Report. Autaptic Excitation Elicits Persistent Activity and a Plateau Potential in a Neuron of Known Behavioral Function

Transcription

1 Current Biology 19, , March 24, 2009 ª2009 Elsevier Ltd All rights reserved DOI /j.cub Autaptic Excitation Elicits Persistent Activity and a Plateau Potential in a Neuron of Known Behavioral Function Report Ravit Saada, 1,2 Nimrod Miller, 1,2 Itay Hurwitz, 1 and Abraham J. Susswein 1, * 1 The Leslie and Susan Gonda (Goldschmied) Multidisciplinary Brain Research Center and The Mina and Everard Goodman Faculty of Life Sciences Bar Ilan University Ramat Gan, Israel Summary Synaptic connections from a neuron onto itself (autapses) are not uncommon [1], but their contributions to information processing and behavior are not fully understood. Positive feedback mediated by autapses could in principle give rise to persistent activity, a property of some neurons in which a brief stimulus causes a long-lasting response [2]. We have identified an autapse that underlies a plateau potential causing persistent activity in the B31/B32 neurons of Aplysia. The persistent activity is essential to the ability of these neurons to initiate and maintain components of feeding behavior. Persistent activity in B31/B32 arises from a voltage-dependent muscarinic autapse and from pharmacologically identical network-based positive feedback. Depolarization via the autapse begins later than networkdriven excitation, and the effect of the autapse is therefore overshadowed by the earlier network-based depolarization. In B31/B32 neurons isolated in culture only the autapse is present, and the autapse functionally replaces the missing network-based feedback. Properties of B31/B32 provide insight into a possible general function of autapses. Autapses might function along with synapses from presynaptic neurons as components of feedback loops. Results The B31/B32 neurons in the Aplysia buccal ganglia are cholinergic central pattern generator (CPG) and motor neurons [3 5] whose activity corresponds to the protraction phase of a consummatory feeding response [5]. Brief depolarization elicits a persistent 40 mv depolarization with superimposed axon spikes that fail to invade the soma [4, 6]. Persistent activity outlasting a stimulus might arise via positive feedback networks or from endogenous plateau potentials [2, 7]. We examined the mechanisms underlying B31/B32 persistent activity. Muscarinic Excitation of B31/B32 In Situ B63 drives B31/B32s via electrical coupling and fast and slow cholinergic excitatory postsynaptic potentials (EPSPs) (Figures 1A and 1C) [8 10; I. Hurwitz et al., 1999, Soc. Neurosci. Abstr., abstract]. B31/B32s excite themselves via the coupling to B63 [8, 10]. B31/B32 persistent activity can be terminated by hyperpolarizing B31/B32 (Figure 1B), indicating that B31/B32s *Correspondence: avy@mail.biu.ac.il 2 These authors contributed equally to this work display bistable states characteristic of neurons with plateau potentials. Persistent B31/B32 activity depends on muscarinic transmission. As reported previously [8, 11; I. Hurwitz et al., 1999, Soc. Neurosci. Abstr., abstract], tetrodotoxin (TTX) (Figure 1D2), or the muscarinic antagonist pirenzepine, blocked persistent B31/B32 activity. A muscarinic agonist restored the persistent activity after TTX (Figure 1D3) or pirenzepine [8]. Neurons treated with oxotremorine and TTX also displayed spontaneous persistent depolarizations (Figure 1D4), which might result from random depolarizations initiating a voltagedependent response. The B63 to B31/B32 connection was explored in a high divalent cation solution that raises spike thresholds, thereby decreasing polysynaptic contributions. B63 elicited fast and slow EPSPs in B31/B32 (Figure 2A). The fast EPSPs are blocked by the nicotinic antagonist hexamethonium (not shown) [12]. The slow EPSP was blocked by pirenzepine (Figure 2A). The slow EPSP was larger when a B31/B32 was held at 240 mv than at 260 mv. At 240 mv, the slow EPSP produced an 8 mv depolarization maintained for s. B63 has a lower firing threshold than B31/B32 axons, and therefore B63 fires earlier than the axons (Figure 1C). If B31/ B32s and B63 act on a common post-synaptic target, B31/ B32 effects will follow those of B63 and will be superimposed on them. It will be difficult to separate the effects of B31/B32s and B63, because B63 and B31/B32s both release acetylcholine (ACh). We examined whether B63 and B31/B32s together act on B31/B32, by stimulating a B31/B32 axon (approximately mm medial to the soma) with brief pulses while recording from the soma (N = 3 preparations). The pulses were too brief to activate B63 spikes reliably. When the stimulus did elicit B63 spiking, the fast EPSPs resulting from these spikes were monitored in B31/B32 (see Supplemental Data, Figure S1, available online). Runs in which B63 fired were discarded. Firing the B31/B32 axon at its physiological frequency [11] caused a depolarization comparable in amplitude and duration to that caused by B63 (Figure 2B). This depolarization was sensitive to pirenzepine (Figure 2B), and was voltage dependent: its amplitude was larger when the B31/B32 soma was held at 240 mv than at 260 mv. Because both the autaptic and the B63-dependent slow muscarinic EPSPs are voltage sensitive, their contributions to the B31/B32 persistent response at the physiological voltage (220 mv) will be larger than those shown in Figure 2. The relative contribution of each component at 220 mv was not tested because the two components could not be separated at this voltage. B31/B32 Morphology Is Consistent with Autaptic Transmission For determining whether their morphology is consistent with the presence of autaptic connections, B31/B32 neurons were filled with 4% lucifer, fixed, and observed via confocal microscopy. B31/B32s had many neurites close to the soma that branch profusely in the area of the soma and proximal axon (Figure 2C). After the initial segment, B31/B32 axons give off relatively few neurites as they innervate the I2 muscle [4]. Neurites close to the soma have many varicosities

2 Current Biology Vol 19 No A B63 B31/B32 ACh ACh I2 muscle (protraction) B 1) 2) -30 na 3 s 30 mv +30 na C axon spike D B63 EPSP axon spike 1) ASW 2) TTX 3) TTX, oxotremorine +20 na 4) spontaneous 2 s 20 mv Figure 1. Properties of B31/B32 (A) The connections between B63 and B31/B32. Triangles represent excitatory synapses; Resistor, electrical connection. Both neurons are cholinergic. (B) Bistable membrane potentials in a B31/B32. (1) A 3 s depolarization induces a plateau depolarization in B31/B32. (2) A 1 s hyperpolarization restores the B31/B32 to rest potential. (C) A brief stimulus to either B63 (upper traces) or a B31/B32 (lower traces) initiates a buccal motor program, which is characterized by firing in B63 that elicits fast EPSPs in the B31/B32. B31/B32 then slowly depolarizes, increasing the firing frequency in B63, which causes still greater depolarization of the B31/B32. B31/B32 then displays a persistent depolarization. Unlike most somata in Aplysia, the B31/B32 somata are inexcitable. Axon spikes that fail to invade the soma are superimposed on the B31/B32 depolarization. The depolarization is terminated by the activity of neuron B64, which inhibits both B31/B32 and B63. (D) The response of a B31/B32 to a 5 s, 20 na stimulus in a number of conditions. (1) In artificial seawater (ASW), the stimulus induced spikes in B63, which produced fast EPSPs in B31/B32, followed by a persistent depolarization. (2) Treatment with TTX blocks B63 spikes and EPSPs, and also blocks the subsequent depolarization. (3) Addition of a muscarinic agonist (oxotremorine) restores the ability of the stimulus to elicit a persistent depolarization. Parts (1 3) of this figure replicate earlier data from reference [8], and are presented as background for the data below. (4) In the presence of TTX and oxotremorine B31/ B32s also generate spontaneous depolarizations. See Hurwitz et al. [9] for details. (Supplemental Data, Movie S1 and Figure S2), which are generally regarded as presynaptic release sites [13 15]. In many cases neurites loop back and contact themselves. In addition, in some cases neurites contact one another, with varicosities located at contact sites (Supplemental Data, Figure S2). Although B31/B32s are postsynaptic targets of many neurons and are electrically coupled to many neurons, to date no postsynaptic neural follower of B31/B32s has been identified. The presence of varicosity-bearing neurites close to the soma is consistent with autaptic transmission. These varicosities are likely to be autaptic release sites. Autaptic Transmission in Culture Proof that B31/B32 muscarinic autapses contribute to persistent activity could be obtained by culturing B31/B32s in isolation. Invertebrate neurons in culture often retain many of the properties and connections seen in situ [17 25]. B31/B32 and other buccal ganglia neurons were marked with fast green, and then removed from the ganglia and cultured in isolation. A brief stimulus to an isolated B31/B32 neuron elicited a persistent depolarization of 30 mv or more. Activity similar to that in Figure 3A was seen repeatedly in 23 of 35 mature (4 7 days in culture) isolated B31/B32 neurons. These neurons displayed overshooting action potentials not present in situ, which could be a consequence of an increase in input resistance of isolated B31/B32 cells. B31/B32s are massively electrically coupled in situ, and their input resistance was 4 MU (60 standard deviation [SD], n = 4 neurons), whereas isolated B31/B32s had an input resistance of 37 MU (67.2 SD, n = 11 neurons). Is persistent activity in isolated B31/B32 neurons dependent on synaptic transmission? Treatment with TTX blocked the persistent depolarization (Figure 3B) (in all four preparations). Bathing in a low Ca 2+ solution that blocks transmitter

3 Autaptic Muscarinic Excitation 481 A B C Figure 2. Autaptic Transmission in B31/B32 (A) B63 elicits fast and slow EPSPs in B31/B32. The slow EPSP is voltage dependent, and is blocked by the muscarinic antagonist pirenzepine. The traces show recordings from the B31/B32 soma as a result of stimulating B63 at 10 Hz. Top traces are recordings while holding B31/B32 at 240 mv, whereas in bottom traces B31/B32 was held at 260 mv. Recordings were made in the presence of ASW and in the presence of the muscarinic blocker pirenzepine, which blocked the slow EPSP. The amplitude of the slow EPSP is larger at 240 mv than at 260 mv. Note that the amplitude of the fast EPSPs is reduced at 240 mv, as the conductance underlying the fast EPSP is ohmic, and the holding potential is closer to the equilibrium potential. (B) Action potentials in the B31/B32 axon at a physiological frequency (22 Hz) produce an autaptic slow muscarinic, voltage-dependent depolarization. Recordings were performed while holding the B31/B32 soma at either 240 mv (top traces) or 260 mv (bottom traces). The B31/B32 axon was penetrated and stimulated with current pulses (10 ms, 60 na) that elicited single action potentials. Voltage transients recorded while stimulating the axon are clipped in the figure. At both voltages the axon was stimulated in the presence of ASW and in the presence of the muscarinic antagonist pirenzepine. The recording at 240 mv shows an averaged trace from two runs in each of three separate preparations. The recording at 260 mv showed an averaged trace from two runs in a single preparation. (C) Morphology of B31/B32 from a whole mount viewed with confocal microscopy. (i) A low-magnification picture of a B31/B32 neuron filled with lucifer yellow. A largediameter axon exits from the soma medially (downward). Many neurites arising from the soma and proximal axon are present. These neurites bear many varicosities, transmitter release sites. (ii) Enlargement of area marked in (1). release also blocked the B31/B32-persistent depolarization (Figure 3C) (in all six preparations). The muscarinic antagonist pirenzepine also blocked persistent activity (in all four preparations) (Figure 3D), and the muscarinic agonist oxotremorine restored persistent activity without spikes after treatment with TTX (in all four preparations) (Figure 3E). Oxotremorine plus TTX also elicited spontaneous depolarizations in B31/ B32 (Figure 3F), as it did in situ (see Figure 1D4). The B31/B32 Autapse Is Not an Artifact of Culturing Culturing neurons might change their properties [17]. In mammals, culturing neurons in isolation causes an increased tendency to form autapses [1, 26, 27]. If B31/B32 neurons develop autapses because they are isolated, isolating other buccal ganglia neurons should cause autapses to develop. Three neurons with properties similar to those of B31/B32 were cultured in isolation. B64 is similar to B31/B32 in that it displays persistent activity in situ [28]. However, B64 activity arises via an endogenous plateau potential not requiring synaptic transmission [28]. Neurons B8 and B61 resemble B31/B32 in that they are cholinergic excitatory motor neurons that are also cholinoceptive [3 5, 29, 30], but they do not display persistent activity in situ. Properties of isolated B64 neurons (n = 3) were similar to those in situ (Figure 4A). A 200 ms depolarization induced a plateau potential and persistent activity in situ (Figure 4A1) and in isolated neurons (Figure 4A2). These were not blocked by TTX that blocked firing (Figure 4A3), indicating that they are not dependent on autaptic transmission. Depolarizing and firing isolated B8 (n = 7) or B61 (n = 4) neurons led to no subsequent persistent activity (not shown), indicating that persistent activity does not routinely form in isolated buccal ganglion neurons. Autapses in B31/B32 might arise because targets for synaptogenesis are not present. If so, coculturing B31/B32 with other cholinoceptive neurons should eliminate B31/B32 autapses. However, in cocultures of B31/B32 and either B8 or B61 (n = 2), B31/B32 displayed autaptic activity (Figure 4B1). Autapses were especially striking when B31/B32 neurons were cocultured (n = 3). The neurons were electrically coupled, and in three runs depolarizing one B31/B32 depolarized the other, causing spiking and a persistent depolarization (not shown). However, in seven runs depolarizing one neuron caused persistent activity only in the stimulated cell (Figure 4B2), indicating that each B31/B32 preferred synapsing onto itself. Discussion In addition to providing insight into the function of autaptic transmission, our data provide insight into how the properties

4 Current Biology Vol 19 No Figure 3. Properties of B31/B32 in Culture (A) A 0.2 na, 5 s depolarization elicited persistent activity. Note the slow rise time of the depolarization (over 8 s from the start of the stimulus to the peak depolarization). In situ B31/B32, the persistent response has a comparable slow rise (12). (B) A series of pulses in 0.2 na increments after treatment with TTX. TTX blocked the ability of depolarizing pulses to induce persistent activity. (C) Treatment with a 0 Ca 2+ solution that blocks synaptic transmission blocks the persistent activity. The neuron was stimulated with a series of pulses that were increased in 0.2 na increments. The 0.2, 0.4, 0.6, and 0.8 na depolarizing pulses induced spiking without persistent activity. (D) Response to a depolarizing pulse much larger than that used in other experiments (3.0 na) in the presence of the muscarinic antagonist pirenzepine. The stimulus elicited spikes, but not persistent activity. (E) After treatment with TTX, the cell was treated with oxotremorine, a muscarinic agonist. A 1 na pulse elicited persistent activity, in spite of the absence of spikes. (F) In the presence of TTX and oxotremorine, spontaneous depolarizations similar to the elicited responses were seen. Note the slow rise time of the depolarization. A B C ASW TTX 0 Ca ++, + Co ++ D E F pirenzepine TTX, oxotremorine, depolarization TTX, oxotremorine, spontaneous 20 mv 2 s of a neuron give rise to behavior. The B31/B32 neurons have a central role in deciding whether to respond to food, and in effecting part of the response [3, 5, 8, 11, 12]. Persistent activity underlies the ability of the cell to decide and to effect behavior [11]. Understanding cellular mechanisms underlying persistent activity lies at the heart of understanding how the properties of B31/B32s produce behavior. The B31/B32 Autapse in Culture Culturing a neuron allows one to examine endogenous properties of neurons that are partially hidden by network properties [20 22]. Culture has been fruitful in examining the properties of persistent plateau-like responses in CPG neurons [18 25]. Growing CPG neurons in culture has been used to demonstrate that some neurons display endogenous plateau potentials, whereas others do not [19, 21, 22, 25]. Isolated B31/ B32 and B64 neurons in culture respond to a brief stimulus with a sustained plateau depolarization (Figures 3 and 4), indicating that endogenous plateau potentials are present in both. The B31/B32 plateau potential was dependent on autaptic muscarinic transmission (Figure 3), whereas the B64 plateau depolarization was not (Figure 4A). A B64 1) 2) 3) in situ B Co-cultures 1) B61 B31 isolated 2) B31 B31 isolated, TTX 2 s 5 s 20 mv Figure 4. The B31/B32 Autapse Is Not an Artifact of Culture (A) Properties of B64 in situ and isolated in culture. In all three traces, a 200 ms depolarization (bar underneath the trace) elicited a plateau potential. (1) In situ, the stimulus amplitude was 3 na. (2) In culture the stimulus amplitude was 0.5 na. (3) The stimulus elicited a plateau potential in culture, even when spikes were blocked by treatment with TTX. Parts (2) and (3) are from the same neuron. (B) A plateau potential in B31/B32 when it is cocultured with other neurons. (1) B31/B32 was cocultured with B61. In culture, B31/B32 becomes electrically coupled to most neurons with which it is cocultured, reflecting the massive coupling of B31/B32 in situ. B61 was hyperpolarized to 2100 mv (for a recording in which B61 was not hyperpolarized, see Supplemental Data, Figure S4). The hyperpolarization prevented B61 from depolarizing and firing in response to the stimulus to B31/B32, and thereby depolarizing B31/B32 via the coupling between these neurons. The figure shows that depolarizing B31/B32 (0.8 na for 5 s) in these conditions still elicits a plateau potential and persistent activity, indicating self-excitatory EPSPs are still created by B31/B32 when another target neuron is present. (2) Two B31/B32 neurons were cocultured. The neurons were electrically coupled (note electrical EPSPs in the top trace in response to firing in the bottom trace). In many cases, the coupling causes depolarization of one neuron to drive plateau potentials in the other neuron. However, in some cases, such as that shown, depolarization (2 na, 5 s) caused a plateau potential and persistent activity in the stimulated B31/B32, but not in the coupled neuron, which was at 260 mv. These traces show that a B31/B32 neuron prefers to develop a synapse onto itself rather than develop synapses onto a second B31/B32 with which it is cocultured.

5 Autaptic Muscarinic Excitation 483 Isolated neurons often display autapses [1, 26, 27], and autapses can cause epileptiform bursts similar to plateau potentials in hippocampal neurons grown in isolation [27], raising the possibility that autaptic plateau potentials in B31/B32 are an artifact of culturing in isolation. A number of points argue against this possibility. First, previous studies that examined isolated Aplysia neurons in culture, including cholinergic and cholinoceptive neurons, did not find neurons synapsing upon themselves [16 18, 31]. Second, culturing in isolation other buccal ganglia cholinergic, cholinoceptive neurons did not lead to plateau potentials. Third, B31/B32s developed a plateau potential even when not isolated (Figure 4B). Fourth, the kinetics and the amplitude of the plateau potential in culture were similar to those of the persistent activity of B31/ B32 in situ. B31/B32s in culture, but not in situ, had overshooting action potentials. In situ, B31/B32 is electrically coupled to many neurons [6, 9], and coupling shunts currents arising from the axon spikes. Isolation increased the input resistance of B31/ B32s by almost an order of magnitude, allowing axon spikes to invade the soma. The increased input resistance will also make the autapses more effective in driving B31/B32 activity. An increase in spike amplitude in culture might also be the result of a change in ion channel distribution. Sodium channels might be inserted closer to the soma in culture than in situ. Studies in other CPG neurons in culture have also shown increases in spike amplitude [25]. Contribution of the B31/B32 Autapse to the In Situ Persistent Activity B31/B32 neurons are electrically coupled to many neurons, particularly to B63, which monosynaptically excites B31/ B32s via fast and slow EPSPs (Figure 2A) [8, 9]. Block of B31/B32 activity by TTX and pirenzepine suggested that the actions of presynaptic neurons accounts for persistent B31/ B32 activity [8]. We have now shown that a voltage-dependent muscarinic autapse contributes to B31/B32 activity. Firing a B31/B32 axon at its physiological frequency depolarizes the B31/B32 soma even when B63 is not active (Figures 2B). In addition, the B31/B32 morphology is consistent with the presence of an autapse (Figure 2C and Supplemental Data). B31/B32s have many neurites with varicosities close to the soma. Connections onto and from B31/B32s have been explored in a number of laboratories. Chemical synaptic connections from B31/B32s to other neurons have not yet been found, suggesting that these varicosities may function only to release transmitter onto B31/B32s. Definitive morphological evidence of autaptic transmission in B31/B32s could be provided by electron microscopy. However, it is difficult to demonstrate a functional chemical synapse via morphology in Aplysia, because Aplysia synapses do not have strong postsynaptic densities. In addition, somatic membranes of Aplysia neurons are highly invaginated, making it difficult to identify a contact. The B31/B32 autapse could not contribute to the initiation of the cell s activity, because B31/B32s fires only after they have been substantially depolarized. The B31/B32 autapse would contribute to maintaining the depolarization. B63 and the B31/B32 autapse at 240 mv produce comparable depolarizations, suggesting that their contributions to the full depolarization to 220 mv are also comparable. Why do B31/B32s require the autapse? If B63 firing alone initiates the B31/B32 depolarization, and B63 and other neurons fire throughout the depolarization, why could they not drive the depolarization without the autapse? The autapse may contribute to the ability of B31/B32s to repolarize quickly. B31/B32s becomes depolarized over a number of seconds (see Figure 1), but repolarize within 0.5 s [28]. B31/B32 depolarization is driven by the slow activation kinetics (over 3 s) of the voltage-dependent inward current activated by muscarinic transmission [11]. B31/B32 repolarization is caused by B64, which causes both presynaptic [32] and postsynaptic inhibition [28] of protraction-phase neurons. The autapse in B31/ B32s means that B64 inhibition is simultaneously presynaptic and postsynaptic, i.e., it both changes the B31/B32 membrane potential and prevents release of transmitter onto itself. Block of ACh release coupled with the change in membrane potential might facilitate a rapid repolarization. An autapse is also found in another prominent Aplysia buccal ganglia neuron. The B4-B5 neurons display a selfinhibitory cholinergic synapse [33] likely to play a role in biasing the output of the CPG toward egestion, rather than ingestion [34]. In a Clione CPG, a GABAergic excitatory interneuron drives other neurons via slow EPSPs, and also excites itself. The possibility that self-excitation is via an autapse, or via electrical coupling to neurons that synapse onto the interneuron, was not resolved. In Clione the self-excitation prolongs neural activity, but does not underlie a plateau potential [35]. Autapses are not uncommon in the nervous system of mammals [1]. Some studies have examined their functional role in information processing [1, 36 42]. Their contribution to synaptic function might be more extensive than that found to date, but difficult to detect, as is the contribution of the muscarinic autapse in B31/B32. As in B31/B32, autapses in the mammalian central nervous system (CNS) might function with pharmacologically similar synapses from presynaptic neuron with lower firing thresholds, making it difficult to isolate their effects. A persistent response to a brief stimulus is a common feature in many neural systems [2]. Such responses might be a mechanism for short-term memory storage [2]. Models have shown that autapses can account for features of a persistent response to a brief stimulus [43]. These models contained traditional EPSPs, and did not consider the possibility that an autapse could also be based on voltage-dependent synaptic transmission, which might then activate currents similar to those underlying endogenous plateau potentials. Autapsebased plateau potentials could have a major role in neural processing in the CNS. In most cases, biophysical mechanisms underlying persistent neural activity have not been explored. Phenomena similar to those in B31/B32 could be widespread, and they await documentation in additional circuits. Supplemental Data Supplemental Data include Supplemental Experimental Procedures, Results, and Discussion, four figures, and one movie and can be found with this article online at S (09) Acknowledgments We thank Samuel Schacher for advice and help in establishing the tissue culture techniques, Leonard J. Cleary and Alexander Perelman for advice on morphology, and John Koester and Alon Korngreen for comments on the manuscript. The research was supported by grant 420/06 from the Israel

6 Current Biology Vol 19 No Science Foundation. All authors contributed to planning and interpreting the experiments. R.S. and I.H. performed in situ intracellular recordings. R.S. performed the microscopy. N.M. and R.S. prepared the cell cultures and recorded from cultured neurons. A.J.S. wrote the manuscript. Received: August 21, 2008 Revised: January 21, 2009 Accepted: January 22, 2009 Published online: March 5, 2009 References 1. Ikeda, K., and Bekkers, J.M. (2006). Autapses. Curr. Biol. 16, R Major, G., and Tank, D. (2004). Persistent neural activity: Prevalence and mechanisms. Curr. Opin. Neurobiol. 14, Hurwitz, I., Cropper, E.C., Vilim, F.S., Alexeeva, V., Susswein, A.J., Kupfermann, I., and Weiss, K.R. (2000). Serotonergic and peptidergic modulation of the buccal mass protractor muscle (I2) in Aplysia. J. Neurophysiol. 84, Hurwitz, I., Goldstein, R.S., and Susswein, A.J. (1994). Compartmentalization of pattern-initiation and motor functions in the B31 and B32 neurons of the buccal ganglia of Aplysia californica. J. Neurophysiol. 71, Hurwitz, I., Neustadter, D., Morton, D.W., Chiel, H.J., and Susswein, A.J. (1996). Activity patterns of the B31/B32 pattern initiators innervating the I2 muscle of the buccal mass during normal feeding movements in Aplysia californica. J. Neurophysiol. 75, Susswein, A.J., and Byrne, J.H. (1988). Identification and characterization of neurons initiating patterned neural activity in the buccal ganglia of Aplysia. J. Neurosci. 8, Marder, E. (1991). Plateaus in time. Curr. Biol. 1, Dembrow, N.C., Jing, J., Brezina, V., and Weiss, K.R. (2004). A specific synaptic pathway activates a conditional plateau depolarization underlying protraction phase in the Aplysia feeding central pattern generator. J. Neurosci. 24, Hurwitz, I., Kupfermann, I., and Susswein, A.J. (1997). Different roles of neurons B63 and B34 that are active during the protraction phase of buccal motor programs in Aplysia californica. J. Neurophysiol. 78, Susswein, A.J., Hurwitz, I., Thorne, R., Byrne, J.H., and Baxter, D.A. (2002). Mechanisms underlying fictive feeding in Aplysia: Coupling between a large neuron with plateau depolarizations activity and a spiking neuron. J. Neurophysiol. 87, Hurwitz, I., Ophir, A., Korngreen, A., Koester, J., and Susswein, A.J. (2008). Currents contributing to decision-making in neurons B31/B32 of Aplysia. J. Neurophysiol. 99, Hurwitz, I., Kupfermann, I., and Weiss, K.R. (2003). Fast synaptic connections from CBIs to pattern generating interneurons in Aplysia: Initiation and modification of buccal motor programs. J. Neurophysiol. 89, Malkinson, G., Fridman, Z.M., Kamber, D., Dormann, A., Shapira, E., and Spira, M.E. (2006). Calcium-induced exocytosis from actomyosindriven, motile varicosities formed by dynamic clusters of organelles. Brain Cell Biol. 35, Hatada, Y., Wu, F., Sun, Z.Y., Schacher, S., and Goldberg, D.J. (2000). Presynaptic morphological changes associated with long-term synaptic facilitation are triggered by actin polymerization at preexisting varicositis. J. Neurosci. 20, RC Zhang, H., Wainwright, M., Byrne, J.H., and Cleary, L.J. (2003). Quantitation of contacts among sensory, motor, and serotonergic neurons in the pedal ganglion of Aplysia. Learn. Mem. 10, Camardo, J., Proshansky, E., and Schacher, S. (1983). Identified Aplysia neurons form specific chemical synapses in culture. J. Neurosci. 3, Kleinfeld, D., Parsons, T.D., Raccuia-Behling, F., Salzberg, B.M., and Obaid, A.L. (1990). Foreign connections are formed in vitro by Aplysia californica interneuron L10 and its in vivo followers and non-followers. J. Exp. Biol. 154, Kleinfeld, D., Raccuia-Behling, F., and Chiel, H.J. (1990). Circuits constructed from identified Aplysia neurons exhibit multiple patterns of persistent activity. Biophys. J. 57, Panchin, Y.V., Arshavsky, Y.I., Selverston, A., and Cleland, T.A. (1993). Lobster stomatogastric neurons in primary culture. I. Basic characteristics. J. Neurophysiol. 69, Saver, M.A., Wilkens, J.L., and Syed, N.I. (1999). In situ and in vitro identification and characterization of cardiac ganglion neurons in the crab, Carcinus maenas. J. Neurophysiol. 81, Straub, V.A., Staras, K., Kemenes, G., and Benjamin, P.R. (2002). Endogenous and network properties of Lymnaea feeding central pattern generator interneurons. J. Neurophysiol. 88, Syed, N.I., Bulloch, A.G., and Lukowiak, K. (1990). In vitro reconstruction of the respiratory central pattern generator of the mollusk Lymnaea. Science 250, Turrigiano, G., Abbott, L.F., and Marder, E. (1994). Activity-dependent changes in the intrinsic properties of cultured neurons. Science 264, Turrigiano, G., LeMasson, G., and Marder, E. (1995). Selective regulation of current densities underlies spontaneous changes in the activity of cultured neurons. J. Neurosci. 15, Turrigiano, G.G., and Marder, E. (1993). Modulation of identified stomatogastric ganglion neurons in primary cell culture. J. Neurophysiol. 69, Bekkers, J.M., and Stevens, C.F. (1991). Excitatory and inhibitory autaptic currents in isolated hippocampal neurons maintained in cell culture. Proc. Natl. Acad. Sci. USA 88, Segal, M.M. (1991). Epileptiform activity in microcultures containing one excitatory hippocampal neuron. J. Neurophysiol. 65, Hurwitz, I., and Susswein, A.J. (1996). B64, a newly identified central pattern generator element producing a phase switch from protraction to retraction in buccal motor programs of Aplysia californica. J. Neurophysiol. 75, Gardner, D. (1977). Interconnections of identified multiaction interneurons in buccal ganglia of Aplysia. J. Neurophysiol. 40, Morton, D.W., and Chiel, H.J. (1993). The timing of activity in motor neurons that produce radula movements distinguishes ingestion from rejection in Aplysia. J. Comp. Physiol. [A] 173, Rayport, S.G., and Schacher, S. (1986). Synaptic plasticity in vitro: Cell culture of identified Aplysia neurons mediating short-term habituation and sensitization. J. Neurosci. 6, Hurwitz, I., Susswein, A.J., and Weiss, K.R. (2005). Transforming tonic firing into a phasic output in the Aplysia feeding system: Presynaptic inhibition of a command-like neuron (CBI-2) by a CPG element (B64). J. Neurophysiol. 93, Gardner, D. (1977). Voltage-clamp analysis of a self-inhibitory synaptic potential in the buccal ganglia of Aplysia. J. Physiol. 264, Ye, H., Morton, D.W., and Chiel, H.J. (2006). Neuromechanics of multifunctionality during rejection in Aplysia californica. J. Neurosci. 26, Norekian, T.P. (1999). GABAergic excitatory synapses and electrical coupling sustain prolonged discharges in the prey capture neural network of Clione limacina. J. Neurosci. 19, Bacci, A., and Huguenard, J.R. (2006). Enhancement of spike-timing precision by autaptic transmission in neocortical inhibitory interneurons. Neuron 49, Bacci, A., Huguenard, J.R., and Prince, D.A. (2003). Functional autaptic neurotransmission in fast-spiking interneurons: A novel form of feedback inhibition in the neocortex. J. Neurosci. 23, Friedman, D., and Strowbridge, B.W. (2000). Functional role of NMDA autoreceptors in olfactory mitral cells. J. Neurophysiol. 84, Nicoll, R.A., and Jahr, C.E. (1982). Self-excitation of olfactory bulb neurones. Nature 296, Smith, T.C., and Jahr, C.E. (2002). Self-inhibition of olfactory bulb neurons. Nat. Neurosci. 5, Tamás, G., Buhl, E.H., and Somogyi, P. (1997). Fast IPSPs elicited via multiple synaptic release sites by different types of GABAergic neurone in the cat visual cortex. J. Physiol. 500, Tamás, G., Buhl, E.H., and Somogyi, P. (1997). Massive autaptic selfinnervation of GABAergic neurons in cat visual cortex. J. Neurosci. 17, Seung, H.S., Lee, D.D., Reis, B.Y., and Tank, D.W. (2000). The autapse: A simple illustration of short-term analog memory storage by tuned synaptic feedback. J. Comput. Neurosci. 9,