J Neurophysiol 98: 2664-2682, 2007.
First published August 29, 2007; doi:10.1152/jn.00696.2007
0022-3077/07 $8.00
Commissural Mirror-Symmetric Excitation and Reciprocal Inhibition Between the Two Superior Colliculi and Their Roles in Vertical and Horizontal Eye Movements
M. Takahashi,
Y. Sugiuchi and
Y. Shinoda
Department of Systems Neurophysiology, Graduate School of Medicine, Tokyo Medical and Dental University, Tokyo, Japan
Submitted 24 June 2007;
accepted in final form 23 August 2007
 |
ABSTRACT
|
|---|
The functional roles of commissural excitation and inhibition between the two superior colliculi (SCs) are not yet well understood. We previously showed the existence of strong excitatory commissural connections between the rostral SCs, although commissural connections had been considered to be mainly inhibitory. In this study, by recording intracellular potentials, we examined the topographical distribution of commissural monosynaptic excitation and inhibition from the contralateral medial and lateral SC to tectoreticular neurons (TRNs) in the medial or lateral SC of anesthetized cats. About 85% of TRNs examined projected to both the ipsilateral Forel's field H and the contralateral inhibitory burst neuron region where the respective premotor neurons for vertical and horizontal saccades reside. Medial TRNs received strong commissural excitation from the medial part of the opposite SC, whereas lateral TRNs received excitation mainly from its lateral part. Injection of wheat germ agglutinin–horseradish peroxidase into the lateral or medial SC retrogradely labeled many larger neurons in the lateral or medial part of the contralateral SC, respectively. These results indicated that excitatory commissural connections exist between the medial and medial parts and between the lateral and lateral parts of the rostral SCs. These may play an important role in reinforcing the conjugacy of upward and downward saccades, respectively. In contrast, medial SC projections to lateral SC TRNs and lateral SC projections to medial TRNs mainly produce strong inhibition. This shows that regions representing upward saccades inhibit contralateral regions representing downward saccades and vice versa.
|
INTRODUCTION
|
|---|
The pathway that interconnects the two superior colliculi (SCs) has been considered to be important for visual-orienting behavior. This functional significance was first revealed by Sprague (1966)
—the so-called Sprague effect—in which transection of the commissure between the SCs allows cats to once again orient to visual targets in a hemifield that had been made blind by prior ablation of the occipitotemporal cortex. This suggested that the commissural connection of the superior colliculus (SC) plays an important role in visual-orienting behavior and may mediate mutual suppression between the two SCs to prevent initiation of competing responses. Edwards (1977) demonstrated that a population of neurons in the cat SC projected through the commissure to the contralateral side. The presence of an inhibitory tectotectal projection was also confirmed by many physiological studies (Hoffman and Straschill 1971; Infante and Leiva 1986; Maeda et al. 1979; Moschovakis and Karabelas 1982). In particular, the existence of an inhibitory commissural connection was first physiologically demonstrated by Maeda et al. (1979) using intracellular recording. Electrical stimulation of the contralateral SC evoked inhibitory postsynaptic potentials (IPSPs) in tectal cells and the latencies of the IPSPs were mainly monosynaptic. Later, Moschovakis and Karabelas (1985) classified efferent neurons in the deeper layers of the cat SC into three major groups (X, T, and I neurons) by use of intracellular HRP staining. Of these, T neurons were defined as those that contained at least one branch that participated in the commissural projection of the SC. They presumed that T neurons were inhibitory because, at that time, a commissural connection between the SCs was considered to be mainly inhibitory. In agreement with these observations, Appell and Behan (1990) found that commissural cells constituted a major source of GABAergic projection to the contralateral SC. Thus it had been considered that tectotectal neurons were mainly inhibitory and allowed the SC on the active side to suppress the opposite SC. Recently, however, a study showed that transection of nontectotectal fibers in the caudal half of the commissure of the SC restored visual orientation to a cat that had been previously rendered hemianopic by a large unilateral cortical lesion (Wallace et al. 1989, 1990). Therefore rostral tectotectal commissural connections must play a functional role different from suppression of the "Sprague effect." In fact, other studies have suggested the existence of an excitatory commissural connection. Fixation cells located in the rostral pole of the SC in the monkey (Munoz and Guitton 1991; Munoz and Wurtz 1993a,b) were activated by electrical stimulation of the fixation zone of the contralateral SC (Munoz and Istvan 1998). Olivier et al. (2000) used a double-labeling method to demonstrate that commissural neurons are either GABAergic or glutamatergic and both have similar populations in the rostral SC of the cat. This supports an earlier ultrastructural study indicating that commissural terminals could be divided into two classes suggestive of both excitatory and inhibitory action (Behan 1985). These studies suggest that both excitatory and inhibitory neurons in the SC project to the contralateral SC. In fact, we recently demonstrated that a strong excitatory monosynaptic commissural connection exists in the rostral SCs, along with different patterns of excitatory and inhibitory commissural inputs to fixation neurons and saccade neurons that are related to horizontal, oblique, and vertical saccades (Takahashi et al. 2005b).
The spatial distribution of commissural neurons and their terminals in the SC has been examined anatomically. Tectotectal cells are generally restricted to the rostral half of the SC, but their projections show diverse patterns (Edwards 1977; Fish et al. 1982; Magalhães-Castro et al. 1978; Olivier et al. 1998). Intracellularly labeled neurons showed a widespread arrangement of labeled axons and terminals of individual neurons in the contralateral SC (Grantyn 1988; Moschovakis and Karabelas 1985; Moschovakis et al. 1988a; Rhoades et al. 1986). Our previous study showed that commissural neurons could be classified into two types: those that project only to the rostral part of the contralateral SC and those that project to the entire rostrocaudal extent of the contralateral SC (Takahashi et al. 2005b). We noted that the former are most likely excitatory and the latter are most likely inhibitory because, in response to contralateral stimulation of the rostral SC, monosynaptic excitatory postsynaptic potentials (EPSPs) were recorded only from the rostral part of the SC and monosynaptic IPSPs were recorded from the entire rostrocaudal extent of the SC. This finding suggested that both excitatory and inhibitory commissural neurons are located mainly in the rostral part of the SC, but they show a difference in the rostrocaudal distribution of their commissural projections. Specifically, inhibitory commissural neurons project widely to the contralateral SC, whereas excitatory commissural neurons project only to the rostral part of the contralateral SC. However, we still do not know how tectotectal commissural excitation and inhibition are distributed with respect to the mediolateral dimension of the SC, which is of importance because the medial and lateral halves of the SC represent saccades into the upper field and lower field, respectively.
In this study, we sought to determine the specific distribution of excitatory and inhibitory commissural inputs to TRNs in the medial or lateral part of the SC from the lateral and medial parts of the contralateral SC in anesthetized cats. TRNs were identified by their antidromic responses to stimulation of the ipsilateral Forel's field H (FFH) and the contralateral inhibitory burst neuron (IBN) region where premotor neurons for vertical and horizontal saccades exist, respectively. Systematic stimulation of the medial and lateral parts of the contralateral SC was performed in anesthetized cats to determine the mediolateral distribution of commissural excitation and inhibition in TRNs in the medial or lateral SC in preparations with or without commissural inputs of nontectal origin. In related anatomical experiments, wheat germ agglutinin–horseradish peroxidase (WGA-HRP) was injected into the medial or the lateral part of the rostral SC and the retrogradely labeled cell bodies were mapped in the contralateral SC. The results show that most of the TRNs examined projected to both the ipsilateral FFH and the contralateral IBN region, and TRNs in the lateral and medial SC received commissural excitation from the lateral and medial SCs on the opposite side, respectively. Furthermore, this commissural excitation was partly conveyed by commissural collaterals of TRNs. In contrast, medial and lateral TRNs received stronger commissural inhibition from the lateral and medial parts of the opposite SC, respectively. This point-to-point commissural excitation in the two SCs and the reciprocal commissural inhibition between the lateral and medial parts of the two SCs will be discussed in relation to the functional roles of the commissural excitation and inhibition in vertical and horizontal saccade generation.
|
METHODS
|
|---|
Experiments were performed in 20 cats weighing 2.5–5.0 kg. Animal experimentation was conducted in accordance with the "Policies on the Use of Animals and Humans in Neuroscience Research," revised and approved by the Society for Neuroscience in 1995, and the "Guiding Principles for the Care and Use of Animals in the Field of Physiological Sciences" (The Physiological Society of Japan, revised in 2001). The experimental protocol was approved by the Animal Care Committee of Tokyo Medical and Dental University. The animals were initially anesthetized with ketamine hydrochloride [Ketalar, 25 mg/kg, administered intramuscularly (im); Parke-Davis] followed by 
-chloralose [40–45 mg/kg, administered intravenously (iv), initial dose, supplemented with additional doses of 10–25 mg/kg, iv throughout the remainder of the experiment]. During recording, the animals were paralyzed by the iv administration of pancuronium bromide (Mioblock, Organon Teknika, Oss, The Netherlands) and artificially ventilated with end-tidal CO2 held at 35–40 mmHg. The heart rate was constantly monitored by an electrocardiogram. The body temperature was kept at 37.0–38.5°C by a heating pad.
The bone over the parietal and occipital cortex was removed and the cerebral cortex was removed bilaterally by aspiration, to allow the introduction of stimulating electrodes into the right SC and a recording electrode into the left SC under direct visual observation. In three cats, four concentric bipolar stimulating electrodes (ID, 0.1 mm; OD, 0.3 mm; interelectrode distance along the longitudinal axis, 1.0 mm) with a 1.0- to 1.2-mm rostrocaudal separation were placed along the presumed horizontal meridian of the motor map in the SC on the right side (McIlwain 1986) (see Fig. 1, A and D). In eight cats, two arrays of four concentric bipolar stimulating electrodes with a 1.0- to 1.2-mm rostrocaudal separation were placed in the medial and lateral parts of the right SC, respectively (see Fig. 4A). In three cats, three concentric bipolar stimulating electrodes with a 1.0- to 1.2-mm mediolateral separation were placed in the most rostral part of the right SC, where the vertical meridian of the motor map is represented (Guitton et al. 1980; McIlwain 1986; Robinson 1972; Sparks and Mays 1983; Stanford et al. 1996) and two electrodes of the same type were placed along the horizontal meridian of the motor map (see Fig. 11, A and E). The tips of the SC electrodes were positioned in the intermediate or deep layer (1.5–2.0 mm from the surface) of the SC (see Fig. 7F) (Izawa et al. 1999; Kawamura and Hashikawa 1978; Moschovakis and Karabelas 1985; Sugiuchi et al. 2005; Takahashi et al. 2005a,b). For antidromic activation of TRNs projecting to the FFH, an array of two concentric bipolar electrodes with a 1.5-mm mediolateral separation was stereotaxically placed on each side in the FFH (A: 7.0–7.5, L: 1.0, 2.5) (see Fig. 7E).
View larger version (22K):
[in this window]
[in a new window]
|
FIG. 1. Properties of commissural inputs to tectoreticular neurons (TRNs) in the caudal part of the superior colliculus (SC). A–C: a caudal TRN projecting to both Forel's field H (FFH) and the inhibitory burst neuron (IBN) region. A: dorsal view of the brain stem showing the experimental setup. Stimulating electrodes were placed in the right SC (sites 1–4), the left and right FFH (sites 11–14), the right omnipause neuron (OPN) region (sites 5–7), and the right IBN region (sites 8–10). VI, abducens nucleus. In this diagram and similar diagrams in the following figures, a recording site of a TRN is indicated by an open cell penetrated with a microelectrode on the surface of the representative SC on the left side and locations of stimulating electrodes are indicated on the surface of the SC based on the picture taken after each experiment. B: antidromic identification (anti ID) of projection sites of a TRN. Antidromic spikes of a TRN in the left SC were evoked by stimulation of the ipsilateral FFH (site 12) and the contralateral IBN region (site 8) at 500 µA. This TRN was not activated from the contralateral FFH at 500 µA (not illustrated). C: properties of inhibitory postsynaptic potentials (IPSPs) evoked by stimulation of the contralateral SC at 200 µA in the same TRN as in B. Number attached to each panel (1–4) corresponds to each stimulation site in the contralateral SC (sites 1–4). Top and bottom traces: intracellular potentials and juxtacellular field potentials recorded just outside the penetrated cell, respectively. Same arrangements for stimulating electrodes and their response traces are used in the following figures, if not stated otherwise. D–F: a caudal TRN projecting only to the IBN region. D: experimental setup. E: antidromic spikes of a TRN in the left SC were evoked by stimulation of the contralateral IBN region (site 8), but only postsynaptic potentials (PSPs) without any spikes were evoked by stimulation of the FFH on either side at 500 µA (sites 11–14). Note that traces of antidromic spikes are clipped in this and following figures. F: properties of IPSPs evoked by stimulation of the contralateral SC at 200 µA in the same TRN as in E.
|
|
View larger version (24K):
[in this window]
[in a new window]
|
FIG. 2. Properties of commissural inputs to TRNs in the rostral SC. A–C: a rostral TRN projecting to both the FFH and the IBN region. A: experimental setup. Open and filled cells in the right SC indicate presumed locations of excitatory and inhibitory commissural neurons, respectively. B: antidromic spikes of a TRN in the rostral SC evoked by stimulation of the ipsilateral FFH (site 12) and the contralateral IBN region (site 8) at 500 µA. C: excitatory postsynaptic potentials (EPSPs) and IPSPs evoked by stimulation of the contralateral SC at 500 and 200 µA in the same TRN as in B. D–F: a rostral TRN projecting solely to the IBN region. D: experimental setup. E: antidromic spikes of a TRN in the rostral SC were evoked by stimulation of the contralateral IBN region (site 9), but not by stimulation of the ipsilateral FFH (site 11) and the contralateral FFH (not illustrated) at 500 µA. F: commissural inputs to the same TRN as in E.
|
|
View larger version (37K):
[in this window]
[in a new window]
|
FIG. 3. Distribution of commissural neurons that were retrogradely labeled after injection of wheat germ agglutinin–horseradish peroxidase (WGA-HRP) into the lateral (A, B) and medial parts (C, D) of the rostral SC. A: dorsal view of the site of WGA-HRP injection into the lateral part of the left SC (black area) and the distribution of retrogradely labeled neurons in the right SC. WGA-HRP solution (0.1 µl) was injected 2.0 mm from the tectal surface in the rostrolateral SC. Survival time, 48 h. Each dot indicates a single labeled neuron. Labeled neurons are plotted on serial transverse sections (70 µm thick) and replotted on the dorsal view of the SCs. B: distribution of the labeled commissural neurons in representative frontal planes of the same SC as shown in A. Representative frontal drawings (a–d) are presented at 350-µm intervals. Each drawing shows neurons contained in 5 consecutive sections. Dots indicate small neurons and open circles indicate larger neurons whose soma cross-sectional area was >500 µm2. SGS, stratum griseum superficiale; SO, stratum opticum; SGI, stratum griseum intermediale; SGP, stratum griseum profundum. C: dorsal view of the site of WGA-HRP injection into the medial part of the left SC (black area) and the distribution of retrogradely labeled neurons in the right SC. WGA-HRP solution (0.1 µl) was injected 1.8 mm from the tectal surface in the rostromedial SC. Survival time, 48 h. D: distribution of the labeled commissural neurons in frontal planes of the same SC as shown in C. Same format as in B.
|
|
View larger version (22K):
[in this window]
[in a new window]
|
FIG. 4. Properties of commissural inputs from the medial and lateral parts of the contralateral SC to a TRN in the lateral part of the caudal SC. A: experimental setup. For stimulation of the contralateral SC, one array of 4 concentric bipolar electrodes was placed at a depth of 1.5 mm in the medial SC (sites 5–8), where upward saccades are represented in the motor map (McIlwain 1986), and another array was placed at a depth of 1.8 mm in the lateral SC (sites 9–12), where downward saccades are represented. B: antidromic spikes of a TRN in the caudolateral SC evoked by stimulation of the ipsilateral FFH (site 1) and the contralateral IBN region (site 16) at 500 µA. C and D: commissural inhibition to the same TRN from the medial part (sites 5–8) (C) and the lateral part (sites 9–12) (D) of the contralateral SC at 175 and 75 µA.
|
|
View larger version (23K):
[in this window]
[in a new window]
|
FIG. 5. Properties of commissural inputs from the medial and lateral parts of the contralateral SC to a TRN in the medial part of the caudal SC (A–C) and a TRN in the medial part of the rostral SC (D–F). A: experimental setup. B: antidromic spikes of a TRN in the medial part of the caudal SC were evoked by stimulation of the ipsilateral FFH (site 2) and the contralateral IBN region (site 16) at 500 µA. C: commissural inputs to the same TRN as shown in B from the medial (sites 5–8) and lateral parts (sites 9–12) of the contralateral SC at 175 µA. Inhibition from the lateral part (sites 10 and 11) was slightly stronger than that from the medial part (sites 6 and 7) because the slopes of the falling phase of the inhibition were steeper. D: experimental setup. Intracellular potentials were recorded from a TRN in the medial part of the left rostral SC. E: antidromic spikes of the TRN evoked in an all-or-none manner at threshold by stimulation of the ipsilateral FFH (site 2) and the contralateral IBN region (site 16) at 500 µA. F: properties of EPSPs and IPSPs evoked by stimulation of the medial (sites 5–8) and lateral parts (sites 9–12) of the contralateral SC at 200 µA in the same TRN as in E.
|
|
View larger version (43K):
[in this window]
[in a new window]
|
FIG. 6. Latency histograms of antidromic spikes and commissural excitation and inhibition in TRNs recorded in the medial part (A–E) (n = 64) and lateral part of the rostral SC (F–J) (n = 67). Arrangement of 4 stimulation electrodes in the medial (sites 5–8) and lateral SC (sites 9–12) is the same as in Fig. 4A. A and F: antidromic spikes evoked by stimulation of the ipsilateral FFH (a) and the contralateral IBN region (b). B, C and G, H: EPSPs evoked by stimulation of the medial (B, G) and lateral (C, H) parts of the rostral SC on the opposite side. D, E and I, J: IPSPs evoked by stimulation of the medial (D, I) and lateral (E, J) parts of the rostral SC on the opposite side. B–E and G–J: total number of PSP latencies is less than the total number of TRNs tested because some latencies could not be determined as a result of the appearance of spikes, masking IPSPs due to the leakage of Cl– from the pipette or the difficulty in reversing IPSPs by Cl– injection.
|
|
View larger version (39K):
[in this window]
[in a new window]
|
FIG. 7. Properties of commissural inputs from the medial and lateral parts of the contralateral SC to a TRN in the lateral part of the SC. A: experimental setup. B: antidromic spikes of a TRN in the lateral part of the rostral SC were evoked by stimulation of the ipsilateral FFH (site 1) at 75 µA and the contralateral IBN region (site 18) at 250 µA. C: commissural inputs to the same TRN as shown in B from the contralateral SC evoked at 300 µA. Note that commissural inhibition was larger from the medial parts (sites 5–7) than from the lateral parts of the contralateral SC (sites 10–12). D: effects of stimulus intensity and stimulus depth at the rostral part of the contralateral SC. a: same stimulation site and polarity as in C5, but at 400 µA. For concentric bipolar stimulation, a cathodal electrode was placed at 1.8 mm and an anodal electrode was placed at 0.8 mm from the SC surface. b: same stimulation site as in C, site 5, but at 550 µA with reversed stimulus polarity for the same electrode. c: same stimulation site as in C, site 9 at 100 µA. Note that the excitation was large enough to evoke orthodromic spikes, although the stimulus intensity was much weaker. E: photomicrograph of the frontal section showing the locations of the stimulating electrode tips (arrowheads) in the ipsilateral (sites 1 and 2) and contralateral FFH (sites 3 and 4). RF, retroplexus bundle; CP, cerebral peduncle. F: photomicrograph of the frontal section of the SC to show the most effective stimulation sites in the right SC (sites 5 and 9). Bottom and top arrowheads for each electrode track indicate the locations of cathodal and anodal electrode tips. CG, central gray.
|
|
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 8. Intraaxonal spikes recorded from an axon of a TRN with a commissural collateral to the contralateral SC. A: experimental setup. Intraaxonal recording was made from an axon of a TRN in the right OPN region. Five stimulating electrodes (sites 5–9) in the left SC and 5 stimulating electrodes in the right SC (sites 10–14). B and C: spikes recorded from an axon in the right OPN region. Note that spikes were not evoked from the rostrolateral and caudal parts of the SCs.
|
|
View larger version (26K):
[in this window]
[in a new window]
|
FIG. 9. Commissural inputs to TRNs with commissural collaterals that are located in the lateral part (A–C) and medial part of the rostral SC (D–F). A–C: a TRN in the lateral part of the rostral SC. A: experimental setup. B: antidromic spikes of a TRN in the lateral part of the rostral SC were evoked in an all-or-none manner at threshold by stimulation of the ipsilateral FFH (site 1) at 250 µA and the contralateral IBN region (site 18) at 250 µA. Same TRN was antidromically activated in an all-or-none manner at 250 µA from the rostrolateral site of the contralateral SC (site 9). C: properties of EPSPs and IPSPs evoked by stimulation of the medial and lateral parts of the contralateral SC in the same TRN as in B. D–F: a TRN in the medial part of the rostral SC. D: experimental setup. E: antidromic spikes of a TRN were evoked in an all-or-none manner at threshold by stimulation of the ipsilateral FFH (site 2) and IBN regions (site 18) at 500 µA. F: commissural inputs evoked by stimulation of the medial and lateral parts of the contralateral SC at 300 µA except for site 5 (250 µA) in the same TRN as in E. Antidromic spikes were evoked only by stimulation of the rostromedial part of the contralateral SC (site 5) at 250 µA.
|
|
View larger version (33K):
[in this window]
[in a new window]
|
FIG. 10. Monosynaptic excitation of a TRN from the contralateral FFH and the ipsilateral IBN region. A: experimental setup. Tectoreticular axons in the medial longitudinal fascicle (MLF) were sectioned transversely at the rostral border of the right OPN region (solid bar). One array of 3 stimulating electrodes with a 1.0-mm dorsoventral separation was placed in the right tectoreticular tract (TRT) just rostral to the section (sites 13–15) and another in the left IBN region (sites 16–18). Intracellular potentials were recorded from a TRN in the rostromedial part of the left SC. B: antidromic spikes evoked by stimulation of the ipsilateral FFH (site 1) (stimulus intensity, 150 µA) and the contralateral TRT (site 14) (stimulus intensity, 400 µA) at thresholds. C: commissural inputs to the same TRN as in B from the medial (sites 5–8) and lateral parts of the contralateral SC (sites 9–12). D: effects of stimulation of the contralateral FFH (site 4) and the ipsilateral IBN region (site 17) on the same TRN as in B. E: photomicrograph of the frontal section of the brain stem showing the section of the right TRT in the MLF just rostral to the OPN region. F: photomicrograph of the frontal section showing the stimulation sites in the left IBN region (arrowheads). Arrows in E and F indicate the midline.
|
|
View larger version (23K):
[in this window]
[in a new window]
|
FIG. 11. Properties of synaptic inputs evoked by stimulation of the contralateral SC and its adjacent midbrain reticular formation to TRNs in the medial part (A–D) and lateral part of the rostral SC (E–G). A–D: a TRN in the medial part of the rostral SC. A: experimental setup. In addition to 5 concentric bipolar electrodes in the right SC (sites 5–9), 4 concentric bipolar stimulating electrodes with a 1.0- to 1.2-mm rostrocaudal separation were placed at depths of 2.0–2.5 mm in the midbrain reticular formation just lateral to the contralateral SC (sites 10–13). B: antidromic spikes evoked in the TRN by stimulation of the ipsilateral FFH (site 2) and the contralateral OPN region (site 14). C: commissural inputs from the contralateral SC to the same TRN as shown in B. D: properties of synaptic inputs from the midbrain reticular formation just lateral to the contralateral SC (extra SC) (sites 10–13). Stimulation of this area evoked small monosynaptic excitation at 500 µA, but the main component of the excitation had a latency of about 1.5 ms, indicating that it was disynaptic. E–G: a TRN in the lateral part of the rostral SC. E: experimental setup. Antidromic spikes were evoked in this TRN from the ipsilateral FFH (site 2) and the contralateral OPN regions (site 15) (not illustrated). F and G: synaptic inputs evoked by stimulation of the contralateral SC and its adjacent lateral midbrain reticular formation in the same TRN as in E. F: small excitation evoked by stimulation of the entire mediolateral extent of the contralateral rostral SC (sites 5–7). In contrast, large inhibition was evoked from all stimulation sites of the contralateral SC (sites 5–9). G: synaptic inputs from the midbrain reticular formation just lateral to the SC. Only small inhibition was evoked by double-pulse stimulation at 300 µA, and the latencies of the IPSPs were >2.0 ms, indicating that they were most likely disynaptic.
|
|
The vermis overlying the fourth ventricle was removed by aspiration to facilitate the placement of stimulating electrodes in the omnipause neuron (OPN) region and the IBN region. We first identified the location of the abducens nucleus by recording a type II response to rotation of the cat in the horizontal plane (Maeda et al. 1971; Shinoda and Yoshida 1974). Stimulation sites in the OPN (Büttner-Ennever et al. 1988; Curthoys et al. 1981; Evinger et al. 1982) and IBN regions (Hikosaka and Kawakami 1977; Hikosaka et al. 1978) were then determined relative to the abducens nucleus (Sugiuchi et al. 2005; Takahashi et al. 2005a,b). For antidromic identification of TRNs projecting to the OPN region or the IBN region, separate electrode arrays were placed in the right OPN and IBN regions contralateral to the recording site in the SC. These electrode arrays consisted of four monopolar electrodes (diameter, 100 µm), insulated except at the tip, that were glued together around a central pillar, so that the tips of the four electrodes were arranged dorsoventrally at 1.0-mm intervals. For stimulation of the IBN region, one such electrode array was placed 0.8 mm lateral to the midline and 0.5–1.0 mm caudal to the caudal border of the abducens nucleus. For stimulation of the OPN region, another such electrode array was placed at the midline and 0.3–1.0 mm rostral to the rostral border of the abducens nucleus. Stimulus currents were delivered between two adjacent tips, and the pillar was grounded to reduce stimulus artifacts. Negative pulses of 0.2-ms duration were delivered at <500 µA for stimulation of the FFH and the OPN and IBN regions and at 
300 µA for stimulation of the SC. Ranck (1975) estimated an effective current spread of 1.0–1.5 mm around an electrode tip by monopolar stimulation at 500 µA (200-µs-duration pulse) in the mammalian CNS. Because we used bipolar stimulation, the effective current spread should be much less than the values estimated by Ranck (1975) (Shinoda et al. 1977). Sasaki et al. (1970, 1972) estimated that 500 µA could not activate fibers or cells beyond 1.0 mm from an electrode tip when a concentric bipolar electrode of the same type as in our present study was used. Positions of the stimulating electrodes in the brain stem and the SC were marked by passing 200-µA negative currents for 20 s after each experiment, and the stimulated sites in the SC, the OPN and IBN regions, and the FFH were histologically confirmed on sections that had been stained with thionine. Histological examination showed that the distance between a stimulating electrode in the IBN region and the presumed caudal border of the OPN region was about 3.0 mm (2.5–3.5 mm) in each experiment. Glass microelectrodes for intracellular recording were filled with 0.4 M KCl or 2 M K-citrate (resistance 10–20 M
). Recording sites in the SC were plotted on a photograph taken during the experiment and the tips of two or more recording glass microelectrodes were broken and left in the recording sites of the SC as landmarks for later histological identification.
To examine the distribution of neurons that projected to the SC, WGA-HRP (Toyobo, Osaka, Japan) was injected into the lateral or medial part of the left SC in six cats. The bone over the parietooccipital cortex was removed under pentobarbital anesthesia (Nembutal, 35 mg/kg, iv; Abbott, Baar, Switzerland) and the parietooccipital cortex overlying the SC was removed by aspiration to identify the left SC under direct visual observation. To eliminate nontectal fibers to the SC, a longitudinal section was made underneath the right SC in one cat in aseptic conditions under pentobarbital anesthesia (35 mg/kg, iv), after removing the overlying parietooccipital cortex by aspiration. As a precautionary measure the cat received antibiotic coverage (cefazoline sodium, 50 mg·kg–1·day–1, im) during the course of the experiment. After 7 days, the animal was prepared for an electrophysiological experiment as described earlier. After the recording session was over, WGA-HRP was injected into the medial SC on the left in this cat with the procedure subsequently described. For injection experiments, a glass micropipette was attached to the needle of a Hamilton syringe and lowered to about 1.5–2.0 mm below the SC surface. Before being withdrawn from the track, the glass micropipette was left in the injection site for 5 min. The volume of solution delivered in each track was about 0.1–0.3 µl. After 48–72 h, animals were deeply anesthetized with ketamine hydrochloride (25 mg/kg, im) followed by pentobarbital sodium (50 mg/kg, iv) and perfused with 2 L of 10% sucrose phosphate buffer (pH = 7.4) followed by 2 L of a fixative solution containing 4% paraformaldehyde and 0.05% glutaraldehyde with 0.2% picric acid in 4% sucrose phosphate buffer. Frozen serial coronal sections (70 or 80 µm thick) were cut on a freezing microtome, incubated in anti-WGA antibody and avidin–biotin complex, and then treated for HRP by the heavy-metal–intensification method (Adams 1981). The sections were lightly counterstained with neutral red. Retrogradely labeled neurons were plotted under a microscope using a camera lucida system and a computer-assisted plotting and reconstruction program (Neurolucida, Micro-BrightField, Colchester, VT).
|
RESULTS
|
|---|
To examine the functional roles of commissural excitation and inhibition between the two SCs, we systematically recorded intracellular potentials from TRNs in the rostral or caudal part, or in the medial or lateral part of the SC, and examined the effects of electrical stimulation of the rostral versus caudal parts and the medial versus lateral parts of the contralateral SC on TRNs. TRNs with their projections to different brain stem areas were identified as projecting to either the IBN region or the FFH, or both, by their antidromic responses to stimulation of the contralateral OPN and IBN regions and/or the FFH on both sides, respectively. All lateralities reported herein are described with reference to the recording site, if not stated otherwise.
Properties of commissural inputs from the contralateral SC to TRNs in the caudal and rostral SC
Our previous study showed that among TRNs that projected to the IBN region, caudal TRNs received only inhibition from the entire rostrocaudal extent of the contralateral SC, whereas rostral TRNs received additional excitation from the contralateral rostral SC (Takahashi et al. 2005b). However, in the previous study we did not identify whether TRNs that projected to the IBN region also projected to the FFH. In this study, we recorded intracellular potentials from TRNs in the rostral one third (mainly one fourth) of the SC (rostral TRNs) or from TRNs in the caudal one half of the SC (caudal TRNs), and further divided TRNs into those that projected only to the IBN region and those that projected to both the FFH and IBN regions. Intracellular potentials were recorded from 320 TRNs, although 53 of them were discarded from analysis of synaptic potentials because EPSPs and reversed IPSPs could not be differentiated due to quick spontaneous diffusion of Cl– into cells. The resting membrane potentials ranged from –40 to –75 mV (mean ± SD, –50 ± 15 mV, n = 267). Figure 1 shows examples of two TRNs recorded in the caudal parts of the SC (Fig. 1, A and D) along the presumed horizontal meridian of the motor map (McIlwain 1986). A TRN in Fig. 1, A–C was antidromically activated from the ipsilateral FFH (Fig. 1B, site 12) and the contralateral IBN area (Fig. 1B, site 8), but not from the contralateral FFH (not illustrated). Stimulation of all four rostrocaudal sites in the contralateral SC evoked only hyperpolarizations in this TRN, and their amplitudes were larger and the slopes of their falling phase were steeper when more rostral sites were stimulated (Fig. 1C). Because these commissural hyperpolarizations were reversed to depolarizing potentials by the injection of Cl– into the cells (Takahashi et al. 2005b), these hyperpolarizations were considered to be IPSPs (Coombs et al. 1955; Eccles 1964). Like this example, most of the caudal TRNs (38 of 45 TRNs examined) projected not only to the contralateral IBN region but also to the ipsilateral FFH. Although most of these cells received only inhibition from the contralateral SC, very small EPSPs that preceded the IPSPs were observed in four TRNs. Because these caudal TRNs were recorded along the horizontal meridian of the motor map, they are most likely related to oblique saccades that have large horizontal and small vertical components. Among the caudal TRNs, some (7 of 45 TRNs examined) projected only to the contralateral IBN area (Fig. 1E, site 8), but not to the FFH on the ipsilateral (Fig. 1E, sites 11 and 12) and contralateral sides (Fig. 1E, sites 13 and 14). These TRNs received only inhibition from the contralateral SC, which was larger from the more rostral SC. However, in this TRN, the largest inhibition was evoked at site 2 (Fig. 1F, site 2), rather than from the most rostral site 1, most likely because the stimulating electrode at site 1 was just rostral to the rostral border of the SC, as shown in Fig. 1D. Because these TRNs projected only to the IBN region, they were most likely related to pure horizontal saccades.
To compare the patterns of commissural inputs to TRNs in the caudal versus rostral SC, we also recorded intracellular potentials from TRNs in the rostral SC. Figure 2 shows examples of two TRNs in the rostral SC. A TRN in Fig. 2, A–C projected to both the ipsilateral FFH and the contralateral IBN region (Fig. 2B, sites 12 and 8). Stimulation of the rostral site of the contralateral SC evoked a depolarization with spikes at 500 µA, which was followed by a hyperpolarization (Fig. 2C, site 1). Even with a decrease in the stimulus intensity to 200 µA, depolarizations were evoked mainly from the rostral sites and hyperpolarizations were evoked from the caudal sites. Our previous study showed that the early depolarization and the late hyperpolarization evoked by stimulation of the contralateral SC were an EPSP and an IPSP, respectively (Takahashi et al. 2005b) because injection of Cl– into the cells changed the hyperpolarization to a depolarizing potential, whereas the early depolarization remained unaffected (Coombs et al. 1955; Eccles 1964). This TRN was recorded near the horizontal meridian of the motor map in the most rostral part of the SC and thus this TRN was most likely an oblique saccade neuron with a small horizontal component (McIlwain 1986). The TRN in Fig. 2, D–F projected only to the contralateral IBN region (Fig. 2E, site 9) without any antidromic spikes evoked by the ipsilateral FFH stimulation (Fig. 2E, site 11). This TRN received only inhibition, without any excitation from the most rostral SC (Fig. 2F, site 1). More caudal sites evoked much smaller inhibitory potentials (Fig. 2F, sites 2–4). Because this TRN was recorded in the center of the most rostral SC where small-amplitude saccades are represented in the motor map, it was most likely that this TRN codes small horizontal saccades. However, in the rostral SC, there were not many such TRNs that projected only to the IBN region (5 of 42 TRNs examined). As shown in Fig. 2, virtually all rostral TRNs that projected to both the IBN region and the FFH received monosynaptic excitation from the rostral part of the contralateral SC and inhibition from its entire rostrocaudal extent, whereas TRNs with a projection only to the IBN region had only monosynaptic inhibition from the contralateral SC.
Spatial distribution of tectotectal connections
Although it has been tacitly assumed that there are point-to-point connections between the rostral SCs on both sides (Behan and Kime 1996a; Edwards 1977; Fish et al. 1982; Magalhães-Castro et al. 1978), there has been no experimental demonstration of a commissural connection between the medial parts of the rostral SCs because Edwards (1977) failed to find such a medial commissural connection. To examine the mediolateral distribution of commissural neurons in the SCs, we injected WGA-HRP into the lateral or medial part of the left rostral SC. Figure 3, A and B shows an example with a WGA-HRP injection into the lateral part of the most rostral SC, which includes the vertical meridian of the motor map for downward saccades (McIlwain 1986). The core of the injection site was located in the intermediate layer of the lateral one third of the rostral SC (Fig. 3, A and B). Retrogradely labeled neurons were distributed most densely in the rostral one third of the right SC, and disappeared rather abruptly at about rostral one half of the total rostrocaudal distance of the SC in the lateral half of the SC. Among the labeled cells, small neurons (filled circles in Fig. 3B) were observed mainly in the intermediate layer and the deep layer and were distributed over the mediolateral extent of the rostral SC, with a lateral bias. However, larger (i.e., cell bodies >500 µm2) labeled neurons (open circles in Fig. 3B) were observed in the intermediate layer, and only in the lateral part of the rostral SC, at a site that was almost symmetrical to the injection site.
To compare the distributions of labeled commissural neurons after a tracer injection into the medial versus the lateral part of the SC, WGA-HRP was injected into the medial part of the left rostral SC. Figure 3, C and D shows a typical injection of WGA-HRP into the medial part of the most rostral SC, which includes the vertical meridian of the motor map for upward saccades (McIlwain 1986). The core of the injection site covered the medial one third of the rostral SC and extended from the intermediate layer to the deep layer (Fig. 3D). Many retrogradely labeled neurons were located in the rostral part of the right SC. Retrogradely labeled neurons of a larger size (open circles in Fig. 3D) were mainly located in the intermediate layer (1.0–1.5 mm from the surface) of the medial half of the rostral SC, and this distribution area of the larger neurons corresponded closely to the symmetrical region of the injection site. Small labeled neurons (filled circles in Fig. 3D) were more widely distributed in the intermediate and deeper layers because they were found in both the medial and lateral parts of the rostral SC. These findings suggest that all commissural connections between the SCs are not strictly a point-to-point connection between mirror-symmetric regions of the SCs. Instead, there are two overlapping patterns of the connection: symmetrical commissural connections between larger neurons and more widespread connections for small commissural neurons that existed in the rostral SC.
Commissural inputs from the medial and lateral parts of the contralateral SC to TRNs in the caudal SC
To reveal the topographic distribution of commissural inhibition in TRNs, we examined the effects of stimulation of the medial and lateral parts of the contralateral SC on caudal TRNs. Figure 4 shows the typical pattern of commissural input to a TRN in the lateral part of the caudal SC (about 4 mm caudal to the rostral border and about 1.5 mm medial to the lateral border of the left SC). Stimulation of all eight sites in the SC evoked only inhibition (Fig. 4, C and D) at 175 µA, and weaker stimulation at 75 µA still evoked inhibition at all stimulation sites, except site 9, which was found to be at the rostral border of the lateral SC after the histological examination. The amplitude and slope of the falling phase of the inhibition were larger and the latency was shorter when more rostral sites were stimulated, except for site 9 (Fig. 4, C, site 5 and D, site 10). In addition, the inhibition was larger from the medial part (Fig. 4C, sites 5–8) than the lateral part of the contralateral SC (Fig. 4D, sites 10–12). As in this example, the commissural inhibition in TRNs in the lateral part of the caudal SC (14 of 14 TRNs examined) was largest from the most rostral part of the medial SC. The latencies of the IPSPs in the caudolateral TRNs evoked by stimulation of the medial and lateral parts of the rostral SC ranged from 1.0 to 1.4 ms (mean ± SD, 1.2 ± 0.1 ms, n = 14) and from 1.0 to 1.8 ms (1.4 ± 0.3 ms, n = 14), respectively.
On the other hand, Fig. 5, A–C shows commissural input to a TRN in the medial part of the caudal SC (about 4 mm caudal to the rostral border of the left SC and about 1.5 mm lateral to the midline). Stimulation of the medial and lateral parts of the contralateral SC evoked inhibition in this TRN (Fig. 5C). The amplitudes and latencies of the inhibition from the medial and the lateral SC were not particularly different, but the falling slopes of the inhibition from the lateral SC were slightly steeper than or were almost equal to those from the medial SC, suggesting that inhibition was stronger from the lateral SC. In 8 of 14 TRNs in the medial part of the caudal SC, inhibition was larger from the lateral SC, and in the other 6 TRNs, inhibition from the lateral and medial parts of the contralateral SC was almost equal. The latencies of the IPSPs in the caudomedial TRNs evoked by stimulation of the medial and lateral parts of the rostral SC ranged from 1.0 to 1.8 ms (1.5 ± 0.4 ms, n = 6) and from 1.0 to 1.8 ms (1.4 ± 0.3 ms, n = 8), respectively. In summary, TRNs in the lateral part of the caudal SC received stronger commissural inhibition from the medial SC, whereas those in the medial part of the caudal SC received stronger commissural inhibition from the lateral SC or almost equal commissural inhibition from the medial and lateral SC. This inhibition was monosynaptic from the rostral part of the contralateral SC.
Commissural inputs from the medial and lateral parts of the contralateral SC to TRNs in the rostral SC
We showed that rostral TRNs received both monosynaptic excitation and inhibition from the contralateral rostral SC (Fig. 2, A–C). Furthermore, the morphological data suggest that larger commissural neurons project to the corresponding part of the contralateral SC: larger neurons in the lateral SC project to the lateral part of the contralateral SC and those in the medial SC project to the corresponding medial part. These connections between the corresponding parts of the SCs seem most likely to be excitatory because commissural neurons involved in these projections are rather large and some TRNs have a commissural collateral to the contralateral SC (Moschovakis and Karabelas 1985; Takahashi et al. 2005b). To analyze the mediolateral distribution of the commissural excitation and inhibition in the rostral SC, we examined the effects of stimulation of the medial and lateral parts of the contralateral SC on TRNs in the medial part of the rostral SC (Fig. 5, D–F). Figure 5, D–F shows a typical example of commissural inputs to a TRN in the medial part of the rostral SC. In this TRN, strong excitation that evoked orthodromic spikes was caused by stimulation of only the most rostral site of the medial SC (Fig. 5F, site 5) and inhibition was evoked by stimulation of the other medial sites (Fig. 5F, sites 6–8) and lateral sites (Fig. 5F, sites 9–12) of the contralateral SC. As exemplified in this TRN, TRNs in the medial part of the rostral SC received strong excitation from the rostromedial part of the contralateral SC and weaker excitation from nearby sites of the contralateral SC. The latencies of EPSPs in the rostromedial TRNs evoked by stimulation of the medial and lateral parts of the contralateral rostral SC ranged from 0.6 to 1.4 ms (0.8 ± 0.2 ms, n = 39) (Fig. 6Ba) and from 0.7 to 1.4 ms (0.9 ± 0.2 ms, n = 27) (Fig. 6Ca), respectively. The difference in the latencies of the EPSPs between the medial and lateral SCs was about 0.1 ms, which was most likely attributed to the current spread from the lateral electrode to excitatory commissural neurons in the medial SC. In contrast, rostromedial TRNs usually received as strong inhibition from the medial as from the lateral part of the contralateral SC. Because the early excitation was usually present, the onset of the IPSPs could be determined only by superimposing reversed IPSPs on the control IPSPs. The latencies of the IPSPs in the rostromedial TRNs evoked by the medial and lateral parts of the rostral SC ranged from 1.0 to 1.7 ms (1.3 ± 0.2 ms, n = 31) (Fig. 6Da) and from 1.0 to 2.0 ms (1.4 ± 0.3 ms, n = 28) (Fig. 6Ea), respectively.
The commissural input pattern to TRNs in the lateral part of the rostral SC was different from that to TRNs in the medial part. Figure 7 shows a typical example of commissural inputs to a TRN in the lateral part of the rostral SC. Strong excitation that triggerd orthodromic spikes was evoked by stimulation of the most rostral site of the lateral SC (Fig. 7C, site 9) and weaker excitation was evoked by stimulation of the more caudal sites of the lateral part (Fig. 7C, sites 10–12) and the most rostral site of the medial part (Fig. 7C, site 5) of the contralateral SC. In most TRNs in the lateral part of the rostral SC, excitation was evoked by stimulation of both the medial (sites 5 and 6) and lateral parts (sites 9 and 10) of the contralateral rostral SC. This might be explained by the fact that commissural fibers from the lateral SC pass through the medial part of the SC to the contralateral side and medial stimulating electrodes might activate such passing fibers. The amplitude of the excitation from the rostromedial part (site 5) was larger when the stimulus intensity was increased from 300 to 400 µA (compare Fig. 7, C, site 5 and Da). When site 5 was stimulated with a reversed stimulus polarity for the same electrode (Fig. 7Db) (the cathodal and anodal electrode tips were 0.8 and 1.8 mm from the surface of the SC, respectively), the excitation was much smaller, and the inhibition was as large as with a normal stimulus polarity. These findings suggested that the excitation evoked by stimulation of the contralateral medial SC originated from neurons in the lateral SC, but not in the medial SC. Stronger stimulation (Fig. 7Da) or stimulation with the deeper cathodal tip (Fig. 7C, site 5) was considered to activate excitatory passing fibers from the lateral SC that were located in the deeper part of the medial SC. This interpretation was further supported by the observation that weaker stimulation (stimulus intensity, 100 µA) of the rostrolateral part (site 9) of the contralateral SC still evoked strong excitation with the normal polarity (Fig. 7Dc). Therefore excitatory commissural neurons terminating on the lateral SC were most likely located in the lateral part of the rostral SC. In contrast, in the rostrolateral TRNs, weaker stimulation of the lateral SC could not evoke large inhibition, whereas stimulation of the more superficial layer of the medial SC with the reversed polarity still evoked large inhibition (Fig. 7Db). Accordingly, it was most likely that medial stimulation activated inhibitory tectotectal neurons in the medial SC rather than inhibitory passing fibers from the lateral SC. This point will be addressed in more detail in relation to other possible pathways that might be involved in commissural inhibition.
Commissural collaterals of TRNs projecting to the contralateral SC
Our previous study showed that some TRNs that projected to the contralateral IBN region also had commissural collaterals to the contralateral SC (Takahashi et al. 2005b). Because these TRNs are excitatory for OPNs (Takahashi et al. 2005a) or IBNs (Sugiuchi et al. 2005), the commissural collaterals of these TRNs must be excitatory and most likely constitute a candidate for mediating commissural excitation described in the previous sections.
To analyze commissural collaterals of TRNs in the SC, we recorded intraaxonal spikes in the tectoreticular tract (TRT) in the ventral part of the right medial longitudinal fascicle (MLF) (Verhaart 1964) at the level of the OPN region, and examined whether TRNs with their stem axons running through the TRT had axon collaterals to the contralateral SC and the FFH by stimulating the SCs and the FFHs on both sides (Fig. 8). Figure 8, B and C shows an example of intraaxonal spikes evoked by stimulation of the left FFH (Fig. 8B, site 2), the left SC (Fig. 8C, sites 5 and 6), and the right SC (Fig. 8C, sites 10 and 11). Spikes were evoked in an all-or-none manner at short and fixed latencies even at threshold by stimulation of the left SC (Fig. 8C, sites 5 and 6, top traces) and the right SC (Fig. 8C, sites 10 and 11, top traces), and followed double-pulse stimuli at short intervals (Fig. 8C, sites 5, 6, 10, and 11, bottom traces). These spikes were considered to be activated directly from the left FFH and SC and the right SC. In addition, this axon was antidromically activated only from the rostromedial sites (sites 10 and 11), but not from the rostrolateral site (site 12) or the caudal sites (sites 13 and 14) in the right SC. Therefore this axon in the right TRT most likely originated from a TRN in the medial part of the left SC with its ascending axon that projected to the left FFH, and also its commissural collateral that terminated in the corresponding medial part of the right rostral SC. In a similar way, we found that 11 TRNs that projected to both the ipsilateral FFH and the contralateral TRT had axon collaterals to the contralateral SC, and lateral and medial TRNs had an axon collateral to the lateral and medial parts of the contralateral rostral SC, respectively.
We further analyzed the properties of commissural inputs to TRNs with a commissural collateral to the contralateral SC by recording intracellular potentials directly from their cell bodies in the SC. Figure 9, A–C shows an example of a TRN in the lateral part of the rostral SC. This lateral TRN was antidromically activated from the lateral part of the opposite rostral SC at 250 µA (Fig. 9B, site 9), as well as from both the ipsilateral FFH (Fig. 9B, site 1) and the contralateral IBN region (Fig. 9B, site 18). Stimulation of the rostral SC (Fig. 9C, sites 5, 6, and 10) evoked antidromic spikes, and stimulation of the lateral SC evoked large EPSPs at 200 µA in this lateral TRN (Fig. 9C, site 9). Commissural inhibition could be seen from the caudal SC (Fig. 9C, sites 7, 8, 11, and 12), but not from the rostral SC. This may have been due to masking of the inhibition by the large EPSPs. Figure 9, D–F shows an example of a TRN in the medial part of the rostral SC. In this medial TRN, antidromic spikes were evoked by stimulation of the medial part of the contralateral SC (Fig. 9F, site 5) as well as the ipsilateral FFH (Fig. 9E, site 2) and the contralateral IBN region (Fig. 9E, site 18). In addition, the commissural inhibition was larger from the lateral part of the contralateral SC (Fig. 9F, sites 9–12). As shown in Fig. 9, lateral TRNs (n = 5) had a commissural collateral that projected to the lateral SC and received commissural excitation from the lateral SC as well as the medial SC of the contralateral SC. In contrast, medial TRNs (n = 6) had a commissural collateral that projected to the medial SC and received larger commissural inhibition from the lateral SC.
If commissural excitation of TRNs is mediated by TRNs with commissural collaterals, stimulation of the FFH and the IBN region should evoke monosynaptic excitation in TRNs of the opposite SC by axon reflex by commissural collaterals. To test this possibility, we recorded postsynaptic potentials in TRNs in the left SC, while stimulating the right FFH and the left IBN region (Fig. 10). To exclude monosynaptic excitation of left TRNs by recurrent collaterals of TRNs, the TRT was transversely sectioned at a level just rostral to the right OPN region (Fig. 10, A and E). A TRN was antidromically activated from the ipsilateral FFH (Fig. 10B, site 1) and the contralateral TRT just rostral to the section (Fig. 10B, site 14). Stimulation of the contralateral rostral SC evoked monosynaptic excitation at a latency of 0.8 ms (Fig. 10C, sites 5 and 6). Stimulation of the contralateral FFH (Fig. 10D, site 4) and the ipsilateral IBN region (Fig. 10D, site 17) evoked excitation at a latency of 0.8 and 0.9 ms in the same TRN, respectively. In similar experiments, monosynaptic excitation was obtained from the contralateral FFH in 20 TRNs and from the ipsilateral IBN region in 15 TRNs. This finding provides supportive evidence that some TRNs with commissural collaterals excite TRNs in the opposite SC.
Existence of excitatory and inhibitory intratectal neurons with commissural axons terminating in the contralateral SC
To estimate the contribution of nontectal inputs to TRNs by the tectal commissure, we examined the effects of stimulation of the mesencephalon ventrolateral to the contralateral SC on commissural inputs to TRNs. In the mesencephalic reticular formation, many neurons and passing fibers of nontectal origin such as cells in the central mesencephalic reticular formation (cMRF), periparabigeminal nucleus (peri PB) (see Fig. 13), and fibers from the prepositus hypoglossi and substantia nigra project to the opposite SC through the tectal commissure. Figure 11, A–D shows an example of properties of commissural inputs to a medial TRN evoked by stimulation of this area. Stimulation of the medial part of the contralateral rostral SC evoked large monosynaptic excitation in this TRN (Fig. 11C, site 5). As the stimulation sites moved laterally in the rostral part of the SC, the evoked excitation became smaller (Fig. 11C, sites 6 and 7). Stimulation of the further lateral area just outside the SC evoked only much smaller excitation even at the stronger stimulus intensity in this TRN, and the latencies of the excitation were >1.5 ms, indicating that this excitation was disynaptic. The same test was carried out in another TRN in the lateral part of the SC, as shown in Fig. 11, E–G. In this TRN, stimulation of the contralateral SC evoked large inhibition (Fig. 11F, sites 5–9) with small excitation from the contralateral rostral SC (Fig. 11F, sites 5–7). However, stimulation of the mesencephalon just lateral to the SC evoked only very small inhibition with longer latencies (Fig. 11G, sites 10–13). Similar results were obtained in eight TRNs that were recorded in the same preparation; the excitation and the inhibition evoked by stimulation of the contralateral SC were much larger and began earlier than those evoked by stimulation of its adjacent mesencephalon where commissural fibers of nontectal origin ran extensively. These results indicate that the commissural excitation and inhibition from the contralateral SC to TRNs are mainly attributed to tectotectal cells rather than to nontectal cells.
View larger version (29K):
[in this window]
[in a new window]
|
FIG. 12. Commissural excitation and inhibition in TRNs after the elimination of extratectal inputs by chronic longitudinal section underneath the SC. A–C: a TRN in the rostromedial SC. A: experimental setup. A longitudinal section was made underneath the right SC from the level of the rostral border of the SC to its caudal border (arrow). Intracellular recording was made from TRNs, 1 wk after sectioning of input fibers to the SC. A recording site and stimulation sites are shown by an open cell and filled circles in the bottom inset of A, respectively. B: antidromic spikes evoked in a TRN by stimulation of the ipsilateral FFH (site 2) and the contralateral IBN regions (site 14). C: commissural inputs to the same TRN as in B from the contralateral SC. Even after extratectal fibers passing through the contralateral SC were eliminated, both excitation and inhibition were still evoked by stimulation of the contralateral rostral SC (sites 5–8) at 500 and 300 µA. D–F: another TRN in the central part of the rostral SC. D: experimental setup in the same animal as in A, and a recording site (open cell) in the bottom inset. E: antidromic spikes evoked by stimulation of the ipsilateral FFH (site 2) and the contralateral IBN region (site 14). F: commissural inputs from the contralateral SC. Inhibition was evoked by stimulation of all sites of the contralateral SC at 300 µA with a normal polarity (ventral tips were cathodal) (left column). Inhibition was smaller and its latency was slightly longer from the same stimulation sites, but with reversed polarity (dorsal tips were cathodal) (right column). Monosynaptic IPSPs were facilitated by double-pulse stimulation, indicating that cell bodies of inhibitory commissural neurons were located within the SC on the stimulated side.
|
|
View larger version (43K):
[in this window]
[in a new window]
|
FIG. 13. Distribution of retrogradely labeled neurons in the mesencephalon and brain stem after injection of WGA-HRP into the left SC with (B) and without longitudinal section underneath the right SC (A, control). Labeled neurons observed in three consecutive 80-µm sections are plotted in representative frontal sections at 960-µm intervals including the SN (a), cMRF (b and c), PB and peri PB (d), and spinal trigeminal nucleus (Tri) (f). Each dot indicates one labeled neuron. Cells in rectangular areas in A are enlarged in corresponding panels in the right column. Dark area in the left SC (A and B) indicates the greatest deposit of HRP reaction product. Dark area in the right SC in B indicates the section underneath the SC. MGB, medial geniculate body; SN, substantia nigra; RN, red nucleus; CP, cerebral peduncle, CG, central gray; PN, pontine nucleus; III, oculomotor nucleus; VI, abducens nerve; VII, facial nerve.
|
|
To further confirm that the commissural excitation and inhibition of TRNs were mediated by tectotectal neurons rather than nontectal neurons, we made a longitudinal section underneath the right SC that extended from 2 mm rostral to the rostral end of the SC to its caudal end in one cat (see the location of the section in Fig. 13B), and eliminated nontectal input fibers that originated from neurons outside the SC, passed through the right SC, and terminated in the left SC (Fig. 12). One week after we prepared the longitudinal section of the input fibers to the SC, we recorded from TRNs in the left SC and examined commissural inputs to them in response to stimulation of the right SC that was on the same side as the section. In the TRN shown in Fig. 12, A–C, the large excitation was evoked at latencies of 0.8–0.9 ms by stimulation of the contralateral rostral SC at 300 µA (Fig. 12C, sites 6–8), whereas the inhibition was small (Fig. 12C, sites 5 and 9). In the TRN shown in Fig. 12, D–F, the large inhibition was evoked by stimulation of all contralateral sites (Fig. 12F, sites 5–9), whereas the small excitation was evoked by stimulation of the medial sites of the rostral SC with a reversed polarity (Fig. 12F, sites 5 and 6, right column). This TRN was located in the center of the rostral SC and the inhibition was the largest from the center of the contralateral rostral SC (Fig. 12F, site 6). Furthermore, temporal facilitation of the monosynaptic IPSPs was induced by double-pulse stimulation of individual sites (Fig. 12F, sites 5–9). The presence of this facilitation supports that tectotectal neurons provide monosynaptic inhibition on TRNs in the opposite SC because this facilitation can occur only when cell bodies near a stimulating electrode are activated rather than passing fibers (Jankowska et al. 1975; Shinoda et al. 1976, 1982, 1987). In a similar way, monosynaptic EPSPs and IPSPs were confirmed in 8 and 13 TRNs, respectively, in the same operated cat, in which the commissural inputs of nontectal origin were eliminated. These results confirmed that the excitatory and inhibitory commissural connections actually originated from commissural neurons located within the SC, although we could not exclude a possibility of the existence of additional excitatory and inhibitory inputs from outside the SC.
To determine which afferents were eliminated by the longitudinal section among the contralateral collicular afferents that project through the commissure (Edwards et al. 1979; Huerta and Harting 1984a,b; May 2006; Wallace et al. 1989, 1990), we injected WGA-HRP into the medial part of the left rostral SC after the recording experiment in the same chronically operated cat as described earlier, and examined the distribution of retrogradely labeled neurons in the mesencephalon and the brain stem. In control experiments in which WGA-HRP was injected into the left SC without any section, labeled neurons were abundantly found bilaterally in the central mesencephalic reticular formation (cMRF), substantia nigra (SN), parabigeminal nucleus (PB), periparabigeminal nucleus (peri PB), and spinal trigeminal nucleus (Fig. 13A). Histological examination of the extent of the longitudinal section revealed that the section extended from the dorsal surface of the mesencephalon to the central gray matter beyond the midline. It extended from the lateral border of the SC and the medial geniculate nucleus, rostrally, and between the lateral border of the SC and the brachium of the inferior colliculus more caudally (Fig. 13B). After the longitudinal section, labeled neurons were found in all the above-cited structures on the ipsilateral side, but on the contralateral side only the PB (Fig. 13B, d and e) and the spinal trigeminal nucleus (Fig. 13Bf) were labeled. In the DISCUSSION we will explain how these two nuclei could still be excluded from the candidate nuclei responsible for the monosynaptic commissural excitation or inhibition because their projections to the opposite SC were not by the tectal commissure (Graybiel 1978b; McHaffie et al. 1986).
|
DISCUSSION
|
|---|
The present study revealed that mirror-symmetric commissural excitatory connections existed between the homonymous parts of the rostral SCs. In contrast, commissural inhibition is the dominant relationship between the medial part of one SC and the lateral part of the other SC, and vice versa. Although there is both behavioral and physiological evidences that suggest that the SCs functionally suppress each other by commissural connections, little is known about the existence of an excitatory commissural connection and its functional role. Most of the TRNs examined projected to both the FFH and the IBN region, where interstitiospinal (Isa and Sasaki 1992) and reticulospinal neurons (Kakei et al. 1994) terminating on neck motoneurons are located. Furthermore, some of these TRNs may be tectospinal neurons with a collateral projection to the IBN region (Grantyn and Grantyn 1982; Muto et al. 1996; Shinoda et al. 2006). Therefore these TRNs are considered to be involved in control of not only eye movements but also head movements. However, the following discussion will be confined to the functional roles of commissural excitation and inhibition on control of saccades.
Mediolateral distribution of the commissural excitation in the rostral SC
In a previous electrophysiological study, we showed that strong excitatory commissural connections exist between the rostral parts of the SCs (Takahashi et al. 2005b). Because the meridian of vertical saccades is represented in the rostral SC in the motor map (McIlwain 1986), we assumed that commissural excitation in the rostral SCs might play an important role not only in fixation but also in vertical saccades (Takahashi et al. 2005b). If upward-saccade neurons in one SC and downward-saccade neurons in the opposi
TOP
ABSTRACT
INTRODUCTION
