving a single noxious stimulus. Sensitization occurs when an animal encounters a harmful stimulus and the animal learns to respond more strongly to not only that stimulus but also other stimuli, including harmless ones (Potential Figure x2). In sensitization, applying a stimulus to one pathway can produce a change in the strength of another reflex pathway. Analogous to habituation, there is both short-term and long-term sensitization. A single aversive event can produce short-term habituation lasting only minutes, whereas multiple aversive events in succession can produce long-term sensitization that lasts from days to weeks.
In Aplysia, the previously habituated GWR response becomes very strong after a single electric shock to its tail (Fig 10). The noxious stimulus to the tail causes an enhancement of synaptic transmission in several areas of the neural circuit that support the GWR. The fact that this occurs in the same synapses that were depressed by habituation, instills the notion that a synapse can store more than one type of memory and can participate in multiple types of learning.
Just like in habituation, short-term sensitization involves transient changes in the amount of neurotransmitter released, whereas long-term sensitization involves larger synaptic reorganization. This increased neurotransmitter release leads to an increase in postsynaptic motor output for the same presynaptic stimulus. There are however differences in the cellular mechanisms that produce these synaptic changes in the GWR. Habituation is homosynaptic, where decreased activity in the sensory neurons leads to the observed decrease in synaptic strength of this reflex path. Sensitization of the GWR is heterosynaptic, where modulatory interneurons that are activated by stimulation of the tail lead to the enhancement in synaptic strength.
Sensitization Experiments
The most widely-studied interneurons involved in the GWR are the ones that release serotonin (5-HT) (Fig 11). These interneurons form synapses on the siphon sensory neurons, as well as axo-axonic connections with their presynaptic terminals. Presynaptic facilitation occurs when these facilitating interneurons enhance neurotransmitter release from the sensory neuron via those axo-axonic connections. After a single tail shock, 5-HT is released from the interneuron, which can then bind to serotonin receptors to induce cellular changes. 5-HT can either bind to two different receptors engaging either the Gs protein pathway or the Go protein pathway. When 5-HT binds to initiate the Gs pathway, the activity of adenylyl cyclase (AC) is increased. AC converts adenosine triphosphate (ATP) to cyclic adenosine monophosphate (cAMP), which increases the amount of cAMP present in the terminal of the sensory neuron. cAMP then binds to the regulatory subunit of cAMP-dependent protein kinase A (PKA), thereby activating it by exposing its active catalytic subunit. PKA can then act via three different pathways. In the first pathway, vesicular release is enhanced where more neurotransmitter vesicles are moved to the active zone so that they are ready to be released. The machinery involved in exocytotic release is also enhanced and becomes more efficient. In the second pathway, PKA phosphorylates K+ channels, which causes a decrease in K+ current that leads to an elongated action potential. This then causes an increase in the Ca2+ influx, which subsequently causes more neurotransmitter vesicles to be released. The third pathway causes L-Type Ca2+ channels to become active. In the second 5-HT signalling point, the Go protein pathway, phospholipase C (PLC) becomes active, which then via diacylglycerol and protein kinase C (PKC) can act on both L-type Ca Channels as well as on the available pool of neutramistter vesicles. All these pathways converge to contribute to presynaptic facilitation, enhancing neurotransmitter release from the sensory neuron in the short term.
Sensitization has also been studied in rodents in the context of pain thresholds, where rodents with tissue injury start to withdraw their paws upon the presentation of a non-noxious stimulus, where normally they would not. (Fig 12). Spinal cord dorsal horn neurons show increased responsiveness to a given stimulus following injury. Before the injury, paw stimulation outside of the injured zone would not cause the neuron to become active, but after the injury, the neuron now responds to stimulation in the non-injured area. After the injury, in the injured zone, the amount of paw stimulation needed to cause paw withdrawal is much lower than before the injury. The non-injured area too has become behaviourally sensitized, and also requires less paw stimulation to cause a paw withdrawal. Central sensitization has also been documented in humans, where tissue injury results in increased pain sensation to mildly noxious stimuli or pain to non-noxious stimuli.
Aplysia: Short vs Long Mechanisms
The number of synapses has been shown to increase in GWR sensitization experiments (Fig 13A). Changes in the sensory neuron have also been shown (Fig 13B).
Short-term changes occur via presynaptic facilitation leading to enhanced vesicular release from the sensory neuron (Fig14). Long-term changes are supported by synaptic growth and require the synthesis of new proteins. PKA translocates to the nucleus and phosphorylates cAMP-response element binding 1 (CREB-1). CREB-1 can then bind to CREB Response Element (CRE) to active gene transduction pathways. This then leads to the growth of new synapses and increased activity of PKA. When PKA is translated to the nucleus, it activates another pathway, the MAPK pathway, where MAPK then phosphorylates CREB-2, which is an inhibitory repressor of transcription. Phosphorylation leads to less suppression by CREB-2, further enhancing these synaptic changes. Short-term sensitization can involve the activation of silent synaptic terminals and long-term sensitization can involve new synapse formation.
NON-ASSOCIATIVE SUMMARY
In a normal scenario, before any changes. A sensory neuron might synapse on a motor neuron like this: where it forms two synaptic connections (Fig15). A stimulus acting on the sensory neuron will have a stereotyped magnitude of response, as long as the connections don’t change and assuming there are no short-term changes. This is the baseline response in the motor neuron output for a given sensory input.
In long-term habituation, where behaviourally there is a less pronounced response of the reflex, the relationship between the sensory and motor neuron might look something like this. Where now there is only one connection from sensory to motor neuron. Now the same stimulus can elicit a smaller response in the motor neuron, thereby driving behavioural habituation.
In long-term sensitization, where there is a more pronounced response of the reflex, the relationship might look like this: where new the sensory neuron forms many more connections, in this case four. Now the same stimulus input can elicit a larger response in the motor neuron, thereby driving the behavioural sensitization.