Regulation Of Synaptic Efficacy Research Paper

1. Multiple Processes Regulate Synaptic Efficacy

The processes regulating synaptic efficacy vary along several dimensions, including (a) the nature of the change itself (e.g., an increase or decrease in efficacy); (b) mechanisms of induction —that is, the nature of the neural signals that initiate the synaptic change; (c) mechanisms of expression, the proximal causes of synaptic changes; (d) mechanisms of maintenance, the steps involved in their persistence; and (e) time course, ranging from transient forms of regulation lasting only a few seconds, to relatively permanent long-term changes persisting a day or more.

2. Simple Forms Of Synaptic Regulation

2.1 Decreases And Increases Of Synaptic Efficacy

Simple forms of synaptic regulation include decreases and increases in synaptic efficacy. Synaptic depression is typically expressed as a decrease in the amplitude or other parameters of a postsynaptic potential (PSP) at a given synaptic connection (Fig. 1). Other things being equal, the larger the synaptic potential, the greater the influence a presynaptic neuron exerts on its postsynaptic target neuron. Hence, in synaptic depression, the presynaptic neuron exerts less control over its postsynaptic target. Facilitation, enhancement, augmentation, and potentiation all refer to an increase in synaptic strength, often expressed as an increase in PSP amplitude (Fig. 1).

Synaptic regulation has been studied most extensively at chemical synapses. However, changes in the efficacy of electrical synapses can also occur.

2.2 Mechanisms Of Induction

A separate but related issue is the type and locus of cellular activity necessary to induce, or trigger, the change in synaptic efficacy.

Homosynaptic changes arise from activity that occurs at the same (‘homo’) synapse that is modified. Thus, homosynaptic depression is a decrease in synaptic strength (‘depression’) that arises from activity at that synapse (hence, ‘homosynaptic’). In contrast, heterosynaptic changes arise from activity in neurons other than those being modified. Both are instances of non-associative induction mechanisms, in that only one type or locus of neural activity is involved in triggering the synaptic change.

In contrast, conjunctive or associative mechanisms of synaptic regulation depend on activity in more than one cell. Associate mechanisms are of particular interest from a theoretical perspective, because they involve a more advanced regulatory logic: changes in synaptic strength occur if and only if two conditions are met.

2.3 Regulation Of Synaptic Efficacy In The Aplysia Siphon-Withdrawal Circuit

The following four examples illustrate these different forms of induction as well as certain other aspects of synaptic regulation in the context of the gilland siphon-withdrawal circuit of the marine mollusk Aplysia (Fig. 2).

As detailed by Eric Kandel and colleagues (Kandel and Schwartz 1982, Kandel and Pittenger 1999), the Aplysia gilland siphon-withdrawal reflex to a tactile siphon stimulus is mediated in part by the LE siphon sensory neurons, which connect both directly and indirectly to LFS and other siphon and gill motor neurons. The strength of LE sensorimotor connections can be modified in several ways. Similar changes have also been found at connections of Aplysia tail mechanosensory neurons.

2.3.1 Homosynaptic Depression And Homosynaptic Facilitation. Low-frequency activation of LE siphon sensory neurons depresses transmitter release at sensorimotor connections and hence also depresses the postsynaptic response (Fig. 1; Castellucci and Kandel 1976). Functionally, homosynaptic depression contributes to behavioral habituation of the gill-and siphon-withdrawal reflex. The siphon is a small fleshy spout that expels seawater from the mantle cavity. A tactile stimulus to the siphon causes both the gill and the siphon to withdraw. After repeated taps of the siphon, however, the amplitude and duration of the reflex diminish. Habituation is an important and ubiquitous form of learning that enables an organism to screen out stimuli of little consequence.

In addition to homosynaptic depression, Aplysia sensorimotor connections exhibit post-tetanic potentiation (PTP) in response to high-frequency activation, which is due at least in part to homosynaptic facilitation (Clark and Kandel 1984).

2.3.2 Heterosynaptic Facilitation And Heterosynaptic Inhibition. In Aplysia siphon-withdrawal circuits, strong noxious stimuli (e.g., tail shock) activate a group of facilitatory interneurons (e.g., L29 cells) that synapse onto LE siphon sensory cells. Activation of facilitatory interneurons modifies sensory cells and increases their transmitter release via multiple mechanisms. Because this increase in synaptic efficacy is regulated by an extrinsic input, it is an example of heterosynaptic facilitation (Fig. 1; Castellucci and Kandel 1976). Behaviorally, presynaptic facilitation contributes to sensitization of the siphon-withdrawal reflex, increasing its amplitude and duration. With very strong stimuli, heterosynaptic inhibition can also be induced.

2.3.3 aCTIVITY-DEPENDENT ENHANCEMENT OF PRESYNAPTIC FACILITATION. In this associative mechanism (also termed activity-dependent neuromodulation and activity-dependent extrinsic modulation), activity in sensory neurons enhances the heterosynaptic presynaptic facilitation normally evoked by facilitatory interneurons (Hawkins et al. 1983). The facilitation is greater than can be accounted for by a simple combination of a homosynaptic increase (PTP arising from sensory neuron activity) and heterosynaptic facilitation (arising from activation of facilitatory interneurons) and thus is associative. This mechanism is believed to contribute to classical conditioning of the siphon-withdrawal reflex, in which a siphon tap conditioned stimulus (CS) that activates siphon sensory neurons is paired with a tail shock unconditioned stimulus (US) that activates neuromodulatory facilitatory interneurons. Such pairings result in greater synaptic facilitation and a greater increase in the siphon-withdrawal response than that produced by unpaired CS and US presentations.

2.3.4 Hebbian Facilitation. Hebbian facilitation (Hebb 1949) also contributes to increases in synaptic efficacy at LE-LFS connections (Glanzman 1995). The critical inductive signal for Hebbian facilitation is conjoint activity of the presynaptic sensory neuron and activity or large depolarization of the post- synaptic motor neuron. The associative properties presumably arise from a glutamate N-methyl-_D -aspartate (NMDA)-type receptor, which acts as a molecular ‘AND-gate’ that requires both neurotransmitter binding and large postsynaptic depolarization (t o remove Mg block of the channel) in order for Ca²⁺ influx to occur through the channel and trigger synaptic change. Behaviorally, Hebbian facilitation also contributes to classical conditioning of the reflex. The tactile siphon CS activates siphon sensory neurons, and the tail shock US activates and depolarizes siphon motor neurons, thereby providing the key conjunctive events.

2.4 Expression, Maintenance, And Time Course Of Synaptic Regulation

In principle, increases and decreases in synaptic efficacy can result from either a presynaptic expression mechanism (e.g., a decrease in transmitter release from the presynaptic cell), or a postsynaptic expression mechanism (e.g., a decrease in the sensitivity or number of transmitter receptors on the postsynaptic target cell), or both.

Although the mechanisms for expression and for maintenance and duration of synaptic regulation are logically and biologically separable, they are often functionally related and will be considered together here. Four broad categories of mechanisms allow the regulation of synaptic efficacy to span a wide range of temporal domains: (a) transient regulation, which is voltage- and Ca²⁺-dependent, lasting milliseconds to seconds; (b) short-term regulation, involving second- messenger mediated covalent modifications of preexisting proteins, typically lasting minutes to an hour; (c) intermediate-term regulation, dependent on protein synthesis, persisting several hours; and (d) long-term regulation, dependent on both RNA synthesis and protein synthesis, persisting for hours and days (Fig. 3).

2.4.1 Transient, Voltage- And Ca2+-Dependent Presynaptic Regulation. As demonstrated by studies at the squid giant synapse, which permits intracellular stimulation and recording from both presynaptic and postsynaptic elements, transmitter release is regulated by the voltage of the presynaptic terminal, which in turn influences the basal concentration of intracellular Ca²⁺, the ion responsible for transmitter release. Depolarization of the presynaptic terminal elevates intracellular Ca²⁺ concentrations and increases release, whereas hyperpolarization decreases intracellular Ca²⁺ concentrations and decreases release. Axo-axonic synapses (connections from one neuron onto presynaptic terminals of another) can depolarize or hyperpolarize presynaptic terminals, thereby regulating release and producing presynaptic facilitation or inhibition, respectively. This form of synaptic regulation is typically transient, as it is dependent on the continuation of the afferent axoaxonic input, so it allows a moment-to-moment gating of information flow. Homosy naptic spike activity can also influence residual Ca²⁺ levels and thus modify transmitter release relatively directly.

2.4.2 Short-Term Regulation Via Second-Messenger Mediated Covalent Modifications. Short-term regulation of synaptic efficacy, lasting minutes to approximately an hour, typically involves second-messenger cascades and the subsequent covalent modification of pre-existing proteins, often via kinase-dependent phosphorylation. For example, in short-term facilitation at Aplysia sensorimotor connections, facilitatory interneurons activate both the cAMP-dependent protein kinase (PKA) and protein kinase C (PKC) in sensory cells. In the cAMP-dependent cascade, the facilitatory transmitter (e.g., serotonin), the ‘first messenger,’ binds to a receptor coupled to G-protein that activates the enzyme adenylyl cyclase. Adenylyl cyclase activation converts adenosine triphosphate (ATP) to cyclic adenosine monophosphate (cAMP), the ‘second messenger.’ cAMP in turn activates PKA by binding to the regulatory subunits of the kinase, causing them to dissociate from the catalytic subunits. The freed catalytic subunits then phosphorylate the substrate proteins, invoking conformational and functional changes. Similarly, in the PKC cascade, serotonin activates phospholipase C via a separate G-protein coupled receptor, which in turn elevates the levels of the second messenger diacylglycerol and activates PKC. Together, these two second-messenger cascades produce a broadening of sensory neuron action potentials, enhanced Ca + influx, and an enhancement of spike-release coupling, thereby increasing transmitter release.

Second-messenger cascades modify a variety of substrate proteins, allowing for coordinated regulation of multiple steps involved in synaptic transmission. Because substrate phosphorylation persists even after the brief input from facilitatory interneurons terminates, synaptic facilitation is maintained beyond the presence of sensitizing stimulus itself. However, biochemical mechanisms for substrate dephosphorylation also exist. Hence, covalent modifications of substrate proteins by themselves are not well suited for more permanent maintenance of synaptic changes.

2.4.3 Intermediate-Term And Long-Term Regulation Of Synaptic Efficacy Involve Translation And Transcription. Intermediate-term (lasting hours) and long-term forms of synaptic regulation (lasting a day or more) typically require the recruitment of additional steps to those implicated in short-term regulation. The relationship among these different forms of cellular memory has been particularly well studied at Aplysia sensorimotor connections. Unlike short-term facilitation, certain forms of intermediate-term facilitation require the synthesis of new proteins (translation of mRNA to protein). Long-term facilitation involves not only protein synthesis, but also the modification of gene expression (transcription of DNA to RNA) (Montarolo et al. 1986). These forms of synaptic regulation are thus distinct on a mechanistic as well as phenomenological level.

With repeated sensitizing stimuli or their analogs (e.g., serotonin application) used to induce long-term facilitation, PKA phosphorylates not only substrate proteins involved in short-term facilitation at the synapse, but also the transcriptional activator cAMPresponse element binding protein (CREB) in the sensory neuron nucleus. CREB then binds to and activates the promoter element CRE (the cAMP response element), inducing gene expression. The new gene products involve not only ‘effector’ molecules directly involved in the expression of long-term facilitation, but also additional transcription factors, such as C EBP, that regulate the expression of other genes, thereby producing a cascade of gene regulation. Interruption of this gene cascade at early time points by inhibitors of RNA or protein synthesis blocks subsequent long-term facilitation, but leaves short-term facilitation intact. In contrast, inhibition of RNA and protein synthesis at later time points has little effect. Thus the dynamics of macromolecular synthesis in long-term facilitation mimic the time-sensitivity of long-term memory itself. cAMP and CREB have been implicated in neural substrates of memory in other systems, such as Drosophila.

Just as short-term facilitation involves the coordinated facilitation of several processes involved in transmitter release, long-term facilitation involves multiple effector genes. Enhanced expression of the enzyme ubiquitin carboxyterminal hydrolase promotes the degradation of the inhibitory regulatory subunits of PKA, thereby making the kinase persistently active, even after cAMP levels return to baseline. Other gene products lead to the growth of new sensory neuron synaptic terminals (which does not occurn short-term facilitation), thus providing a stable, enduring substrate for long-term memory. Structural changes in sensory neurons are accompanied by corresponding growth in postsynaptic targets, providing further evidence for the coordinated regulation of synaptic efficacy on multiple levels.

3. Functional Consequences Of Simple Synaptic Regulation

3.1 Modulation Of Input–Output Pathways

The functional consequences of simple forms of synaptic regulation can be relatively straightforward in circuits where there is a relatively uniform direction of information flow from input to output. For example, as described above, decreases in synaptic strength at Aplysia sensorimotor connections contribute to decreases in the siphon-withdrawal response (e.g., habituation), and increases in synaptic strength contribute to increases in the siphon-withdrawal response (e.g., sensitization and classical conditioning).

3.2 Reconfiguration Of Multifunctional Networks

In other instances, however, even relatively simple forms of synaptic change can have more complex functional consequences. Many neural networks (or their components) are multifunctional networks, participating in more than one behavior. Because the operation of such networks depends on interactions among multiple non-linear processes at the cellular, synaptic, and network levels, regulation of synaptic efficacy can profoundly alter network operation, reconfiguring the network into functionally distinct modes.

3.2.1 Escape Swimming In Tritonia. One interesting example is the reconfiguration of a premotor network from a resting reflex withdrawal circuit into a rhythmic central pattern generator (CPG) swim circuit in the marine mollusk Tritonia (Katz and Frost 1996). None of the three major interneuron types in this circuit have endogenous bursting properties; the oscillation of the network instead arises from, and hence depends upon, their synaptic interactions. The dorsal swim interneurons (DSIs) are essential elements of the CPG circuit for escape swimming. DSI activity contributes directly to the generation of the swim motor program; in addition, it rapidly enhances the synaptic efficacy and excitability of a second interneuron, C2. C2 has been implicated in the reconfiguration mechanism; its activity disinhibits the DSI neurons, enabling rhythmic dorsal flexions. Thus, the rapid enhancement of C2 contributions may functionally rewire the network, allowing the reconfiguration from simple reflex withdrawal to rhythmic swimming to occur. Because the regulation of C2 synaptic efficacy develops and decays dynamically with DSI activity, it may also play a role in the cycle-by-cycle generation of the swim rhythm itself.

3.2.2 Cpgs In Lobster Stomatogastric Ganglion. The gastric mill CPG and pyloric CPG in the lobster stomatogastric ganglion are also polymorphic networks that can be functionally reconfigured to produce fundamental variations in output under different circumstances (Harris-Warwick and Marder 1991). Such alterations arise both from regulation of synaptic efficacy in the circuit and from changes in the intrinsic properties of the component neurons. For example, stimulation of the anterior pyloric modulator (APM) neuron activates the pyloric motor pattern by enhancing the plateau potential capabilities of pyloric CPG neurons. The platean maintains neurons in a depolarized state and so allows them to fire repetitively. As a consequence, a brief excitatory synaptic input that was previously ineffective can trigger a transition of the neuron to the tonically active plateau condition. In the plateau state, neurons also actively resist hyperpolarization, thus rendering inhibitory inputs less effective. The combined enhancement of synaptic excitation and diminution of inhibition result in a more active network.

Modulatory transmitters produce bursting pacemaker potentials in the anterior burster (AB) neurons, the most important pacemaker of the pyloric CPG. Synaptic interactions within the stomatogastric ganglion are mediated partly by non-spiking release, which varies continuously with membrane potential. Hence, larger bursting potentials enhance transmitter release from AB terminals, increasing AB inhibition of certain postsynaptic CPG targets and removing them from the functional circuit, thereby altering the pyloric network and rhythm. Because different modulatory amines exert different effects on AB bursts, the anatomical circuit can be sculpted in a variety of ways to yield a number of different functional networks.

4. Qualitative Regulation Of Synaptic Efficacy

The decreases and increases in synaptic efficacy described above involve primarily quantitative modifications of synaptic strength, rendering synapses either more or less potent. However, regulation of synaptic efficacy can also involve a number of qualitative changes in synaptic transmission, thus extending the range of adaptive mechanisms that the nervous system utilizes.

4.1 Regulation Of Transmitter Type

An important additional form of synaptic regulation involves changes in the type, rather than the amount, of transmitter that a neuron releases. Such changes can arise either from shifts in the relative release of different co-transmitters resulting from different rates and patterns of neural firing, or from changes in transmitter synthesis (Simerly 1990). Release of one co-transmitter or another can produce qualitative changes in the overall functional effect of synaptic transmission on a particular postsynaptic target (e.g., from excitation to inhibition), rather than simple quantitative changes in a pre-existing synaptic response. Regulation of the transmitter type can also effectively add or remove a particular postsynaptic pathway from operation, thus allowing functional rewiring of the neural circuit without the need for anatomical reorganization. In principle, similar molecular switching might be accomplished by regulation of the relative number or sensitivities of different postsynaptic receptors.

4.1.1 Lobster Stomatogastric Ganglion. Changes in co-transmitter release affect the pyloric CPG in the lobster stomatogastric ganglion. The inferior ventricular nerve (IVN) cells evoke both excitatory and inhibitory responses in the AB neurons, the most important pacemaker of the pyloric CPG, as described above. At low frequencies of IVN activation, one (unknown) excitatory transmitter is released, and the frequency of the pyloric rhythm increases. In contrast, at higher frequencies, the inhibitory transmitter histamine is preferentially released, and the pyloric rhythm is disrupted. In a variation on this theme, the neuromodulatory transmitter dopamine changes the sign of the synaptic connection between certain identified neurons in the CPG from inhibitory to excitatory by reducing the efficacy of the inhibitory synapse and uncovering a weak electrical connection.

4.1.2 Feeding In Aplysia. Differential release of cotransmitters may also provide for intrinsic self-regulation of a network. For example, the motor neurons in the feeding circuit of Aplysia use a variety of peptide co-transmitters that modulate neuromuscular transmission, extending the dynamic range over which the muscles can operate effectively. High frequency activation of motor neuron B15 enhances the co-release of small cardioactive peptide B. This neuromodulatory peptide increases the amplitude and quickens relaxation time of the muscle, thereby enabling larger muscle contractions to be driven at a higher rate. Similarly, in the leech, the myogenic heart muscle is innervated by heart excitatory motor neurons that co-release acetylcholine and FMRFamide, a modulatory peptide that regulates amplitude and duration of heart beat. Amines also modulate the amplitude of potentials at many crustacean neuromuscular junctions.

4.2 Regulation Of Synaptic Plasticity: Metaplasticity

An additional form of synaptic regulation, recently termed ‘metaplasticity’ (Abraham and Bear 1996), is manifested as a modification of the ability of the synapse to undergo changes in synaptic efficacy, rather than as a change in synaptic efficacy per se. For example, long-term potentiation (LTP) not only increases the population spike evoked in dentate granule cells of mammalian hippocampus by afferent stimulation, it also reduces non-linear response properties so that synaptic efficacy is less affected by the recent history of activity at that synapse. LTP thus increases the fidelity as well as the gain of the system (Berger and Sclabassi 1985). Regulation of more longlasting forms of synaptic plasticity also occurs at hippocampal synapses, involving, for example, shifts in the tendency of synapses to exhibit LTP or long-term depression (LTD).

Metaplasticity has recently been implicated at identified inhibitory synapses of the siphon-withdrawal circuit of Aplysia (Fischer et al. 1997). The inhibitory interneuron neuron L30 exhibits multiple forms of synaptic enhancement that control the gain of the siphon-withdrawal reflex, including frequency facilitation during initial phases of high-frequency activation and synaptic augmentation and potentiation that persist after activation. Tail shock reduces the baseline strength of L30 connections onto L29 cells; it also attenuates augmentation potentiation while leaving frequency facilitation relatively intact. Behaviorally, tail shock decreases the inhibition normally produced by cutaneous stimuli, perhaps by reducing augmentation of L30 inhibition, thus suggesting a functional role for metaplasticity within a defined neural circuit.

5. Regulation Of Synaptic Efficacy And Of Other Neural Properties

Although not the explicit subject of this research paper, changes in neural excitability and other intrinsic neural properties can occur in addition to modifications of synaptic efficacy. Synaptic and non-synaptic changes can co-occur in the same neurons, and can be functionally and even mechanistically related. For example, both Aplysia mechanosensory cells and Hermissenda type B photoreceptors exhibit learning related increases in excitability (increased firing to a given stimulus), as well as enhanced synaptic strength (Schuman and Clark 1994). These two modifications work in concert to increase both the number and the potency of the synaptic signals conveyed to their postsynaptic targets. Similarly, as described above, reconfiguration of the Tritonia escape swim CPG and the pyloric CPG in lobster stomatogastric ganglion involve both synaptic and non-synaptic forms of neural regulation.

One important distinction between regulation of synaptic efficacy and of other neural properties is the greater specificity that synaptic changes may exhibit. Because any given neuron may form preor postsynaptic contacts with thousands of other cells, the ability to modify only selected connections of a neuron confers a high degree of precision in the modification of neural pathways. In contrast, intrinsic changes such as enhanced excitability are often expressed cell-wide or nearly cell-wide, enhancing responses to virtually all inputs and enhancing virtually all outputs. In Aplysia siphon sensory cells, several short-term forms of synaptic plasticity, including heterosynaptic facilitation, heterosynaptic inhibition, and Hebbian facilitation, exhibit synapse specificity (Clark and Kandel 1984, Clark 2001). Synapse specificity is perhaps not surprising in short-term regulation, which involves covalent modifications of pre-existing proteins. However, long-term heterosynaptic facilitation, which depends on regulation of transcriptional and translational events, is also in part synapse-specific (Clark 2001, Martin et al. 1997). Synapse-specific long-term facilitation raises the intriguing question of how new gene products are preferentially utilized at facilitated synapses, compared with other synapses of the same neuron. An emerging answer, consistent with comparable findings in hippocampal LTP, is that synapse-specific long-term facilitation involves synaptic tagging and depends in part on local protein synthesis. According to this hypothesis, relatively short-lived, local regulatory events at treated synapses allow the subsequent capture of new gene products that are shipped cell-wide, thus conferring the specificity. Irrespective of the particular mechanistic details, long-term synapse-specific facilitation provides an additional illustrative example of the rich interplay and elegance of the multiple biological processes involved in synaptic regulation.

Bibliography:

1. Abraham W C, Bear M F 1996 Metaplasticity: The plasticity of synaptic plasticity. Trends in Neurosciences 19: 126–30
2. Berger T W, Sclabassi R J 1985 Nonlinear systems analysis and its application to study of the functional properties of neural systems. In: Weinberger N M, McGaugh J L, Lynch G (eds.) Memory Systems of the Brain. Guilford Press, New York
3. Castellucci V, Kandel E R 1976 Presynaptic facilitation as a mechanism for behavioral sensitization in Aplysia. Science 194: 1176–8
4. Clark G A 2001 The ins and outs of classical conditioning: Maxims from a mollusk. In: Steinmetz J E, Gluck M, Solomon P (eds.) Model Systems and the Neurobiology of Associative Learning. Erlbaum, Hillsdale, NJ
5. Clark G A, Kandel E R 1984 Branch-specific facilitation in Aplysia siphon sensory cells. Proceedings of the National Academy of Sciences USA 81: 2577–81
6. Fischer T M, Blazis D E J, Priver N A, Carew T J 1997 Metaplasticity at identified inhibitory synapses in Aplysia. Nature 389: 860–5
7. Glanzman D L 1995 The cellular basis of classical conditioning in Aplysia californica—It’s less simple than you think. Trends in Neurosciences 18: 30–6
8. Harris-Warrick R M, Marder E 1991 Modulation of neural networks for behavior. Annual Review of Neuroscience 14: 39–57
9. Hawkins R D, Abrams T W, Carew T J, Kandel E R 1983 A cellular mechanism of classical conditioning in Aplysia: Activity-dependent amplification of presynaptic facilitation. Science 219: 400–5
10. Hebb D O 1949 The Organization of Behavior. Wiley, New York
11. Kandel, E R, Pittenger C 1999 The past, the future and the biology of memory storage. Philosophical Transactions of the Royal Society London B 354: 2027–52
12. Kandel E R, Schwartz J H 1982 Molecular biology of learning: Modulation of transmitter release. Science 218: 433–43
13. Katz P S, Frost W N 1996 Intrinsic neuromodulation: Altering neuronal circuits from within. Trends in Neurosciences 19: 54–61
14. Martin K C, Casadio A, Zhu H E-Y, Rose J C, Chen M, Bailey C H, Kandel E R 1997 Synapse-specific, long-term facilitation of Aplysia sensory to motor synapses: A function for local protein synthesis in memory storage. Cell 91: 927–38
15. Montarolo P G, Goelet P, Castellucci V F, Morgan J, Kandel E R, Schacher S 1986 A critical period for macromolecular synthesis in long-term heterosynaptic facilitation in Aplysia. Science 234: 1249–54
16. Schuman E M, Clark G A 1994 Synaptic facilitation at connections of Hermissenda type B photoreceptors. Journal of Neuroscience 14: 1613–22
17. Simerly R B 1990 Hormonal control of neuropeptide gene expression in sexually dimorphic olfactory pathways. Trends in Neurosciences 13: 104–10