INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Rhythmic behaviors such as running, walking, and breathing are thought to be controlled by neural networks called central pattern generators (CPGs) (Getting 1989; Marder and Calabrese 1996; Pearson 1993; Selverston and Moulins 1985). While anatomically determined by their synaptic connections, it is now clear that a single CPG network can be reconfigured to produce a variety of motor outputs under different modulatory conditions (Harris-Warrick and Marder 1991; Marder and Calabrese 1996; Marder and Weimann 1992). The cellular and ionic mechanisms by which neuromodulators reconfigure circuits are now a major field of research (Katz 1999).

The crustacean stomatogastric nervous system (STNS) is a very useful model system for examining the properties and modulation of CPGs; it has been studied most extensively in lobsters and crabs (Harris-Warrick et al. 1992a; Selverston and Moulins 1987). The stomatogastric ganglion (STG) contains 30 neurons that control the rhythmic movements of the foregut to grind and filter food (Johnson and Hooper 1992). All the neurons are physiologically identifiable, and all the synaptic connections between them are known. The cells of the STG make up two CPG networks that control the gastric mill and pyloric regions of the foregut. We are studying the pyloric network, which contains 14 neurons composed of the anterior burster (AB), two pyloric dilators (PD), the lateral pyloric (LP), the inferior cardiac (IC), the ventricular dilator (VD), and eight pyloric constrictors (PY).

Numerous neuromodulators in the STNS alter the output from the pyloric and gastric mill CPGs (Harris-Warrick et al. 1992b; Marder 1991; Marder et al. 1997). Our work has focused on the mechanisms by which the amines dopamine (DA), serotonin (5-HT), and octopamine (OCT) reconfigure the pyloric network (Harris-Warrick et al. 1993, 1995a,b, 1998). These amines both influence the intrinsic firing properties of the neurons and change the strength of synaptic connections in this network (Ayali and Harris-Warrick 1999; Flamm and Harris-Warrick 1986a,b; Harris-Warrick et al. 1995a,b; Johnson and Harris-Warrick 1990; Johnson et al. 1993-1995; Kloppenburg et al. 1999). A general principle that has emerged from our previous research is that a particular modulator affects a neural network at a variety of sites and in a variety of ways (Harris-Warrick et al. 1998). In different STG neurons, amines directly modify several ionic currents in different ways, including the transient K⁺ current (I_A) (Harris-Warrick 1993; Harris-Warrick et al. 1993; Kloppenburg et al. 1999), the calcium-dependent outward current [I_O(Ca)] (Kiehn and Harris-Warrick 1992), the hyperpolarization-activated inward current (I_h) (Harris-Warrick et al. 1995b; Kiehn and Harris-Warrick 1992) and the voltage-dependent calcium current (I_Ca) (Kloppenburg et al. 2000; Zhang and Harris-Warrick 1995). By modifying cellular currents, amines can change postinhibitory rebound, plateau potential capability, oscillatory properties of pyloric cells, and synaptic transmission.

We have been studying the role of I_A in the electrical activity of pyloric neurons. I_A is a rapidly activating and inactivating K⁺ current whose functions include modulating synaptic transmission and regulating repetitive spiking, postinhibitory rebound and cycle frequency (Connor and Stevens 1971; Harris-Warrick et al. 1995a,b; Tierney and Harris-Warrick 1992). Hartline (1979) suggested that the sequence and phasing of firing of pyloric neurons in the rhythmic pyloric motor pattern could, in part, be determined by cell-specific differences in expression of I_A. In support of Hartline's hypothesis, we have found that there are different amounts of I_A in the different pyloric cell types (Baro et al. 1997). Reducing I_A with low concentrations of 4-aminopyridine (4-AP) changes overall cycle frequency and phasing among the cells in the pyloric network as well as differentially affecting an individual cell's spike frequency, spikes per burst, and burst duration (Tierney and Harris-Warrick 1992). We have previously shown that DA modulates I_A in the PD, LP, and PY neurons, but in different ways. In the PY and LP cells, DA decreases I_A, leading to excitation and phase advances in firing (Harris-Warrick et al. 1995a,b). In contrast, in the PD cell, DA increases I_A, thus inhibiting and phase delaying its activity (Kloppenburg et al. 1999).

In this paper we examine the role of DA, 5-HT, and OCT in the modulation of I_A in the AB, VD, and IC cells of the pyloric circuit. These studies complete our survey of amine modulation of I_A in the pyloric network. Our results indicate that each neuron shows a unique pattern of I_A modulation by these amines, and this can help explain how the amines shape the pyloric motor pattern.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Materials

California spiny lobsters, Panulirus interruptus, were obtained from Don Tomlinson Commercial Fishing (San Diego, CA) and maintained for up to 4 wk in artificial seawater at 16°C. Chemicals were obtained from Sigma Chemical, St. Louis, MO.

Dissection

Lobsters were packed in ice for 30 min before dissection. The foregut was removed, and the STNS was pinned out in a silicone elastomer (Sylgard)-lined Petri dish as described previously (Selverston et al. 1976). The STG was desheathed, and a petroleum jelly (Vaseline) well was placed around it. The STNS was covered with Panulirus saline (composition, in mM: 479 NaCl, 12.8 KCl, 13.7 CaCl₂, 3.9 Na₂SO₄, 10.0 MgSO₄, 2.0 glucose, 11.1 Tris base, and 5.1 maleic acid, pH 7.35) (Mulloney and Selverston 1974), and the STG was continuously perfused with 15°C, oxygenated saline at 3 ml/min.

Cell and channel isolation

Using standard intracellular techniques (3 M KCl-filled microelectrodes, 10-25 M), cells were identified by matching the firing patterns of intracellular recordings to extracellular recordings of axons innervating identified muscles, and by the timing and pattern of intracellular spikes. Synaptic isolation was accomplished by photoinactivating cholinergic cells or neurons electrically coupled to the cell being studied (Miller and Selverston 1979), and by the addition of picrotoxin (PTX, 5 × 10⁶ M) to block glutamatergic synapses. Currents other than I_A were greatly reduced by the addition of tetrodotoxin (TTX, 10⁷ M, to block voltage-gated Na⁺ currents), cadmium (CdCl₂, 0.2 mM, to block Ca²⁺ and Ca²⁺-activated currents), cesium (CsCl, 7.5 mM, to block the hyperpolarization-activated inward current), and tetraethyl ammonium (TEA chloride, 20 mM, to block rectifying voltage-gated K⁺ currents).

Voltage clamp

Voltage-clamp protocols were carried out using an Axoclamp 2A amplifier controlled by pClamp software (Axon Instruments) running on a PC. Three voltage protocols were used to measure I_A. For the voltage activation protocol, the cells were held at 40 mV; inactivation was removed by a 200-ms step to 90 mV followed by an activating 500-ms step to between 40 mV and +20 or +30 mV in 10-mV incrementing steps. The data were leak subtracted using a P/8 protocol with steps opposite in sign to the activation protocol. A control protocol for activation of non-I_A currents was used that was the same as the activation, but without the deinactivating step to 90 mV. The control protocol currents were digitally subtracted from the activation protocol currents to produce the isolated I_A activation curves. For the voltage inactivation protocol, the cell was held at 40 mV; inactivation was incrementally removed by 200-ms steps to between 90 and 20 mV in 10-mV steps. On each trial, the cell was then stepped to +30 mV for 500 ms to measure the degree of removal of inactivation. The inactivation steps were leak subtracted as described above.

Because I_A activates and inactivates rapidly in the AB and VD neurons (Baro et al. 1997), we minimized the capacitance coupling artifact at the beginning of the voltage steps. We used short shank, 7- to 15-M electrodes made from thick-walled 1.0-mm glass. The electrodes were kept at low penetration angles (usually >90° apart), and a grounded metal plate was inserted between them to reduce capacitative coupling between the electrodes. The electrodes were inserted in the cell as far apart as possible. The headstage of the current injecting electrode was insulated and grounded. These techniques improved our ability to measure very rapidly rising and falling currents.

Data were analyzed by measuring peak currents at each voltage step using Clampfit protocols (Axon Instruments). Peak currents were converted to conductance [using an estimated V_rev of 86 mV (Hartline and Graubard 1992)], and the conductance-voltage relationship was fit using Kaleidagraph 3.0 (Synergy Software) to a Boltzmann equation of the following form
where (g_max) is the maximal conductace, V_act is the voltage of half activation, s is the slope of activation, n = 3 for the activation relation (Baro et al. 1997; Kloppenburg et al. 1999; Willms et al. 1999). Inactivation peak currents during the +30-mV voltage step were plotted against prestep amplitude and fit to a first-order Boltzmann equation (n = 1 in the equation above) to produce values for voltage of half inactivation (V_inact) and slope of inactivation.

Percentage changes in g_max during an amine were calculated from paired recordings before and during amine administration. The control values in Tables 1-3 give only the first control value before any amine administration in each preparation; thus our reported percent changes may differ from calculations made directly from the table. Calculation of steady-state currents was done using Kaleidagraph. The activation function (which is a cubic relation in the Boltzmann relation) was multiplied by the inactivation function to yield a Hodgkin-Huxley-like m³h product.

Amine application

DA (10⁴ M), 5-HT (10⁵ M), and OCT (10⁵ M) were dissolved in saline immediately before use and individually bath applied for 5 min. I_A was measured before the amine was applied, after 5 min of application and after a minimum 30-min wash out, before the next amine was applied. For any given experiment, the order of amine application was randomized and not every amine was applied to every preparation, due to cell death, or damage. Data were used only if the amine effect was fully reversed during the wash out.

Statistics

Tests for statistical significance were carried out using ANOVA and subsequent protected t-tests. Data analysis which produced t or F values with a P < 0.05 were accepted as statistically significant. Means are presented as means ± SE.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Effects of DA, OCT, and 5-HT on the AB, IC, and VD neurons in the intact pyloric network

Figure 1 shows simultaneous intracellular recordings from the AB, IC, and VD neurons during the pyloric pattern with descending modulatory inputs intact. In the control condition, the neurons fire in rhythmic bursts, each with a characteristic phasing and intensity relative to the others. The AB neuron is the primary pacemaker for the pyloric rhythm, generating the highest frequency rhythmic oscillations and inhibiting all the other pyloric neurons except the PDs. The VD and IC neurons synapse on fewer pyloric neurons and play less prominent roles in setting the phasing and cycle frequency of the pyloric rhythm under normal experimental conditions.

View larger version (22K): [in this window] [in a new window]   Fig. 1. The effects of dopamine (DA), octopamine (OCT), and serotonin (5-HT) on the anterior burster (AB), inferior cardiac (IC), and ventricular dilator (VD) neurons in the pyloric circuit with descending connections intact. All intracellular recordings were made from the same preparation before amine administration (control) or after 5 min of amine administration. There was a 30-min wash between amines. The dashed lines show the beginning of each burst with the 1st spike of the AB neuron, to demonstrate the phase changes of the IC and VD neurons.

As described previously (Ayali and Harris-Warrick 1998; Flamm and Harris-Warrick 1986a,b), DA, OCT, and 5-HT all modify the ongoing pyloric rhythm in characteristic ways. During DA application, the cycle frequency slows, and some of the neurons, including the VD neuron, reduce their average firing frequency, are phase-delayed, or stop firing altogether, due to direct inhibition by DA (Fig. 1, DA). In contrast, most of the neurons are directly excited by DA. Both the AB and IC neurons increased their number of spikes per burst, and the IC neuron was phase-advanced in its bursting in the cycle (Fig. 1, DA; to make this phase advance more clear, dashed lines show the first spike of each AB burst relative to the other neurons in control and DA conditions). OCT has a more subtle effect on the pyloric rhythm, with a small decrease in cycle frequency (Fig. 1, OCT). The AB, IC, and VD neurons all show variable changes in spikes per burst relative to the control condition; often there is a small increase in spike frequency, although this was not seen in the experiment shown in Fig. 1. The VD usually shows more rapid recovery from inhibition (Fig. 1, OCT). Finally, 5-HT exerts a modest increase in cycle frequency and, like DA, excites the AB and IC neurons while inhibiting the VD neuron (Fig. 1, 5-HT). In some experiments, such as that shown in Fig. 1, the IC fires nearly tonically, with only brief interruptions due to AB/PD inhibition.

Voltage-clamp measurements of I_A

AB NEURON. Figure 2 shows I_A traces from a voltage-clamped AB neuron in response to a series of depolarizing voltage clamp steps as described in METHODS, before, during, and after bath application of the three amines. DA causes a significant reduction of I_A at all voltages, which is fully reversible after 20-min wash with normal saline. Boltzmann analysis of averaged voltage activation data (Fig. 3A, Table 1) showed that DA () significantly reduced the maximal conductance (g_max) by 39% (t = 6.1, P < 0.05) relative to control (). DA also shifted the V_act by 6 mV in the depolarizing direction (assuming n = 3 in the Boltzmann relation); this resulted in a 3-mV depolarizing shift in the voltage for half-maximal activation of the current. The slope of the voltage activation curve was not significantly affected by DA. DA shifted the voltage dependence of steady-state inactivation, V_inact, by 13 mV in the depolarizing direction, again with no significant change in the slope of the inactivation relation (Table 1). Both of these shifts in voltage dependence were significant (t = 3.1, P < 0.05). These results show that following hyperpolarization at the end of a burst, less I_A will be activated in the AB during the depolarization to the next burst, allowing the burst to occur sooner. This will accelerate the cycle frequency of the isolated AB neuron, as previously reported (Ayali and Harris-Warrick 1999).

View larger version (35K): [in this window] [in a new window]   Fig. 2. Voltage-clamp examples of the effect of DA, OCT, and 5-HT on IA in the AB neuron. The AB neuron used in the DA example was stepped in 10-mV increments from 50 to +40 mV and shows a reduction in IA. The OCT and 5-HT examples are from a different AB neuron that was stepped from 40 to +30 mV and show no change in IA.

View larger version (18K): [in this window] [in a new window]   Fig. 3. Analysis of the effects of DA on IA in the AB neuron. A: g/V curves. Conductance at each voltage step for each cell was normalized by converting it to g/gmax (control). Data at each voltage step were averaged, and voltage activation and inactivation curves for control () and DA () were generated using the Boltzmann equation. B: IA "window currents" for control and DA were calculated by multiplying the activation curve with the inactivation curve. The curves show the voltage range over which IA is tonically active under control and DA conditions.

IC NEURON. Figure 4 shows the voltage activation currents in the IC neuron before, during, and after bath application of DA, OCT, and 5-HT. As can be seen from the figure, both DA and 5-HT modestly reduced the amplitude of I_A in this neuron. There were no significant effects on the kinetics of activation or inactivation of the current. Boltzmann analysis of the effect of DA on the voltage dependence of activation and inactivation is shown in Fig. 5A (see also Table 2). Dopamine significantly reduced g_max by 14% (t = 3.2, P < 0.05). As in the AB neuron, DA also shifted the V_act by 4 mV in the depolarizing direction (assuming a 3rd-order relation), resulting in a 3-mV shift in the voltage for half activation of the current. The slope of the voltage activation curve was unchanged by DA. V_inact was also shifted by 4 mV in the depolarizing direction, again with no effect on the slope (Table 2). As was seen in the AB neuron, the net effect of these reductions in I_A is to allow the IC neuron to recover from inhibition significantly more rapidly, phase advancing its activity in the pyloric network and allowing it to fire at higher frequencies (Fig. 1). In addition the steady-state I_A is also reduced by DA (Fig. 5B). As with the AB neuron, the maximal window current was reduced by 15%, and its peak voltage was shifted from 55 mV under control conditions to 50 mV during dopamine. These results show that during DA, I_A will contribute less tonic outward current to set the resting potential of the IC neuron.

View larger version (34K): [in this window] [in a new window]   Fig. 4. Voltage-clamp examples of the effect of the amines DA, OCT, and 5-HT on IA in the IC neuron. The examples were all taken from the same IC neuron and show IA reduction during both DA and 5HT. Voltage steps in this cell were in 10-mV increments from 40 to +30 mV.

View larger version (22K): [in this window] [in a new window]   Fig. 5. Analysis of the effects of DA and 5-HT on IA in the IC neuron. A and C: g/V curves were obtained as described in the legend to Fig. 3. , control; , DA (A) or 5-HT (C). B and D: IA "window currents" for control and DA (B) or control and 5-HT (D) were calculated by multiplying the respective activation and inactivation curves together.

View larger version (13K): [in this window] [in a new window]   Fig. 6. Voltage-clamp examples of the effect of DA, OCT, and 5-HT on IA in the VD neuron. The examples were taken from the same VD neuron and show no modulation of IA for any amine. Voltage steps in this cell were in 10-mV increments from 40 to +40 mV.

VD NEURON. In the VD neuron, I_A is much smaller and inactivates about 10 times more rapidly than in the other pyloric neurons (Baro et al. 1997) [compare Fig. 6 (VD) to Fig. 2 (AB) or Fig. 4 (IC)]. Figure 6 and Table 3 show our results for the effects of the amines on I_A parameters in the VD neuron. We expected that DA would enhance I_A, since it hyperpolarizes the VD neuron in a way that is similar to the PD neuron, where DA enhances I_A (Kloppenburg et al. 1999). Some VD neurons showed slight increases in I_A with DA, but others showed slight decreases, and there was no statistically significant net effect of DA on the VD neuron. The other amines also had no reproducible effect on I_A in this neuron.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It is now a general principle that both vertebrate and invertebrate motor pattern generator networks are regulated by modulatory inputs using monoamines, peptides, and other transmitters (Calabrese 1998; Harris-Warrick and Marder 1991; Marder and Calabrese 1996). These transmitters reconfigure the network by altering the intrinsic electrophysiological properties of the component neurons and changing the strength of their synaptic connections (Harris-Warrick 1988). The ionic currents affected by these neuromodulators in motor networks have been more difficult to analyze and have only been studied in detail in a few preparations (Grillner 1999; Harris-Warrick et al. 1998; Kiehn and Katz 1999; Kleinhaus and Angstadt 1995). For example, in the lamprey spinal generator for swimming, serotonin enhances and prolongs motoneuron bursts by decreasing an apamin-sensitive calcium-activated potassium current (Wallén et al. 1989), and serotonin and dopamine depress reticulospinal synaptic transmission by reducing presynaptic HVA-type calcium currents (Grillner et al. 1998; Wikstrom et al. 1999). In addition, substance P enhances dorsal cell excitability in part by decreasing a 4-AP-sensitive current (Parker et al. 1997). In the leech heartbeat generator, the peptide FMRFamide accelerates the cycle frequency in part by activating a very slow potassium current (Nadim and Calabrese 1997) and calcium-dependent inward currents (Schmidt et al. 1995). In the lobster gastric mill network, we showed previously that serotonin can evoke bistability in a motor neuron by a combination of enhanced hyperpolarization-activated inward current (I_h), decreased I_K(Ca) (Kiehn and Harris-Warrick 1992), enhanced I_Ca with consequent enhancement of the calcium-activated nonselective current, I_CAN and decreasing an additional potassium current (Zhang and Harris-Warrick 1995). In the lobster pyloric network, a number of peptides excite selected motor neurons by a common enhancement of an inward current with maximal amplitude near the resting potential (Swenson and Marder 2000). This paper continues our work toward understanding the ionic mechanisms by which dopamine, serotonin, and octopamine reconfigure the pyloric network. Previous work has shown that many ionic currents are modulated by these amines (Harris-Warrick et al. 1998) and that modulation of I_A plays a particularly important role.

Hartline (1979) first suggested that cell-specific differences in I_A could contribute to neuronal differences in spike frequency, postinhibitory rebound, and neuronal phasing in the pyloric rhythm. Following his reasoning, we hypothesized that modulatory increases in I_A should lead to hyperpolarization of the resting potential, decreases in oscillatory and spike frequency and phase delay, while decreases in I_A should do the opposite. Our previous research has shown that I_A is a primary target of amine modulation in pyloric neurons. Consistent with Hartline's ideas, DA decreases I_A in the LP and PY cells, and this leads to depolarization and higher spiking rates as well as increases in the rate of postinhibitory rebound and phase advances in the pyloric rhythm (Harris-Warrick et al. 1995a,b). In contrast, Kloppenburg et al. (1999) showed that DA increases I_A in the PD neurons, leading to hyperpolarization and phase delay in the pyloric rhythm. Our goal in this work was to determine whether DA, OCT, and 5-HT modulate I_A in the remaining pyloric neurons, AB, IC, and VD.

AB neuron

Since the AB neuron is the major pacemaker of the pyloric circuit, its modulation should have important consequences for the rhythm of the circuit. DA decreased the maximal conductance of I_A in the AB neuron by 40%. This should lead to enhanced oscillating amplitude, and increases in cycle frequency and spike rate, which are in fact seen with synaptically isolated AB neurons (Ayali and Harris-Warrick 1999; Flamm and Harris-Warrick 1986b). In the synaptically isolated, silent AB, reduction of I_A with 4-AP is sufficient to evoke bursting (Harris-Warrick and Johnson 1987). In the intact network, however, the cycle frequency usually slows somewhat under DA (Fig. 1). As demonstrated by Ayali and Harris-Warrick (1999), this is explained by the fact that the PDs are electrically coupled to the AB, and while DA enhances AB oscillatory properties, it simultaneously hyperpolarizes the PDs. The PDs essentially act as a current sink to slow the AB oscillations. However, the excitation of AB is still seen by the increased amplitude of its slow-wave oscillation during DA (Fig. 1).

DA also shifted the V_act and V_inact in the depolarizing direction in the AB neuron (Fig. 3, Table 1). This has two major consequences. First, it combines with the reduction in g_max to reduce the I_A activated in the critical subthreshold voltage range during the rising phase of the oscillation. As a consequence, the AB neuron can oscillate more rapidly in isolation (Ayali and Harris-Warrick 1999; Flamm and Harris-Warrick 1986b). Second, it shifts the window of "tonic" I_A in the depolarizing direction, leading to a reduction in tonic outward current. Without the accompanying shift in V_inact, the cell might be in danger of losing all its I_A as the membrane depolarized. Thus the shift in V_inact could be a protective mechanism to ensure that that there is sufficient I_A available even at depolarized levels of the neuron to help shape the firing properties of the AB neuron.

In the isolated AB cell, 5-HT and OCT also enhance bursting oscillations and increase the spike frequency (Ayali and Harris-Warrick 1999; Flamm and Harris-Warrick 1986b). However, our experiments suggest that this is not due to modulation of I_A and that other currents must be involved in these changes. This conclusion is consistent with earlier current-clamp experiments (Harris-Warrick and Flamm 1987), which concluded that DA, 5-HT, and OCT each enhance AB bursting by different ionic mechanisms.

IC neuron

In the IC neuron, I_A was decreased by both DA and 5-HT. DA and 5-HT depolarize the isolated IC neuron and increase spiking (Flamm and Harris-Warrick 1986b), and this is also seen in the intact preparation (Fig. 1). The IC burst duration was also longer and was phase advanced relative to the AB neuron. These effects are all consistent with a reduction in I_A by DA and 5HT. The IC neuron only synapses onto the VD neuron in the pyloric network, which is already inhibited by DA, so the IC does not play a major role in organizing the DA-induced pyloric rhythm. However, the IC neuron constricts the valve that controls the flow of nutrients between the gastric mill and the pylorus (Johnson and Hooper 1992), so amine modulation of its activity could significantly affect the functioning of the pylorus.

In addition to reducing g_max in the IC neuron, DA shifted both the V_act and V_inact in the depolarized direction while 5-HT shifted the V_inact in the depolarized direction and showed a trend to shift V_act as well. These changes would make it more difficult to activate I_A in the subthreshold range, allowing more rapid postinhibitory rebound. In addition, they shift the I_A window current to a more depolarized voltage and cause the cell to depolarize and start firing. While OCT excites the IC neuron directly (Flamm and Harris-Warrick 1986b), this does not appear to be mediated by changes in I_A.

VD neuron

Previous research (Flamm and Harris-Warrick 1986b) showed that the isolated VD neuron is inhibited by both DA and 5-HT and weakly excited by OCT. Our experiments show that I_A was not modulated by any amine in the VD neuron (Fig. 6, Table 3). I_A is very small and has very rapid kinetics in the VD relative to the other neurons in the circuit (Baro et al. 1997). Thus I_A may not contribute as much to shaping VD firing activity as it does in the other pyloric neurons. Other currents must be better targets for amine modulation in this cell.

Earlier work on the LP, PY, and PD neurons showed I_A as a major target of DA action (Harris-Warrick et al. 1995a,b; Kloppenburg et al. 1999), and indeed, I_A is modified by DA in the AB and IC neurons. The VD neuron is the only pyloric neuron whose I_A is unaffected by DA. 5-HT and OCT tend to have less dramatic effects on pyloric activity than DA and also fewer effects on I_A. The only effects of these amines we have found are 5-HT's reduction of I_A in the IC neuron and a trend for OCT to reduce I_A in the AB neuron. These effects on I_A will contribute, along with other ionic changes, to the changes in neuronal firing properties that the amines evoke. It is, of course, possible that OCT and 5-HT do modify I_A in these neurons, but only in distal regions of the neuropil that are not detectable by our somatic voltage clamp. However, it is clear that the amines must act on other ionic currents in addition to I_A when they alter pyloric neuron activity.