Amygdala

Long-term synaptic potentiation (LTP) has emerged as the likeliest synaptic substrate for rapid associative learning (Brown, Chapman, Kariss, and Keenan, 1988; Kelso and Brown, 1986). LTP has been reported in numerous brain structures, including the amygdala, where it may participate in emotional learning (Blair et al., 2001; Chapman, Kairiss, Keenan, and Brown, 1990). This article examines an unexpected property of LTP in synapses of the amygdala and the implications of this finding on attempts to link amygdalar LTP to emotional learning.

Amygdalar Neuroanatomy and Neurophysiology

Synaptic transmission and its plasticity seem to depend jointly on the identity of the presynaptic and postsynaptic neurons. A consideration of LTP in the amygdala thus requires some knowledge of the anatomy and physiology of the neurons in this brain region. The functional neuroanatomy is also important for interpreting electrophysiological recordings.

The amygdaloid nuclear complex is a collection of subnuclei that fall into two major classes based on cell morphology: those containing cortexlike neurons and those containing striatallike neurons (Swanson and Petrovich, 1998). The cortexlike group includes the lateral, basal, and basolateral nuclei, sometimes collectively termed the basolateral amygdaloid complex (BLA). The more medially located nuclei, including the central and medial nuclei, are part of the ventromedial expanse of the striatum.

At a circuit level, the functional neuroanatomy of the amygdala remains a mystery, but there have been a few studies of the anatomy and physiology of individual neurons (Chapman, Kairiss, Keenan, and Brown, 1990; Faulkner and Brown, 1999; Washburn and Moises, 1992). The organization of cells in BLA, although different from that of any other brain region, perhaps resembles most closely a combination of the adjacent perirhinal cortex and the CA3 region of the hippocampus. While the unusual cell-firing types are similar to those in the perirhinal cortex (Faulkner and Brown, 1999; McGann, Moyer, and Brown, 2001), the massive system of excitatory recurrents and consequent tendency toward epileptiform activity when pharmacologically disinhibited is reminiscent of the CA3 region of the hippocampus (Brown and Zador, 1990).

Combined studies of BLA by Brown and coworkers (Chapman, Kairiss, Keenan, and Brown, 1990; Faulkner and Brown, 1999) and Washburn and Moises (1992) suggest that there may be as many as seven firing patterns, including regular spiking, single spiking, late spiking, burst spiking, slow charging, irregular spiking, and fast spiking. Small, aspiny stellates tend to be fast-spiking cells, presumably inhibitory interneurons. Pyramids and large stellates can be regular spiking, single spiking, burst spiking, or late spiking. Little is known about other structure-function relationships.

Spontaneous postsynaptic potentials (PSPs) and postsynaptic conductances (PSCs) are commonly much larger than those of the CA1 region of the hippocampus but similar to those in CA3 pyramidal neurons (Faulkner and Brown, 1999; Johnston and Brown, 1984; Xiang and Brown, 1998). Electrical stimulation of the external capsule (EC) commonly elicits in BLA neurons an excitatory-inhibitory conductance sequence (Chapman, Kairiss, Keenan, and Brown, 1990), similar to that which Johnston and Brown and coworkers reported in hippocampal region CA3 (Barrionuevo, Kelso, Johnston, and Brown, 1986; Griffith, Brown, and Johnston, 1986; Johnston and Brown, 1983).

Discovery of LTP in the Amygdala

Brown's team looked for LTP elicited by EC stimulation in horizontal brain slices containing BLA (Chapman, Kairiss, Keenan, and Brown, 1990; Keenan, Chapman, Chang, and Brown, 1988). The gross cytoarchitecture of the slice was vividly revealed through the use of differential-interference contrast (DIC) optics and infrared (IR) illumination, which Brown and cowokers had been developing for this purpose (Keenan, Chapman, Chang, and Brown, 1988). At low magnification, IR-DIC optics sharply resolved the key landmarks needed to position the stimulating and recording electrodes to repeatable locations that can be reliably matched to corresponding plates of a rat brain atlas. Critically important were the borders of EC, the basolateral nucleus, the lateral nucleus, and the central nucleus (Chapman, Kairiss, Keenan, and Brown, 1990).

EC was selected as the placement site for the stimulation electrode for four reasons. First, it was the only visually discernible large population of known afferent inputs to BLA. Second, in contrast to other possible stimulation sites, this allowed a more homogeneous population of synapses, thus partaking of some of the virtues of studies of the Schaeffer collateral/commissural (Sch/com) inputs to pyramidal neurons in hippocampal region CA1. Third, because EC contains fast-conducting fibers, these were presumed to produce the earliest detectable postsynaptic responses. Fourth, neurophysiological results suggested that EC-elicited PSCs in BLA neurons contain monosynaptic responses from the stimulation site.

Regarding this last point, the critical question was not whether EC fibers project to BLA—they are known to do so. The question was whether the PSCs recorded in BLA neurons in response to EC stimulation reflected direct synaptic inputs from stimulated EC fibers. The obvious alternative was that PSCs recorded in response to EC stimulation were produced by inputs from other BLA neurons whose firing was caused by EC stimulation. The difference can be important for designing and interpreting neurophysiology experiments (Xiang and Brown, 1998).

To minimize activation of recurrent circuitry and to avoid epileptiform activity (Johnston and Brown, 1983; Xiang and Brown, 1998), the slices were not disinhibited pharmacologically. The excitatory component of the synaptic conductance waveform was first evaluated (Griffith, Brown, and Johnston, 1986) during baseline testing (thirty to ninety responses at 0.1 Hz). An attempt was then made to induce LTP by delivering three to ten trains (300 msec each) of high-frequency (100 Hz, "tetanic") electrical currents (each 0.1 msec) through the stimulation electrode that was positioned in EC. Following tetanic stimulation, testing was resumed at 0.1 Hz for as long as two hours.

The tetanic stimulation produced an early synaptic enhancement that relaxed back to a sustained level of synaptic potentiation. Because of posttetanic potentiation (PTP), the early enhancement lasted less than fifteen minutes, as in the hippocampus (Barrionuevo, Kelso, Johnston, and Brown, 1986; Kelso and Brown, 1986; Kelso, Ganong, and Brown, 1986). LTP was measured conventionally as the stable enhancement that clearly outlasted PTP. The tetanic stimulation protocol induced LTP in 80 percent of the neurons studied, which averaged a 46 percent increase.

Kinds of LTP in the Amygdala

Two obvious questions arise from the countless studies of LTP in the hippocampus. The first question is whether there is a kind of LTP in the amygdala whose induction depends on the activation of N-methyl-d-aspartate receptors (NMDARs). NMDAR-dependent LTP in the Sch/com input to CA1 is one of the most commonly studied forms of LTP. There is not much neurophysiological evidence for a similar kind of plasticity in the amygdala.

Chapman and Bellavance (1992) found that APV (50 μM), a competitive antagonist for glutamate at the NMDAR, did not block LTP induction produced by tetanic stimulation of EC. These investigators report that LTP induction could be inhibited by APV when the concentration was increased to 100 μM, but this concentration also strongly attenuated EC-evoked PSPs. The recording methods used by Chapman and Bellavance presumably tended to sample from a population of larger BLA neurons, but other researchers have reported similar results in BLA using selection criteria designed to sample small interneurons.

Mahanty and Sah (1998) examined LTP in a subset of cells that have fast-action potentials and that lack spike-frequency adaptation. EC stimulation (two trains of 100 stimuli presented at 30 Hz) resulted in LTP in seven of seven of these putative interneurons. Pharmacological and electrophysiological analysis suggested that NMDARs contribute relatively little to the EC-evoked PSCs. Application of APV (50 μM) had no effect on LTP induction in these putative interneurons.

LeDoux's team (Weisskopf, Bauer, and LeDoux, 1999) reported similar results in a previously unexplored synaptic circuit. They positioned a stimulating electrode in a region of the striatum immediately dorsal to the central nucleus and medial to the lateral nucleus, where they recorded evoked PSPs. LTP induction was unaffected when 50 μM APV was added to the bathing medium.

The second question to emerge from research on hippocampal LTP concerns the role of backpropagating dendritic Ca²⁺ spikes (Zador, Koch, and Brown, 1990). Research in Johnston's laboratory has shown that NMDARs are only part of the story behind LTP induction in the hippocampus (Magee, Hoffman, Colbert, and Johnston, 1998). Recent work by Johnston and coworkers (Magee and Johnston, 1997) demonstrated the critical importance of backpropagating Ca²⁺ spikes in controlling the induction of one form of LTP in the Sch/com synaptic input to CA1 pyramidal neurons. Blocking L-type Ca²⁺ channels prevented or greatly inhibited LTP induction in these pyramidal neurons (Magee and Johnston, 1997). In a parallel finding, Teyler and coworkers (Cavus and Teyler, 1996) have shown that the Sch/com synapses on CA1 pyramidal neurons can undergo two forms of LTP: one that is dependent on NMDARs and another that depends on L-type Ca²⁺ channels.

Teyler and coworkers noted that NMDAR-dependent LTP and Ca²⁺ channel-dependent LTP engage distinct signal transduction cascades and thus could have different functional significance for encoding (Cavus and Teyler, 1996; Morgan, Coussens, and Teyler, 2001). Teyler hypothesized that NMDAR-dependent LTP might be more critically involved in short-term memory, whereas Ca²⁺-channel-dependent LTP could be more important in long-term memory. This suggestion was based partly on findings by Teyler and coworkers (Morgan, Coussens, and Teyler, 2001) that Ca²⁺-channel-dependent LTP has a slower onset, it is more persistent, and it is less subject to elimination by long-term depression (LTD)—at least in brain slices.

Some additional findings from the previously mentioned study by LeDoux and coworkers (Weisskopf, Bauer, and LeDoux, 1999) are relevant to Teyler's hypothesis. Recall that LeDoux and coworkers examined LTP in a pathway to the lateral nucleus by stimulation of a region of the striatum that is near the central nucleus of the amygdala. Although they found no effect of bath application of 50 μM APV on LTP induction, bath application of the L-type Ca²⁺ channel blocker, Nifedipine (30 μM), prevented LTP.

Results from the handful of studies that have been done on the neuropharmacology of LTP in BLA neurons do not seem to fit the pattern seen in the most commonly studied form of LTP that can be induced in the Sch/com synaptic input to CA1 pyramids. With the Sch/com inputs, 100 μM APV effectively prevents LTP induction but has no detectable effect on evoked PSPs and no detectable effect on the expression of previously induced LTP (see Brown et al., 1989). NMDARs play a key role in inducing this form of LTP in Sch/com synapses (Brown, Chapman, Kairiss, and Keenan, 1988). However, under normal experimental conditions, NMDARs are not required for evoking PSPs or for expressing this form of LTP after it has been established.

Behavioral Implications of Amygdalar LTP

Information about amygdalar synapses has obvious implications for studies on the role of the amygdala in emotional learning. In one set of conditioning experiments, Davis and coworkers (Campeau, Miserendino, and Davis, 1992; Miserendino, Sananes, Melia, and Davis, 1990) perfused the amygdala with APV and claimed to have shown that the drug affects learning but not ordinary synaptic transmission nor the ability to recall previous learning. Direct studies of amygdalar synapses, however, predict that infusing BLA with APV at a concentration that blocks amygdala-dependent learning could also impair retrieval of previously established memories by interfering with synaptic signalling and therefore access to memory.

Other laboratories (Fendt, 2001; Lee, Choi, Brown, and Kim, 2001; Lindquist and Brown, 2002) convincingly demonstrated that APV infusion directly into the amygdala does impair its functioning in a more general manner than had been previously claimed by Davis and coworkers. As expected, infusing the amygdala with APV during the time of conditioning impaired learning (Lee, Choi, Brown, and Kim, 2001), as revealed by subsequent testing (done in the absence of APV). Contrary to claims by Davis and coworkers (Campeau, Miserendino, and Davis, 1992; Miserendino, Sananes, Melia, and Davis, 1990), these laboratories also found a profound memory impairment when the amygdala was perfused with APV at the time of testing (but not during the time of conditioning).

Multiple behavioral measures have shown that infusing BLA with APV profoundly impairs both the acquisition and expression of several classical, BLA-dependent, conditioned-fear responses (CRs). Perhaps even more surprising and dramatic was the finding that infusing BLA with APV also greatly attenuated or even eliminated unconditioned responses (URs) that are normally elicited during conditioning trials. UR reactivity during conditioning predicted CR production during subsequent testing (Lee, Choi, Brown, and Kim, 2001). The collective research on LTP in amygdalar neurons correctly anticipated that the effect of APV on amygdala function is much less specific than Davis and coworkers contended (Campeau, Miserendino, and Davis, 1992; Miserendino, Sananes, Melia, and Davis, 1990).

Conclusion

There are multiple forms of LTP/LTD that can be spatially segregated in different neurons or even colocalized in the same synapses. Since the discovery of amygdalar LTP by Brown and coworkers (Chapman, Kairiss, Keenan, and Brown, 1990) it has become clear that this form of LTP can be induced in the presence of APV unless the concentration is high enough to interfere with experimentally evoked PSPs. This form of LTP is especially interesting in the context of Teyler's hypothesis that APV-resistant LTP is preferentially involved in long-term memory, an idea that could be relevant to the persistence of emotional memory. Since1990, when LTP was first observed in amygdala brain slices, the reliability and resolution of in vitro technology has matured enough to hasten progress in subsequent research (Faulkner and Brown, 1999; Moyer and Brown, 2002).

See also:GUIDE TO THE ANATOMY OF THE BRAIN: AMYGDALA; NEURAL SUBSTRATES OF CLASSICAL CONDITIONING: FEAR-POTENTIATED STARTLE; NEURAL SUBSTRATES OF EMOTIONAL MEMORY

Bibliography

Barrionuevo, G., Kelso, S. R., Johnston, D., and Brown, T. H. (1986). Conductance mechanism responsible for long-term potentiation in monosynaptic and isolated excitatory synaptic inputs to hippocampus. Journal of Neurophysiology 55, 540-550.

Blair, H. T., Schafe, G. E., Bauer, E. P., Rodrigues, S. M., and Le- Doux, J. E. (2001). Synaptic plasticity in the lateral amygdala: A cellular hypothesis of fear conditioning. Learning and Memory 8, 229-42.

Brown, T. H., Chapman, P. F., Kairiss, E. W., and Keenan, C. L. (1988). Long-term synaptic potentiation. Science 242, 724-728.

Brown, T. H., Ganong, A. H., Kairiss, E. W., Keenan, C. L., and Kelso, S. R. (1989). Long-term potentiation in two synaptic systems of the hippocampal brain slice. In J. H. Byrne and W. O. Berry, eds., Neural models of plasticity. San Diego: Academic Press.

Brown, T. H., and Zador, A. M. (1990). The hippocampus. In G. Shepherd, ed., Synaptic organization of the brain. New York: Oxford University Press.

Campeau, S., Miserendino, M. J., and Davis, M. (1992). Intraamygdala infusion of the N-methyl-D-aspartate receptor antagonist AP5 blocks acquisition but not expression to an auditory conditioned stimulus. Behavioral Neuroscience 106, 569-574.

Cavus, I., and Teyler, T. J. (1996). Two forms of long-term potentiation in area CA1 activate different signal transduction pathways. Journal of Neurophysiology 76, 3,038-3,047.

Chapman, P. F., and Bellavance, L. L. (1992). Induction of long-term synaptic potentiation in the basolateral amygdala does not depend on NMDA receptor activation. Synapse 11, 310-318.

Chapman, P. F., Kairiss, E. W., Keenan, C. L., and Brown, T. H. (1990). Long-term synaptic potentiation in the amygdala. Synapse 6, 271-278.

Faulkner, B., and Brown, T. H. (1999). Morphology and physiology of neurons in the rat perirhinal-lateral amygdala area. Journal of Comparative Neurology 411, 613-642.

Fendt, M. (2001). Injections of the NMDA receptor antagonist aminophosphonopentanoic acid into the lateral nucleus of the amygdala block the expression of fear-potentiated startle and freezing. The Journal of Neuroscience 21, 4,111-4,115.

Griffith, W. H., Brown, T. H., and Johnston, D. (1986). Voltageclamp analysis of synaptic inhibition during long-term potentiation in hippocampus. Journal of Neurophysiology 55, 767-775.

Johnston, D., and Brown, T. H. (1983). Voltage-clamp analysis of mossy fiber synaptic input to hippocampal neurons. Journal of Neurophysiology 50, 464-486.

—— (1984). Biophysics and microphysiology of synaptic transmission in hippocampus. In R. Dingledine, ed., Brain slices. New York: Plenum Press.

Keenan, C. L., Chapman, P. F., Chang, V., and Brown, T. H. (1988). Videomicroscopy of acute brain slices from hippocampus and amygdala. Brain Research Bulletin 21, 373-383.

Kelso, S. R., and Brown, T. H. (1986). Differential conditioning of associative synaptic enhancement in hippocampal brain slices. Science 232, 85-87.

Kelso, S. R., Ganong, A. H., and Brown, T. H. (1986). Hebbian synapses in hippocampus. Proceedings of the National Academy of Sciences of the United States of America 83, 5,326-5,330.

Lee, H. J., Choi, J.-S., Brown, T. H., and Kim, J. J. (2001). Amygdalar N-methyl-D-aspartate (NMDA) receptors are critical for the expression of multiple conditioned fear responses. Journal of Neuroscience 21, 4,116-4,124.

Lindquist, D. H., and Brown, T. H. (2001). Antagonizing NMDA receptors in the basolateral complex of the amygdala prevents conditioned enhancement of the rat eyeblink reflex. Integrative Physiological and Behavioral Science 36, 171.

Magee, J., Hoffman, D., Colbert, C., and Johnston, D. (1998). Electrical and calcium signaling in dendrites of hippocampal pyramidal neurons. Annual Review of Physiology 60, 327-346.

Magee, J. C., and Johnston, D. (1997). A synaptically controlled associative signal for Hebbian plasticity in hippocampal neurons. Science 275, 209-213.

Mahanty, N. K., and Sah, P. (1998). Calcium-permeable AMPA receptors mediate long-term potentiation in interneurons in the amygdala. Nature 394, 683-687.

McGann, J. P., Moyer, J. R., Jr., and Brown, T. H. (2001). Predominance of late-spiking neurons in layer VI of rat perirhinal cortex. Journal of Neuroscience 21, 4,969-4,976.

Miserendino, M. J. D., Sananes, C. B., Melia, K. R., and Davis, M. (1990). Blocking of acquisition but not expression of conditioned fear-potentiated startle by NMDA antagonists in the amygdala. Nature 345, 716-718.

Morgan, S. L., Coussens, C. M., and Teyler, T. J. (2001). Depotentiation of vdccLTP requires NMDAR activation. Neurobiology of Learning and Memory 76, 229-238.

Moyer, J. R., Jr., and Brown, T. H. (2002). Patch-clamp techniques applied to brain slices. In A. Walz, A. Boulton, and G. B. Baker, eds., Advanced techniques for patch-clamp analysis. Totowa, NJ: Humana Press.

Swanson, L. W., and Petrovich, G. D. (1998). What is the amygdala? Trends in Neuroscience 21, 323-331.

Washburn, M. S., and Moises, H. C. (1992). Electrophysiological and morphological properties of rat basolateral amygdaloid neurones in vitro. Journal of Neuroscience 12, 4,066-4,079.

Weisskopf, M. G., Bauer, E. P., and LeDoux, J. E. (1999). L-type voltage-gated calcium channels mediate NMDA-independent associative long-term potentiation at the thalamic input synapses to the amgydala. Journal of Neuroscience 19, 10,512-10,519.

Xiang, Z., and Brown, T. H. (1998). Complex synaptic current waveforms evoked in hippocampal pyramidal neurons by extracellular stimulation of dentate gyrus. Journal of Neurophysiology 79, 2,475-2,484.

Zador, A., Koch, C., and Brown, T. H. (1990). Biophysical model of a Hebbian synapse. Proceedings of the National Academy of Sciences of the United States of America 87, 6,718-6,722.

Thomas H.Brown

Derick H.Lindquist