INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Omnipause neurons (OPNs) located near the midline of the pons discharge tonically during fixation and pause during saccades in all directions. OPNs play a pivotal role in saccade generation; their tonic discharge inhibits premotor burst neurons, and the pause allows the burst neurons to fire. Intracellular recordings from OPNs in alert cats (Yoshida et al. 1999) have shown that a pause in spike activity is caused by inhibitory postsynaptic potentials (IPSPs). Analyses of the timing and profiles of IPSPs have suggested that the duration of the pause is controlled by afferents conveying eye-velocity signals from the brain stem burst generator and that the onset of the pause is controlled by input other than the burst generator. One of the most likely structures providing the latter input is the superior colliculus (SC) (for review, see Fuchs et al. 1985). Electrical stimulation of the SC induces monosynaptic excitation followed by suppression of OPN spikes (Kaneko and Fuchs 1982; Raybourn and Keller 1977). The monosynaptic excitation is more efficiently induced from the rostral than from the caudal SC (Paré and Guitton 1994). It was reported that suppression of spikes after SC stimulation had a delay of 6-8 ms (Raybourn and Keller 1977) and that multiple shocks from the SC had a strong additive effect on the duration of suppression (Kaneko and Fuchs 1982). However, the synaptic mechanism of the spike suppression of OPNs following SC stimulation and the number of synaptic linkages between the SC and OPNs remain to be clarified. The present study investigates postsynaptic potentials (PSPs) in OPNs with intracellular recordings in alert cats. We show that SC stimulation elicits disynaptic IPSPs, evidence supporting the idea that SC activity induces the saccadic pause by active inhibition rather than by disfacilitation. Extracellular recordings from a larger sample of OPNs were also made to supplement the intracellular data regarding the efficiency of the rostral versus caudal SC in production of the inhibition.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experiments were performed with three adult cats. All experimental protocols complied with the guidelines of the University of Tsukuba policy on the humane care and use of laboratory animals. The surgical procedures and methods for recording eye movements and painless immobilization of the animal's head have been described previously (Yoshida et al. 1999). A pair of microelectrodes made of Elgiloy wire (Suzuki and Azuma 1976) was chronically implanted in the SC on each side for stimulation. One electrode was placed in the caudal part of the SC where a short-pulse train (10-30 µA) evoked contraversive saccades with a large horizontal component (Fig. 1Ab). The other electrode was placed 2.5-3.5 mm rostral to the first, where stimulation evoked small saccades (Fig. 1Aa). Intra- and extracellular recordings were made from OPNs with glass micropipettes filled with 4 M NaCl solution (Yoshida et al. 1999). Responses of OPNs to single-pulse stimulation (0.2 ms, 200 µA) of the SC were recorded. No eye movement was evoked by this stimulation. Histological reconstructions located the stimulation sites in the intermediate or deep layer in the rostral and caudal of the SC.

View larger version (39K): [in this window] [in a new window]   Fig. 1. A: saccades evoked by short-pulse train stimulation (top: 30 µA, 20 pulses, 400 Hz) of the rostral (a) and caudal SC (b). Hor and Ver, horizontal and vertical eye position traces. B: identification of an omnipause neuron (OPN). , artifacts of superior colliculus (SC) stimulation. C: intracellular records from an OPN showing hyperpolarization (top) associated with a saccade. Radial eye velocity trace (middle) is inverted. D: SC-induced extracellular spike responses (a), intracellular potentials (b), and extracellular control records (c) for the same OPN as in B. E: spike suppression by SC-induced inhibitory postsynaptic potentials (IPSPs). F: reversal of SC-induced IPSPs by chloride injection: before injection (a), during passage of hyperpolarizing current (4 nA) (b) and after cessation of the current (c).

Analyses of excitation and inhibition were obtained from both intracellular and extracellular records. Whereas intracellular responses could be assessed by changes in the membrane potential and the ability to reverse responses by current injection, extracellular assessment required statistical inference. For extracellular records, inhibition was inferred by spike suppression and excitation by an increase in activity. Responses to stimulation were summed over 50-200 trials to obtain a peristimulus spike distribution, and the number of spikes occurring in a window of length T (0.5 ms for excitation, 1.5-3.5 ms for suppression) was counted. The decision concerning whether these counts reflected excitation or suppression was made by determining the deviations of these counts from 95% confidence intervals of activity preceding stimulation (baseline). These confidence intervals were constructed by determining the average number of spikes during baseline, with the variance measure reflecting the sum of squares of deviations from this average evaluated at the same window size used for evaluating excitation or suppression following stimulation.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

OPNs were identified by a saccade-related pause of spikes (Fig. 1B) and, after the cell was impaled, by a hyperpolarization associated with each saccade (e.g., Fig. 1C) (cf. Yoshida et al. 1999). The membrane potential of OPNs was 30-40 mV, at which level spike generation mechanisms were often inactivated. Figure 1D shows extracellular spikes (a) and intracellular responses (b) of the same cell following stimulation of the rostral SC. The initial excitatory postsynaptic potentials (EPSPs; Fig. 1Db) were curtailed by a potential change in the hyperpolarizing direction, which suppressed tonic discharges for a certain period (Fig. 1Da). Spike suppression without preceding excitation was also associated with a membrane hyperpolarization (Fig. 1E). The hyperpolarization (Fig. 1Fa) was reversed to a depolarization during passage of a hyperpolarizing current (Fig. 1Fb) and after injection of Cl ions into the cell (Fig. 1Fc), indicating that the hyperpolarization was the IPSP. The IPSPs were often followed by late EPSPs that increased firing rates (Fig. 1E; see also Fig. 2). The initial EPSPs had latencies of 0.8-1.2 ms (mean, 1.0 ms, n = 4) and were regarded as monosynaptic. The IPSPs were induced with latencies ranging from 1.3 to 1.9 ms (mean, 1.6 ms, for 25 IPSPs in 19 OPNs). Latencies of the late EPSPs were 3-4 ms, a rough estimate because their onset or the transition from IPSPs was gradual.

View larger version (26K): [in this window] [in a new window]   Fig. 2. A: intracellular records from an OPN showing PSPs following stimulation of the rostral and caudal sites of the bilateral SC. B: suppression of extracellular spikes induced by SC stimulation. Spikes in 150 superimposed trials (top) are summed to produce the cumulative spike distribution (middle) and peristimulus time histogram (bottom).

The patterns of PSPs induced by SC stimulation depended on the stimulation site (Fig. 2). Figure 2A shows the PSPs evoked in an OPN following stimulation of the rostral and caudal parts of the SC on both sides. Stimulation of the rostral-left SC elicited IPSPs preceded by monosynaptic EPSPs (Fig. 2Aa), whereas stimulation of the caudal-left SC induced IPSPs without preceding EPSPs (Fig. 2Ab). Table 1A shows the efficiency of the rostral and caudal SC for production of IPSPs in 19 OPNs. Notably, IPSPs were induced from the rostral as well as caudal SC in almost all OPNs examined (9/9 and 13/14 cells, respectively).

Since the sample of intracellular data was small and might have been biased toward larger cells, making it unrepresentative, both excitatory and inhibitory effects were examined in a larger sample of OPNs using extracellular recording (Table 1B). Overall, our extracellular observations complemented what we observed intracellularly (cf. Table 1). Stimulation of the rostral SC was more efficient at eliciting monosynaptic excitation (0.8- to 1.6-ms latencies) than stimulation of caudal sites (63 vs. 31%, ² test, P < 0.0001). Disynaptic inhibition, exhibited by spike suppression (Fig. 2B), began 1.3-2.4 ms following stimulation and lasted 1.5-3.5 ms. Unlike excitation, SC stimulus location had no influence on the probability of observing spike suppression (64 vs. 67%). When early spikes were induced, postspike refractoriness would participate in the succeeding suppression. However, its influence on the incidence of spike suppression was probably minor in this study because SC stimulation elicited IPSPs in most OPNs regardless of whether there was a preceding EPSP. Thus both intra- and extracellular data indicated that the rostral and caudal sites were equally effective for production of OPN inhibition.

Late excitation was observed intra- and extracellularly in a majority of OPNs following stimulation of the rostral and caudal SC, frequently of the SC on both sides (Table 1).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

It has recently been shown that a saccade-related pause of activity in OPNs is initiated by IPSPs (Yoshida et al. 1999). The current study provides direct evidence that signals originating in the SC are capable of inducing the IPSPs. The IPSPs induced by SC stimulation had latencies as short as 1.3-1.9 ms. They are thought to be induced disynaptically because the difference in mean latency between monosynaptic EPSPs (1.0 ms) and IPSPs (1.6 ms) is appropriate for the intercalation of single inhibitory interneurons. We suggest that the disynaptic inhibitory pathway from the SC mediates a saccadic signal that initiates an OPN pause. Such short-latency inhibition appears favorable to rapid initiation of saccades.

The spatial pattern of inhibition was different from that of excitation. The rostral as well as caudal SC stimulation sites in this study were able to induce saccades, and both were equally effective for producing inhibition. This fact is consistent with the idea that the first element of the disynaptic pathway consists of saccade-related burst neurons, which are distributed widely in the SC. In contrast, monosynaptic excitation was more efficiently induced from the rostral SC than from the caudal SC. This supports the previous finding of Paré and Guitton (1994) and the notion that the rostral pole of the SC is involved with fixation (e.g., Munoz and Guitton 1991).

In our intracellular study, the incidence of monosynaptic EPSPs was lower than that of extracellularly studied excitation, particularly for caudal stimulation (Table 1). This is probably due to a fairly depolarized state of impaled cells, which would make EPSPs smaller and more difficult to detect. It is likely that monosynaptic EPSPs induced by caudal stimulation are small and easily missed under such conditions.

Inhibitory interneurons mediating SC-induced IPSPs remain to be identified. One candidate is reticular long-lead burst neurons (LLBNs), which are monosynaptically activated from the SC (Raybourn and Keller 1977). Intracellular HRP techniques have shown that some pontine LLBNs project to the OPN area (Scudder et al. 1996b). Another candidate may be neurons at the dorsomedial margin of the nucleus reticularis tegmenti pontis, as suggested by Langer and Kaneko (1990). This region receives direct projections from the SC and contains neurons that are labeled retrogradely after HRP injection in the OPN area. A third possibility is that the disynaptic inhibition is mediated by local interneurons close to OPNs. It has been shown that axonal branches of saccade-related SC burst neurons reach the OPN region (Scudder et al. 1996a). The fourth candidate, medium-lead inhibitory burst neurons (IBNs), probably do not relay the disynaptic IPSPs, as evidenced by the fact that single-pulse stimulation of the SC during intersaccadic intervals did not induce spikes of IBNs (Chimoto et al. 1996; Raybourn and Keller 1977) but did induce disynaptic IPSPs in OPNs.

Since OPNs pause for saccades in all directions, we first expected that individual OPNs might receive inhibitory effects following stimulation of the colliculus on each side. However, nearly half of the OPNs examined in our study were inhibited from one colliculus and not from the other (Table 1B), possibly implying that for a given saccade, not all OPNs are evenly inhibited by a signal from the SC. A partial reduction in OPN population activity could be sufficient to allow medium-lead burst neurons to begin firing, as suggested by Scudder (1988), and the activity of these burst neurons could, in turn, inhibit a whole population of OPNs, including those that have not been inhibited by SC signals. An alternative explanation is that all OPNs receive disynaptic inhibitory connections from the colliculi on both sides but single-pulse stimulation is not always sufficient to activate inhibitory interneurons. Further studies are needed to distinguish between these possibilities.