THE JOURNAL OF COMPARATIVE NEUROLOGY 366~149-162 ( 1996)

Morphology and Soma-Dendritic

Distribution of Synaptic Endings From the
Rostra1 Interstitial Nucleus of the Medial
Longitudinal Fasciculus (riMLF) on
Motoneurons in the Oculomotor and
Trochlear Nuclei in the Cat
SHWU-FEN WANG AND ROBERT F. SPENCER
Departments o f h a t o m y (S.-F.W.,R.F.S.) and Otolaryngoloa-Head and Neck Surgery (R.F.S.),
Medical College of Virginia, Virginia Commonwealth University, Richmond, Virginia 23298

ABSTRACT
The morphology and soma-dendritic distribution of anterograde biocytin-labelled rostra1
interstitial nucleus of the medial longitudinal fasciculus (riMLF) synaptic endings in the
oculomotor and trochlear nuclei have been examined by electron microscopy by using both
preembedding immunoperoxidase and postembedding immunogold methods. The results
indicate that three morphological types of riMLF synaptic endings are distinguishable on the
basis of synaptic vesicle morphology (spheroidal, pleiomorphic, or ellipsoidal) and postsynaptic
membrane specializations (asymmetrical or symmetrical j. All three morphological types of
riMLF synaptic endings establish synaptic connections predominantly with dendrites. Synaptic
endings that contain ellipsoidal synaptic vesicles have a more proximal soma-dendritic
distribution than those that contain either spheroidal or pleiomorphic synaptic vesicles.
Furthermore, all three morphological types of synaptic endings are encountered in the same
motoneuron subdivisions of the oculomotor and trochlear nuclei in the same experiments. The
findings suggest that subregions of the riMLF contain coexistent populations of excitatory and
inhibitory premotor neurons that are related to opposite directions of vertical saccadic eye
movements but that project to the same motoneuron subgroups on the ipsilateral side. Both the
morphology and the mode, pattern, and soma-dendritic distribution of saccade-related riMLF
synaptic endings that establish synaptic connections with vertical motoneurons differ from
those of excitatory and inhibitory second-order vertical vestibular synaptic endings. These
differences in the synaptic organization of riMLF and second-order vestibular inputs to
oculomotor and trochlear motoneurons may be related to differences in the information
transferred by each source, the riMLF input conveying eye-velocity signals, and the vestibular
input conveying eye-position signals. , IUSC;WiIey-Iiss. Inc.

Indexing terms: mesencephalic reticular formation, excitatory synapses, inhibitory synapses, biocytin,
electron microscopy

Motoneurons in the extraocular motor nuclei are the
final common pathway upon which afferents converge from
brainstem premotor areas that are related intimately to the
control of different types of eye movements (e.g., vestibulo-
ocular reflex, optokinetic nystagmus, smooth pursuit, sac-
cadic, and vergence). For vertical eye movements, the major
inputs to motoneurons in the oculomotor (inferior rectus,
superior rectus, and inferior oblique) and trochlear (supe-
rior oblique) nuclei are derived from the vestibular nuclei
for vertical vestibuloocular eye movements (for review,
see Spencer et al., 1992), the interstitial nucleus of Cajal
(INC; Carpenter et al., 1970; Graybiel and Hartwieg, 1974;
Buttner-Ennever and Biittner, 1978; Steiger and Buttner-
Accepted October 10, 1996.
Dr. Shwu-Fen Wang is currently at the School of Physical Therapy.
College of Medicine, National Taiwan University, Taipei, Taiwan R.O.C.
Address reprint requests to Dr. Robert F. Spencer, Department of
Anatomy, Medical College of Virginia, Virginia Commonwealth University,
1101 East Marshall Street. Richmond, VA 23298-0709.

( 1996 WILEY-LISS, INC.

 150
Ennever, 1979; Labandeira-Garcia et al., 1989; Carpenter
et al., 1992; Li et al., 1993; Robinson et al., 1994), and the
rostra1 interstitial nucleus of the medial longitudinal fascicu-
lus (riMLF; Graybiel and Hartwieg, 1974; Graybiel, 1977;
Biittner-Ennever and Biittner, 1978; Steiger and Biittner-
Ennever, 1979; Nakao and Shiraishi, 1983, 1985; Isa et al.,
1988; Labandeira-Garcia et al., 1989; Moschovakis et al.,
1990, 1991a,b; Nakao et al., 1990; Carpenter et al., 1992;
Isa and Itouji, 1992; Li et al., 1993; Moschovakis and
Highstein, 1994; Robinson et al., 1994; Shiraishi et al.,
1994). The INC and riMLF in the mesencephalic reticular
formation are related specifically to the control of vertical
saccadic eye movements (Biittner-Ennever et al., 1982;
Fukushima, 1987, 1991; Buttner-Ennever and Biittner,
1988; Hepp et al., 1989; Fukushima and Fukushima, 1992;
Isa and Naito, 1994; Moschovakis and Highstein, 1994).
Reciprocal excitatory and inhibitory synaptic connections
of second-order vestibular neurons with motoneurons in
the oculomotor and trochlear nuclei provide the physiologi-
cal basis for the vertical vestibuloocular reflex (Highstein
and Ito, 1971; Precht and Baker, 1972; Berthoz et al., 1973;
Highstein, 1973; Uchino et al., 1978; Iwamoto et al., 1990).
The excitatory second-order vestibular input terminates
predominantly on the distal dendrites of oculomotor and
trochlear motoneurons (Dememes and Raymond, 1980;
Spencer and Baker, 1983). By contrast, second-order inhibi-
tory vestibular neurons establish synaptic connections pre-
dominantly with the somata and proximal dendrites of
motoneurons in the oculomotor and trochlear nuclei (Bak
et al., 1976; Dememes and Raymond, 1980; Spencer and
Baker, 1983). Although both excitatory and inhibitory
synaptic effects have been observed in oculomotor and
trochlear motoneurons following electrical stimulation of
the region of the riMLF (Schwindt et al., 1974; Nakao and
Shiraishi, 1983, 1985), neither the morphology nor the
mode (i.e., single vs. multiple synaptic contact zones),
pattern (i.e,, single vs. multiple postsynaptic targets), or
soma-dendritic distribution of these inputs have been exam-
ined. Thus, it is unknown how the saccade-related inputs
interact with the vestibular inputs and to what extent these
inputs correspond to the known types of synaptic endings in
the oculomotor (Tredici et al., 1976; Waxman and Pappas,
1979) and trochlear (Bak and Choi, 1974) nuclei.
In the present study, the morphology and synaptic
organization of riMLF synaptic endings in the oculomotor
and trochlear nuclei labelled by anterograde transport of
biocytin have been examined by electron microscopy by
using preembedding peroxidase and postembedding colloi-
dal-gold localization procedures. Quantitative analyses have
been used to determine the soma-dendritic distribution of
presumed excitatory and inhibitory riMLF synaptic end-
ings on oculomotor and trochlear motoneurons.
MATERIALS AND METHODS
The data obtained in this study were derived from
experiments performed in a previous study (Wang and
Spencer, 1995) in which biocytin was injected stereotax-
cally into different regions of the riMLF to anterogradely
label synaptic endings in the oculomotor and trochlear
nuclei. Two groups of sections from those experiments,
each representing 150 IJ-m intervals through the oculomo-
tor and trochlear nuclei, were processed for the ultrastruc-
tural localization of biocytin by using preembedding or
postembedding procedures, as described below.
S.-F. WANG AND R.F. SPENCER
Preembedding localization of biocytin
One group of sections was incubated for 2 hours in avidin
D-horseradish peroxidase (HRP; Vector; 1 5 0 0 ) in 0.1 M
sodium phosphate buffer containing 0.1% Triton X-100.
For the histochemical demonstration of HRP, sections were
treated in 0.05% 3,3’-diaminobenzidine tetrahydrochloride
(DAB; Aldrich) and 0.01% hydrogen peroxide with 0.005%
cobalt acetate and 0.005% nickel chloride in 0.1 M sodium
phosphate buffer, pH 7.2, for 5-10 minutes. After washing
through several changes of buffer, sections were processed
for electron microscopy. Sections were postfixed in 1.0%
osmium tetroxide in 0.1 M phosphate buffer with 7%
dextrose for 1 hour at 4°C and were stained en bloc with 2%
uranyl acetate in 0.05 M maleate buffer for 1 hour at 4°C.
Then, sections were dehydrated in graded methanols and
propylene oxide, infiltrated with plastic resin (Fullam),and
flat embedded between two glass microscope slides coated
with a water-soluble release agent (Electron Microscopy
Sciences). After curing, the embedded sections were at-
tached to prelabelled BEEM capsules, removed from the
slides, and trimmed to include regions of the oculomotor
and trochlear nuclei that were determined in the previous
light microscopic study to contain anterogradely labelled
terminals. Semithin (0.2 IJ-m)sections were cut with glass
knives on an LKB Ultratome IV ultramicrotome and were
stained with 0.1% toluidine blue in 1.0%sodium borate for
light microscopy as a reference for the electron microscopic
analysis. Ultrathin (60-80 nm) sections were cut with a
diamond knife on the ultramicrotome, collected on Formvar-
coated, single-slot copper grids, and stained with 1.0%
uranyl acetate in distilled water and 0.1% lead citrate in 0.1
N sodium hydroxide. Sections were examined and photo-
graphed using a Zeiss EM-10CA electron microscope.
Postembedding localization of biocytin
Another group of untreated sections was processed for
electron microscopy, as described above. Serial ultrathin
(60-90 nm) sections were cut on the ultramicrotome and
collected on Formvar-coated, single-slot, gold grids. The
sections were pretreated for 6-7 minutes in 1.0% sodium
metaperiodate (Fluka) in distilled water, washed through
three changes of distilled water, and then immersed in two
changes of 0.05 M Tris-buffered saline (TBS) containing
0.1%bovine serum albumin (BSA)for 15 minutes each. The
sections were placed on a drop of 15 nm gold-tagged
streptavidin (Jansen)diluted 1:20 with 0.05 M TBS contain-
ing 0.1% BSA, pH 8.2, in a ceramic dish for 2 hours with
constant agitation a t room temperature. The grids were
washed in two changes of 0.05 M TBS followed by three
changes of distilled water. All solutions used in the process-
ing were filtered through a 0.2 IJ-m Acrodisc (Gelman)
syringe filter before use. Sections were counterstained with
uranyl acetate and lead citrate and then were examined and
photographed with the electron microscope.
Quantitative analysis
The cross-sectional areas of biocytin-labelled synaptic
endings localized by both the preembedding and postembed-
ding procedures and the diameters of the postsynaptic
profiles were measured from electron micrographs at a
magnification of ~ 2 4 , 0 0 0using a BIOQUANT (R and M
Biometrics) digital image analysis system operating on a
Hewlett-Packard Vectra microcomputer with standard mor-
phometric, spreadsheet, and statistical software. Colloidal-
 riMLF SYNAPTIC CONNECTIONS WITH OCULOMOTOR AND TROCHLEAR MOTONEUKONS 151
gold particles overlying synaptic endings labelled by the postsynaptic dendritic profile (Fig. SA,B,D). Occasionally,
postembedding method were counted, and their density multiple synaptic contact zones were associated with axoso-
was calculated as the number of particles per ym2. Diam- matic synaptic endings of this variety (Fig. 3C).
eter measurements of the proximal dendrites of choline A second population of riMLF synaptic endings contained
acetyltransferase (ChAT)-immunoreactive trochlear moto- predominantly flattened or ellipsoidal synaptic vesicles and
neurons were made directly from the light microscope at a established symmetrical synaptic contacts (Fig. 4A-D).
magnification of x 1,350 by using images captured from a Synaptic contact zones were characterized by a modest or
DageiMTI Newvicon NC-70 video camera with a Truevi- inconspicuous postsynaptic densification (Fig. 4D, inset).
sion Targa M8 frame-grabber board interfaced to the Synaptic endings of this variety most often established
microcomputer and image-analysis software. synaptic contacts with dendrites (Fig. 4B-E) and, to a lesser
extent, with somata (Fig. 4A) and spine-like profiles (Fig.
4A,B).
RESULTS The third population of riMLF synaptic endings con-
Morphology of biocytin-labelled riMLF tained pleiomorphic synaptic vesicles and established synap-
synaptic endings tic contacts with a modest postsynaptic thickening along
the postsynaptic membrane (Fig. 5A, inset). Whereas most
With the preembedding method, biocytin-labelled riMLF
of the synaptic vesicles were spheroidal, these were smaller
synaptic endings were recognized easily by dense peroxi-
than the uniformly spheroidal synaptic vesicles in the
dase reaction product that permitted their unequivocal
synaptic endings that established obvious asymmetrical
identification compared to unlabelled synaptic endings.
contacts. Synaptic endings in this category established
Labelled synaptic endings varied considerably in profile in
synaptic contacts predominantly with one (Fig. 5B,C) or
apparent direct relation to their soma-dendritic distribu-
more (Fig. 5A,D) dendrites.
tion. Most of the labelled synaptic endings associated with
somata (Fig. 1A-C) and proximal dendrites (Fig. 1D) were
dome shaped and tended to form shallow depressions in the
Size and labelling density of biocytin-labelled
surface membrane of the postsynaptic profile. At these
riMLF synaptic endings
proximal sites, labelled synaptic endings occurred either The riMLF synaptic endings in the oculomotor and
singly (Fig. 1A) or infrequently in a series reminiscent of en trochlear nuclei, as a population, were relatively heteroge-
passant boutons that arise from a single axonal arboriza- neous in size. The cross-sectional areas of 168 riMLF
tion (Fig. 1B).Labelled synaptic endings also were observed synaptic endings that were labelled by using the preembed-
occasionally in association with somatic spine-like append- ding procedure ranged from 0.30 to 6.11 Fm2 (Fig. 61, with a
ages (Fig. l C ) , particularly in the superior rectus subdivi- mean of 1.67 i- 1.19 Fm'. Without regard to differences in
sion of the oculomotor nucleus and in the trochlear nucleus. synaptic vesicle morphology, as a group, the cross-sectional
In all cases, riMLF synaptic endings on the motoneuron areas of 384 riMLF synaptic endings that were labelled by
somata and proximal dendrites, which were identified on using the postembedding procedure ranged from 0.03 to
the basis of elaborate arrays of granular endoplasmic 9.46 Fm2 (Fig. 6), with a mean of 2.47 2 1.46 Fm2.
reticulum, were associated with only one postsynaptic Although the synaptic endings that were labelled by the
profile. preembedding procedure were significantly smaller than
The overwhelming majority of riMLF labelled synaptic those labelled with the postembedding procedure ( P <
endings were encountered in the neuropil of the oculomotor 0.01; t test), the difference between the two samples was
and trochlear nuclei. Synaptic endings almost invariably attributable most likely to sampling bias toward larger
were dome shaped when they were associated with only a terminals. Smaller terminals labelled with the small colloi-
single postsynaptic dendritic profile (Fig. 2A,C,E). Most of dal-gold particles were more difficult to find in the electron
the labelled synaptic endings, however, were associated microscope.
with two or more small- and/or medium-caliber dendrites No significant differences were observed in the gold-
(Fig. 2A-D) and, consequently, had irregular profiles. Irre- particle density associated with the three categories of
spective of their soma-dendritic location, riMLF synaptic riMLF synaptic endings labelled with the postembedding
endings typically contained numerous small mitochondria procedure. The particle density of synaptic endings that
clustered near the center of each terminal. Labelled myelin- contained spheroidal synaptic vesicles ranged from 4.44 to
ated axons also were observed in the neuropil, usually in 8.86 particles/pm2, with a mean of 6.04 * 1.02 particles/
close proximity to the labelled synaptic endings (Figs. lD, km2. The density of gold particles overlying synaptic end-
2E). ings that contained flattened or ellipsoidal synaptic vesicles
The unobstructed view of the ultrastructural features of ranged from 4.86 to 8.04 particlesipm2, with a mean of
the synaptic endings labelled by the postembedding colloidal- 5.96 ? 1.32 partic1esiIJ-m'. Synaptic endings that contained
gold procedure permitted the classification of riMLF synap- pleiomorphic synaptic vesicles exhibited a gold-particle
tic endings into three populations on the basis of synaptic density that ranged from 4.23 to 8.33 particles/pm2, with a
vesicle morphology. All three populations of anterograde mean of 6.46 * 1.22 particles/km2. For comparison, back-
biocytin-labelled riMLF synaptic endings coexisted within ground levels measured from unlabelled synaptic endings
the same motoneuron subdivisions that were targeted by in neighboring regions of the oculomotor nucleus (e.g.,
injections in different regions of the riMLF. One population medial rectus subdivision) ranged from 0.22 to 1.21 par-
*
of riMLF synaptic endings contained large spheroidal synap- ticles/pm2, with a mean of 0.41 0.31 particles/Fm2. The
tic vesicles and formed asymmetrical synaptic contacts (Fig. quantitative data, therefore, indicate that biocytin labelled
3A-D). Synaptic contact zones were characterized by a the different populations of riMLF synaptic endings in an
prominent postsynaptic densification with occasional sub- approximately equivalent manner without regard to their
junctional dense bodies (Fig. 3A, inset). In most instances, presumed physiological action or neurotransmitter con-
only a single synaptic contact was observed with each tent.
 Fig. 1. A-D: Electron micrographs of biocytin-labelled (preembed-
ding) synaptic endings from the rostra1 interstitial nucleus of the
medial longitudinal fasciculus (riMLF) associated with somata (A-C)
and a proximal dendrite (d; D ) in the oculomotor and trochlear nuclei.
Although relatively few in number, labelled axosomatic synaptic end-
ings from the riMLF a r e observed either in isolation (A) or as a series
(B). Less frequently, synaptic connections a r e established with somatic
spine-like ( s )appendages (C). Note the labelled myelinated axons in D.
Asterisks indicate unlabelled synaptic endings. Scale bars = 1 km.
riMLF SYNAPTIC CONNECTIONS WITH OCULOMOTOR AND TROCHLEAR MOTONEURONS 153

Fig. 2. A-E: Electron micrographs of biocytin-labelled (preembedding) synaptic endings from the
riMLF contacting medium-and small-diameter dendrites (d) in the oculomotor and trochlear nuclei.
Individual synaptic endings in A-D are associated with two or more postsynaptic profiles. Note the labelled
myelinated axon ( a )in E. Scale bars = 1 ym.
154
Fig. 3 . A-D: Electron micrographs of presumed excitatory biocytin-
labelled riMLF synaptic endings containing spheroidal synaptic vesicles
and forming asymmetric synaptic contact profiles (large arrows) in the
oculomotor and trochlear nuclei. Biocytin is localized by the postembed-
ding procedure using 15 nm colloidal-gold particles (small arrows).
Soma-dendritic distribution of
biocvtin-labelled riMLF svnaDtic endinmY I Y
Most biocytin-labelled riMLF synaptic endings were en-
countered in the neuropil, and comparatively fewer labelled
synaptic endings were associated with the somata of oculo-
motor and trochlear motoneurons. A quantitative analysis
of the postsynaptic diameters of profiles that were in
synaptic contact with labelled synaptic endings was under-
taken to determine the soma-dendritic distribution of riMLF
synaptic endings. With the preembedding procedure, the
diameters of postsynaptic profiles ranged from 0.17 to
11.79 pm (Fig. 7), with a mean of 3.22 2 2.63 pm. With the
postembedding procedure, postsynaptic profiles ranged from
0.20 to 14.51 pm in diameter (Fig. 71,with a mean of 2.54 2
2.36 pm.
S.-F. WANG AND R.F. SPENCER
Labelled synaptic endings often contact more than one medium-
diameter and/or small-diameter dendrite (A,B,D). Labelled axosomatic
synaptic endings with similar synaptic vesicle morphology (C) were
encountered only rarely. d, Dendrite; psd, postsynaptic densification.
Scale bars = 0.5 pm in A-D; 0.25 pm in inset.
For comparison, the diameters of 252 proximal dendrites
oftrochlear motoneurons that were labelled by the immuno-
histochemical localization of ChAT (McHaffie et al., 1991)
were measured by light microscopy. The diameters of
proximal dendrites ranged from 2.42 to 9.59 pm (Fig. 7),
with a mean of 5.36 2 1.48 pm. The 95% confidence
interval (CI) ranged from 2.06 to 8.66 pm. Using the CI of
ChAT-immunoreactive proximal dendrites as a reference,
postsynaptic profiles were characterized as being somata
( > 8.66 pm), proximal dendrites (2.06-8.66 pm), and distal
dendrites ( < 2.06 pm), including spines. These results
indicated that the majority (92.86%)of the synaptic endings
labelled with the preembedding procedure contacted den-
drites, although very few (7.14%)contacted somata (Figs. 7,
8 ) .Axodendritic synaptic endings established synaptic con-
 Fig. 4. A-E: Electron micrographs of presumed inhibitory biocytin-
labelled riMLF synaptic endings containing ellipsoidal synaptic vesicles
and forming symmetrical synaptic contact profiles (large arrows) in the
cat oculomotor and trochlear nucleus. Biocytin is localized by the
postembedding procedure using 15 nm colloidal-gold particles (small
arrows). Labelled synaptic endings are associated occasionally with
somata or spine-like appendages (A),but, more frequently, they contact
medium- and small-diameter dendrites id; B-E). d, Dendrite; s, spine-
like appendage; psd, postsynaptic densification. Asterisks indicate
unlabelled synaptic endings. Scale bars = 0.5 pm, 0.25 pm in inset.
 156 S.-F. WANG AND R.F. SPENCER

Fig. 5. A-D: Electron micrographs of biocytin-labelled riMLF syn-
aptic endings containing pleiomorphic synaptic vesicles and forming
synaptic contacts (large arrows) in the oculomotor and trochlear nuclei.
Biocytin is localized by the postembedding procedure using 15 nm
colloidal-gold particles (small arrows). Labelled synaptic endings in A,
B, and D contact more than one medium- and small-diameter dendrite.
The ultrastructural features (e.g., size and morphology of synaptic
vesicles, postsynaptic densification) vary but are distinct from those
that characterize typical excitatory and inhibitory synaptic endings (see
Figs. 2 , 4 ) .d, Dendrite; psd, postsynaptic densification. Scale bars = 0.5
p m , 0.25 pm in inset.

tacts with a greater weighting toward proximal (52.98%) spines were characterized by their sac-like protrusions from
than distal (39.88%)dendrites. However, the three popula- a soma or a dendrite, respectively (Figs. lC, 4A,B). Classify-
tions of synaptic endings labelled with the postembedding ing the postsynaptic profiles by their ultrastructural fea-
procedure overall made synaptic contacts with a greater tures, the soma-dendritic distributions were determined in
weighting toward distal (62.83%) than proximal (32.74%) the trochlear nucleus for the three categories of riMLF
dendrites (Figs. 7, 8). synaptic endings characterized on the basis of synaptic
Because the diameters measured from single ultrathin vesicle morphology (Fig. 9). Using these criteria, the major-
sections were dependent on the plane of the section through ity (53.10%) of the labelled riMLF synaptic endings con-
the process, postsynaptic profiles were characterized subjec- tacted distal dendrites. A smaller proportion of the synaptic
tively on the basis of their ultrastructural features. For endings contacted proximal dendrites (27.43% ), somata
example, a distinction was made between proximal and (7.96%), and spine-like appendages (11.50%).The majority
distal dendrites by the content of granular endoplasmic of labelled synaptic endings that contained spheroidal
reticulum, Golgi apparatus, and polyribosomes. Proximal (61.11%,)and pleiomorphic (72.41%)synaptic vesicles con-
dendrites were distinguished from somata by their content tacted distal dendrites. Very few of these two types of
of parallel arrays of microtubules. Somatic or dendritic synaptic endings contacted proximal dendrites ( 16.67%
riMLF SYNAPTIC CONNECTIONS WITH OCULOMOTOR AND TROCHLEAR MOTONEURONS 157

0
g 0
~0 0
~ 0 0
y 0
"~
0 0
x"
0 0
~" 0
,~
0 0
,~ , & ~
A

Cross-Sectional Area (pm2)

Fig. 6. Distribution of cross-sectional areas of biocytin-labelled and the median is 1.28 pm2. The mean size of 384 riMLF synaptic
riMLF synaptic endings in the trochlear nucleus localized by pre- and endings labelled with the postembedding procedure is 2.47 ? 1.46 pm2,
postembedding procedures. The mean size of 168 riMLF synaptic and the median is 2.19 km2.
endings labelled with the preembedding procedure is 1.67 t 1.19 pm',

Post-embedding

ChAT

9 ~ 9 ~ 9 ~ 9 ~ 9 " 9 ~ 9 ~ 9 ~ 9 ~
o o - - ~ ~ m o ~ ~
7
A
Postsynaptic Diameter (vm)
Fig. 7. Diameter distribution of postsynaptic profiles contacted by that are smaller in diameter than t h e proximal dendrites of the
riMLF synaptic endings labelled with the pre- and postembedding motoneurons. Synaptic endings labelled by the preembedding proce-
procedures in the trochlear nucleus compared to the diameter distribu- dure have a wider postsynaptic distribution than those labelled by the
tion of proximal dendrites of choline acetyltransferase (ChATI- postembedding procedure. Very few of the labelled synaptic endings
immunoreactive trochlear motoneurons. Note that the overall distribu- make synaptic contact on the soma.
tion ofriMLF synaptic endings is weighted toward postsynaptic profiles

spheroidal, 17.24% pleiomorphic). The synaptic endings synaptic endings of each type were located on the somata
that contained ellipsoidal synaptic vesicles were weighted (11.11% spheroidal, 3.45% pleiomorphic, 8.33% ellipsoidal) of
more toward proximal dendrites (41.67%) than toward trochlearmotoneuronsoron somaticordendriticspines (11.11%
distal dendrites (35.42%). By contrast, significantly fewer spheroidal, 6.90% pleiomorphic, 14.58% ellipsoidal).
158 S.-F. WANG AND R.F. SPENCER

95 % Confidence IntervaI Distribution asymmetrical, with a prominent postsynaptic densification.

A second population of riMLF synaptic endings contains
predominantly ellipsoidal synaptic vesicles and exhibits a
symmetrical pre-/postsynaptic membrane profile. The third
category of riMLF synaptic endings has ultrastructural
features that are intermediate between those of the spheroi-
dal and those of the ellipsoidal types. These synaptic
endings contain pleiomorphic synaptic vesicles and are
associated with a modest postsynaptic specialization. All
three morphological types of riMLF synaptic endings estab-
lish synaptic connections predominantly with dendrites.
Synaptic endings containing ellipsoidal synaptic vesicles
have a more proximal soma-dendritic distribution than
those that contain either spheroidal or pleiomorphic synap-
tic vesicles. Furthermore, all three morphological types of
synaptic endings were encountered in the same motoneu-
ron subdivisions of the oculomotor and trochlear nuclei in
the same experiments.
A <2.06 2.06-8.66
Postsynaptic Diameter (pm)
~3.66
To characterize the soma-dendritic distribution of riMLF
synaptic endings, two methods have been used to classify
proximal and distal dendrites. One method is based on the
~ p ~ . ~ ~ - . diameter of the postsynaptic dendrites with reference to the
I Pre-embedding i CI of the diameter distributions of proximal dendrites of
i ~
ChAT-immunoreactive trochlear motoneurons. The other
~ Post-embedding method is based on the ultrastructural features of the
lp-pppp- 1 postsynaptic dendrites, including the content of granular
endoplasmic reticulum, Golgi apparatus, and ribosomes.
For the sample population in this study, both methods
Ultrastructural Distribution produced very similar results. In general, therefore, at least
6o ~~ ~- ~~
in the present situation, the diameters of the postsynaptic
1
p ~ - ~ -~p~~

dendrites can serve as a useful indicator of the proximodis-

Distal Proximal Soma
B Soma-Dendritic Distribution
Fig. 8. Comparison of the soma-dendritic distributions of riMLF
synaptic endings labelled by the preembedding and postembedding
methods with reference to the 95% confidence interval of postsynaptic
diameters of ChAT-immunoreactive proximal dendrites of trochlear
motoneurons ( A )and the identification of postsynaptic profiles based
on ultrastructural criteria ( B ) .
DISCUSSION
The findings in this study indicate that three morphologi-
cal types of anterograde, biocytin-labelled, riMLF synaptic
endings, which are distinguishable on the basis of synaptic
vesicle morphology and synaptic specializations, establish
synaptic connections with motoneurons in the cat oculomo-
tor and trochlear nuclei. One type of riMLF synaptic ending
contains a uniform population of large, spheroidal, synaptic
vesicles, and the synaptic contact zones are distinctly
tal location of synaptic endings on the soma-dendritic
surfaces.
Two of the morphological types of riMLF synaptic end-
ings in the trochlear nucleus are similar in ultrastructural
features to unlabelled synaptic endings that were identified
previously (Bak and Choi, 1974; Tredici et al., 1976;
Waxman and Pappas, 1979). The riMLF synaptic endings
that contain ellipsoidal synaptic vesicles appear to corre-
spond to types I and I1 (Bak and Choi, 1974), which are
characteristic of inhibitory synapses (Uchizono, 1965; Lar-
ramendi et al., 1967). By contrast, riMLF synaptic endings
that contain spheroidal synaptic vesicles are similar to type
I11 synapses in the trochlear nucleus (Bak and Choi, 1974),
which are considered to be excitatory in nature (Uchizono,
Spine 1965; Larramendi et al., 1967). The third population of
riMLF synaptic endings contains pleiomorphic synaptic
vesicles and does not make typical symmetrical or asym-
metrical synaptic contacts. The nature of this population of
synaptic ending is unknown.
Synaptic endings of types 1-111 have been identified
previously as originating from the vestibular nuclei (Bak et
al., 1976; Demcmes and Raymond, 1980). However, the
ipsilateral second-order, inhibitory, vestibular input to
oculomotor and trochlear motoneurons is distributed pre-
dominantly on the somata and the proximal dendrites (Bak
et al., 1976; Demcmes and Raymond, 1980; Spencer and
Baker, 1983), whereas presumed inhibitory synaptic end-
ings from the ipsilateral riMLF have a more widespread
distribution on proximal and distal dendrites. Further-
more, the mode (i.e., single vs. multiple synaptic contact
zones) and pattern (i.e., single vs. multiple postsynaptic
targets) of synaptic termination of riMLF inhibitory synap-
tic endings differ from those previously demonstrated for
the second-order inhibitory vestibular synapses. Each inhibi-
riMLF SYNAPTIC CONNECTIONS WITH OCULOMOTOR AND TROCHLEAR MOTONEURONS
Distal Proximal
Soma-Dendritic Distribution
Fig. 9. Soma-dendritic distribution of different categories of riMLF synaptic endings characterized by
synaptic vesicle morphology Le., ellipsoidal, pleiomorphic, and spheroidal) in the trochlear nucleus labelled
with the postembedding procedure. The postsynaptic profiles are classified into distal and proximal
dendrites, soma, and spine according to their ultrastructural features.
tory riMLF synaptic ending exhibits only a single synaptic
contact zone with each postsynaptic profile, but synaptic
connections usually are established simultaneously with
multiple postsynaptic dendrites. By contrast, each inhibi-
tory vestibular synaptic ending exhibits multiple, spatially
separated synaptic contact zones with only one postsynap-
tic profile (Spencer and Baker, 1983). Consistent with the
notion of multiple types of inhibitory synaptic endings,
recent findings have demonstrated at least two types of
synaptic endings in the cat oculomotor and trochlear nuclei
that are immunoreactive toward glutamate decarboxylase,
the synthesizing enzyme of the putative inhibitory neuro-
transmitter GABA, and that differ on the basis of the mode,
pattern, and soma-dendritic distribution of their synaptic
connections (Spencer et al., 1992). Furthermore, these
findings are compatible with the notion that synaptic
endings with distinctly different ultrastructural and/or
synaptic features originate from different sources (Rose and
Neuber-Hess, 1991).
The type I11 contralateral, excitatory, second-order, ves-
tibular synaptic endings on oculomotor (DemGmes and
Raymond, 1980; Spencer and Baker, 1983) and trochlear
(Bak et al., 1976; Spencer and Baker, 1983) motoneurons
are distributed predominantly on dendrites, although the
overall soma-dendritic weighting k e . , proximal vs. distal
dendrites) of this input has not been determined. Although
the excitatory riMLF synaptic endings also are distributed
on the dendrites of oculomotor and trochlear motoneurons,
this input terminates preferentially on small- and medium-
caliber distal dendrites. Based on evidence for the spatial
segregation of inputs on the soma-dendritic surface of
spinal motoneurons (Rose and Neuber-Hess, 1991), the
proximodistal distribution of vestibular and riMLF inputs
to oculomotor and trochlear motoneurons might be ex-
pected to differ as well. Such a segregation of inputs on
different portions of the soma-dendritic extent of a neuron,
159
1 0 Pleiomorphic I
1I ‘1
r’
Spheroidal
. -
.-
,
I
1
Soma Spine
however, does not preclude the possibility of interactions
between the inputs that might influence specific individual
or combined synaptic effects (White et al., 1990; Tomasulo
et al., 1993).
The soma-dendritic distribution of riMLF synaptic end-
ings is likely to have a significant role in influencing the
postsynaptic physiological responses of oculomotor and
trochlear motoneurons. Given the relatively constant elec-
trotonic location of the riMLF excitatory input onto oculo-
motor and trochlear motoneurons, the shape of the EPSPs
produced by activation of the vertical saccadic premotor
neurons should be similar for all motoneurons that receive
excitatory synaptic inputs from the riMLF (Liischer and
Clamann, 1992). The time course of the excitatory postsyn-
aptic potentials (EPSPs) also should be similar for all
motoneurons, because the spatial and temporal dispersion
of the input appears to be minimal (Walmsley and Stuklis,
1989). Other factors, such as the size-related rheobase and
input resistance of the postsynaptic motoneurons, might
influence the modulation of EPSP amplitude by high-
frequency stimulation (Koerber and Mendell, 1991) and,
presumably, play a role in the recruitment of the motoneu-
rons during eye movements.
Given the divergence of individual riMLF synaptic inputs
to oculomotor and trochlear motoneurons, as indicated by
single synaptic endings that establish simultaneous synap-
tic connections with multiple postsynaptic profiles, the
presynaptic control of motoneuron recruitment during
vertical saccadic eye movements is likely to be less impor-
tant than that proposed for second-order excitatory vestibu-
lar inputs that function in the vertical vestibuloocular
reflex (Spencer and Baker, 1983). One intriguing possibility
is that the differences in the synaptic organization of riMLF
and second-order vestibular inputs to oculomotor and
trochlear motoneurons may be related to differences in the
information transferred by each source, the riMLF input
 160
conveying eye-velocity signals (Buttner et al., 1977; King
and Fuchs, 1979; Vilis et al., 1989; Moschovakis et al.,
1991a,b) and the vestibular input conveying eye-position
signals (McCrea et al., 1980; 1987a,b; Ohgaki et al., 1988;
Iwamoto et al., 1990; Scudder and Fuchs, 1992). Further-
more, these morphological and physiological differences
between the riMLF and vestibular inputs to vertical moto-
neurons translate to distinctive clinical deficits in vertical
gaze following lesions in the midbrain a t the thalamomesen-
cephalic junction (Christoff, 1974; Cogan, 1974; Halmagyi
et al., 1978; Jacobs et al., 1978; Kompf et al., 1979;
Trojanowski and LaFontaine, 1981; Buttner-Ennever et
al., 1982; Pierrot-Deseilligny et al., 1982;Moffie et al., 1983;
Bogousslavsky and Regli, 1984; Ranalli et al., 1988; Deleu
et al., 1989; Bogousslavsky et al., 1990; Thomke and Hopf,
1992; Green et al., 1993) vs. lesions in the medulla (Meien-
berget al., 1978).
The synaptic organization of riMLF terminals in the
oculomotor and trochlear nuclei that are related to vertical
saccadic eye movements also differs substantially from the
organization of pontomedullary reticular inputs to abdu-
cens neurons that are related to horizontal saccadic eye
movements. In the present study, riMLF synaptic endings
establish synaptic contacts predominantly with dendrites
(particularly distal dendrites). By contrast, synaptic end-
ings from excitatory and inhibitory burst neurons in the
pontomedullary reticular formation terminate on the so-
mata and proximal dendrites of abducens neurons
(Destombes and Rouviere, 1981). The reason for this
difference in the soma-dendritic distribution of reticular
inputs to vertical vs. horizontal motoneurons is unclear,
unless the electrolytic lesions that were employed in the
latter study disrupted fibers of passage (particularly those
from the vestibular nuclei). A more intriguing possibility is
that these differences also may be related to different
neurotransmitters that are utilized by premotor neurons
involved in the control of horizontal vs. vertical eye move-
ments (e.g., glycine vs. GABA; Spencer et al., 1989, 1992;
Spencer and Baker, 1992; Spencer and Wang, 1996).
More significantly, the present study has demonstrated
that both excitatory and inhibitory inputs from the riMLF
establish synaptic connections with the same motoneuron
subgroups ipsilaterally. This arrangement, as such, con-
trasts with the reciprocal organization of excitatory and
inhibitory connections that characterize the horizontal
eye-movement systems (Escudero and Delgado-Garcia,
1988) and the vertical vestibuloocular reflex (Highstein and
Ito, 1971; Precht and Baker, 1972; Berthoz et al., 1973;
Highstein, 1973; Uchino et al., 1978; Iwamoto et al., 1990).
The results from a previous study indicate a topographical
organization of premotor neurons in the riMLF that project
to vertical-upward or vertical-downward motoneurons in
the oculomotor and trochlear nuclei (Wang and Spencer,
1996). For example, rostral regions of the riMLF target
predominantly inferior rectus and superior oblique moto-
neurons. The present study has documented that this
projection to the trochlear nucleus has both excitatory and
inhibitory components. Consistent with this finding, electri-
cal stimulation of the riMLF region elicits both monosynap-
tic EPSPs and inhibitory postsynaptic potentials (IPSPs) in
oculomotor and trochlear motoneurons (Schwindt et al.,
1974; Nakao and Shiraishi, 1983, 1985).Given the fact that
physiological studies have demonstrated that neurons dis-
charging in relation to vertical-upward or vertical-down-
ward saccadic eye movements are intermingled within the
S.-F. WANG AND R.F. SPENCER
riMLF (Buttner et al., 1977; King and Fuchs, 1979; Vilis et
al., 1989; Moschovakis et al., 1991a,b; Crawford and Vilis,
1992), it is logical to assume that a single region of the
riMLF contains excitatory neurons that are related to one
vertical axis of movement and contains inhibitory neurons
that are activated during movements in the opposite direc-
tion. This arrangement would require that supranuclear
inputs to the riMLF must distinguish between the two
populations of neurons that are related to opposite direc-
tions of vertical saccadic eye movements.
ACKNOWLEDGMENTS
This study was supported by U S . Public Health Service
MERIT Award EY02191 from the National Eye Institute,
National Institutes of Health. The excellent technical assis-
tance of Lynn Davis and Nancy Smith is greatly appreci-
ated.
LITERATURE CITED
Bak, I.J., and W.B. Choi (1974) Electron microscopic investigation of
synaptic organization of the trochlear nucleus in cat. I. Normal ultrastruc-
ture. Cell Tissue Kes. 150:409-423.
Bak, I.J., R. Baker, W.B. Choi, and W. Precht (1976) Electron microscopic
investigation of the vestibular projection to the cat trochlear nuclei.
Neuroscience 1.477-482.
Berthoz, A., R. Baker, and W. Precht (1973)Labyrinthine control of inferior
oblique motoneurons. Exp. Brain Res. 18:225-241
Bogousslavsky, J., and F. Regli (1984) Upgaze palsy and monocular paresis
of downward gaze from ipsilateral thalamo-mesencephalic infarction: A
vertical “one-and-a-half’ syndrome. J . Neurol. 231:4345.
Bogousslavsky, J . , J . Miklossy, F. Regli, and R. Janzer (1990)Vertical gaze
palsy and selective unilateral infarction of the rostral interstitial nucleus
of the medial IonL~tudinalfasciculus (riMLF).J . Neurol. Neurasurg.
Psychiatr. 53r67-71.
Buttner, U., J.A. Biittner-Ennever, and V. Henn (1977) Vertical eye
movement related unit activity in the rostral mesencephalic reticular
formation of the alert monkey. Brain Res. 130:239-252.
Buttner-Ennever, J.A., and U. Buttner (1978) A cell group associated with
vertical eye movements in the rostral mesencephalic reticular formation
of t h e monkey. Brain Res. 151:3147.
Biittner-Ennever, J.A., and U. Biittner (1988) The reticular formation. In
J.A. Biittner-Ennever (ed): Neuroanatomy of the Oculomotor System.
Amsterdam: Elsevier Science Publishers BV, pp. 119-176.
Biittner-Ennever, J.A., U. Buttner, B. Cohen, and G. Baumgartner (1982)
Vertical gaze paralysis and the rostral interstitial nucleus of the medial
longitudinal fasciculus. Brain 105:125-149.
Carpenter, M.B., J.W. Harbison, and P. Peter (1970) Accessory oculomotor
nuclei in the monkey: Projections and effects nfdiscrete lesions. J. Comp.
Neurol. 140:131-154.
Carpenter, M.B., A.B. Periera, and N. Guha (1992)Immunocytochemistry of
oculomotor afferents in the squirrel monkey (Saimiri sciureus). J .
Hirnforsch. 33151-167.
Christoff, N.A. (1974) A clinicopathologic study of vertical eye movements.
Arch. Neurol. 31:1-8.
Cogan, D.C. (1974) Paralysis of down-gaze. Arch. Ophthalmol. 91:192-199.
Crawford, J.D., and T. Vilis (1992) Symmetry of oculomotor burst neuron
coordinates about Listing’s plane. J. Neurophysiol. 68:432-448.
Deleu, D., T. Buisseret, and G. Ebinger (1989) Vertical one-and-a-half
syndrome: Supranuclear downgaze paralysis with monocular elevation
palsy. Arch. Neurol. 46:1361-1363.
Dememes, D., and J. Raymond (1980) Identification des terminaisons
vestibulaires dans les noyaux oculomoteurs communs chez le chat par
radioautographie en microscopie electronique. Brain Kes. 196:381-345.
Destombes, J . , and A. Kouviere (1981) Ultrastructural study of vestibular
and reticular projections to the abducens nucleus. Exp. Brain Res.
43253-260.
Escudero, M., and J . M . Delgado-Garcia (1988)Behavior of reticular, vestibu-
lar and prepositus neurons terminating in the abducens nucleus of the
alert cat. Exp. Brain Kes. 71:218-222.
riMLF SYNAPTIC CONNECTIONS WITH OCULOMOTOR AND TROCHLEAK MOTONEURONS
Fukushima, K. (1987)The interstitial nucleus of Cajal and its role in the
control of movements of head and eyes. Progr. Neurobiol. 29:107-192.
Fukushima, K. (1991) The interstitial nucleus of Cajal in the midbrain
reticular formation and vertical eye movement. Neurosci. Res. 10:159-
187.
Fukushima, K., and J . Fukushima (1992)The interstitial nucleus of Cajal is
involved in generating downward fast eye movement in alert cats.
Neurosci. Res. 15:299-303.
Graybiel, A.M. 11977) Organization of oculomotor pathways in the cat and
rhesus monkey. In R. Baker and A. Berthoz ieds): Control of Gaze by
Brainstem Neurons. Amsterdam: ElsevieriNorth-Holland Biomedical
Press, pp. 79-88.
Grayhiel, A.M., and E.A. Hartwieg (1974) Some afferent connections of the
oculomotor complex in the cat: An experimental study with tracer
techniques. Brain Res. #1:543-551.
Green. J .P., N.J. Newman, and J.S. Winterkorn (1993) Paralysis of down-
gaze in two patients with clinical-radiologic correlation. Arch. Ophthal-
mol. 111:219-222.
Halmagyi, G.M., A.W. Evans, and J.M. Hallinan 11978) Failure of downward
gaze: The site and nature of the lesion. Arch. Neurol. 35.22-26.
Hepp, K., V. Henn, T. Vilis, and B. Cohen (1989) Brainstem r e g o n s related
to saccade generation. In R.H. Wurtz and M.E. Goldberg ieds): The
Neurobiology of Saccadic Eye Movements. Amsterdam: Elsevier Science
Publishers BV, pp. 105-212.
Highstein, S.M. ( 1973) The organization of the vestibulo-oculomotor and
trochlear reflex pathways in the rabbit. Exp. Brain Res. 17:285-300.
Highstein, S.M., and M. Ito (1971) Differential localization within the
vestibular nuclear complex of the inhibitory and excitatory cells innervat-
ing IIIrd nucleus oculomotor neurons in rabbit. Brain Res. 29358-362.
Isa. T., and T. Itouji (1992) Axonal trajectories of single Forel’s field H
neurones in the mesencephalon, pons and medulla oblongata in the cat.
Exp. Brain Kes. 89:484495.
Isa. T., and K. Naito (19941 Activity of neurons in Forel’s field H during
orienting head movements in alert head-free cats. Exp. Brain Res.
100:187-199.
Isa, T., T. Itouji, S. Nakao, and S.Sasaki (1988) Subtypes of neurones in
Forel’s field H as defined by their axonal projection. Neurosci. Lett.
90:95-99.
Iwamoto, Y., T. Kitami, and K. Yoshida (1990) Vertical eye movement-
related secondary vestibular neurons ascending in medial longitudinal
fasciculus in cat. 11. Direct connections with extraocular motoneurons. J .
Neurophysiol. 63:918-935.
Jacobs, L., P.L. Anderson, and M.B. Bender (1978) The lesion producing
paralysis of downward but not upward gaze. Arch. Neurol. 28.319-323.
King, W.M., and A.F. Fuchs 11979) Reticular control of vertical saccadic eye
movements by mesencephalic burst neurons. J. Neurophysiol. 42:861-
876.
Koerber. H.R., and L.M. Mendell (1991) Modulation of synaptic transmis-
sion at Ia-afferent fiber connections on motoneurons during high-
frequency stimulation: Role of postsynaptic target. J. Neurophysiol.
65590-597.
Kompf, D., T. Pasik, P. Pasik, and M.B. Bender (1979) Downward gaze in
monkeys. Stimulation and lesion studies. Brain 102:527-538.
Labandeira-Garcia, J.L., M.J. Guerra-Seijas, and J.A. Labandeira-Garcia
( 1989, Oculnmotor nucleus afferents from the interstitial nucleus of
Cajil and t he region surrounding the fasciculus retroflexus in the rabbit.
Neurosci. Lett. 101:ll-16.
Larramendi, L.M.H., L. Fickenscher, and N. Lemkey-Johnston 11967)
Synaptic vesicles of inhibitory and excitatory terminals in the cerebel-
lum. Science 156:967-969.
Li, W.-B.. Y. Shiraishi, and S. Nakao (1993) Topographical organization of
cat mesodiencephalic areas for monosynaptic activation of vertical
oculomotoneurons. Exp. Brain Res. W43-52.
Luscher. H - R . , and H.P. Clamann (1992) Relation between structure and
function in information transfer in spinal monosynaptic reflex. Physiol.
Rev. 7271-99.
McCrea, K.A., K. Yoshida, A. Berthoz, and R. Baker (1980) Eye movement
related activity and morphology of second order vestibular neurons
terminating in the cat abducens nucleus. Exp. Brain Res. 40t468-473.
McCrea, R.A., A. Strassman. and S.M. Highstein l1987aJ Anatomical and
physiological characteristics of vestibular neurons mediating the vertical
vestibulo-ocular reflexes of the squirrel monkey. J. Comp. Neurol.
264.57 1-594.
McCrea. R.A., A. Strassman, E.May, and S.M. Highstein l1987b) Anatomi-
cal and physiologcal characteristics of vestibular neurons mediating the
161
horizontal vestibulo-ocular reflex of the squirrel monkey. J . Comp.
Neurol. 264:547-570.
McHaffie, J.G., M. Beninato, B.E. Stein, and K.F.Spencer (1991)Postnatal
development of acetylcholinesterase in, and cholinergic projections to,
the cat superior colliculus. J . Comp. Neurol. %13:113-131.
Meienberg, O., J. Rover, and G. Kommerell (1978) Prenuclear paresis of
homolateral inferior rectus and contralateral superior oblique eye
muscles. Arch. Neurol. 35231-233.
Moffie, D., B.W.O. de Visser, and S.Z. Stefanko (1983)Parinaud’s syndrome.
J . Neurol. Sci. 581175-183.
Moschovakis, A.K., and S.M. Highstein (19941 The anatomy and physiology
of primate neurons that control rapid eye movements. Annu. Rev.
Neurosci. 17:465-488.
Moschovakis, A.K., C.A. Scudder, and S.M. Highstein (1990) A structural
basis for Hering’s law: Projections to extraocular motoneurons. Science
248: 1118- 1 119.
Moschovakis, A.K., C.A. Scudder, and S.M. Highstein (1991al Structure of
the primate oculomotor burst generator. I. Medium-lead burst neurons
with downward on-direction. J . Neurophysiol. 65.203-217.
Moschovakis, A.K., C.A. Scudder, S.M.Highstein, and J.D. Warren (1991b)
Structure of the primate oculomotor burst generator. 11. Medium-lead
burst neurons with upward on-direction. J . Neurophysiol. 65:218-229.
Nakao, S., and Y. Shiraishi (1983)Excitatory and inhibitory synaptic inputs
from the medial mesodiencephalic junction to vertical eye movement-
related motoneurons in the cat oculomotor nucleus. Neurnsci. Lett.
42: 12.5-130.
Nakao, S., and Y. Shiraishi 11985) Direct excitatory and inhibitory synaptic
inputs from the medial mesodiencephalic junction to motoneurons
innervating extraocular oblique muscles in t h e cat. Exp. Brain Res.
61:62-72.
Nakao, S.,Y. Shiraishi, W.-B. Li, and T. Oikawa 11990)Mono- and disynaptic
excitatory inputs from the superior colliculus to vertical saccade-related
neurons in the cat Forel’s field H. Exp. Brain Res. 82:222-226.
Ohgaki, T., I S . Curthoys, and C.H. Markham 11988) Morphology of
physiologically identified second-order vestibular neurons in cat, with
intracellularly injected HRP. J. Comp. Neurol. 276:387-411.
Pierrot-Deseilligny, C., F. Chain, F. Gray, M. Serdaru, R. Escourolle, and F.
Lhermitte ( 1982) Parinaud’s syndrome. Electro-oculographic and ana-
tomical analyses of six vascular cases with deductions about vertical gaze
organization in the premotor structures. Brain 105:667-696.
Precht, W., and K. Baker (1972) Synaptic organization of the vestibulo-
trochlear pathway. Exp. Brain Res. 14:158-184.
Ranalli, P.J., J.A. Sharpe, and W.A. Fletcher (1988) Palsy of upward and
downward saccadic, pursuit, and vestibular movements with a unilateral
midbrain lesion: Pathophysiologic correlations. Neurology 3 8 114-122.
Robinson, F.R., J.O. Phillips, and A.F. Fuchs (1994) Coordination of gaze
shifts in primates: Brainstem inputs to neck and extraocular motoneu-
ron pools. J . Comp. Neurol. 346:43-62.
Rose, P.K., and M. Neuher-Hess 11991) Morphology and frequency of axon
terminals on the somata, proximal dendrites, and distal dendrites of
dorsal neck motoneurons in the cat. J . Comp. Neurol. 307:259-280.
Schwindt, P.C., W. Precht, and A. Richter ( 1974) Monosynaptic excitatory
and inhibitory pathways from medial midbrain nuclei to trochlear
motoneurones. Exp. Brain Kes. 20:223-238.
Scudder, C.A., and A.F. Fuchs 11992)Physiological and behavioral identifica-
tion ofvestibular nucleus neurons mediating the horizontal vestibuloocu-
lar reflex in trained rhesus monkeys. J. Neurophysiol. 682444264.
Shiraishi, Y., S. Matsuo, and S. Nakao (1994) Medium-lead burst neurons
related to vertical saccades in cat Forel’s field H: Input from vestibular
nucleus. Neurosci. Lett. 17224-26.
Spencer, R.F., and K. Baker (1983)Morphology and synaptic connections of
physiologcally-identified second-order vestibular axonal arborizations
related to cat oculomotor and trochlear motoneurones. Soc. Neurosci.
Abstr. 9:1088.
Spencer, K.F., and Baker, R. (1992) GABA and glycine as inhibitory
neurotransmitters in the vestibuloocular reflex. Ann. N.Y. Acad. Sci.
656:602-611.
Spencer, R.F., and Wang, S:F. ( 1996) Immunohistochemical localization of
neurotransmitters utilized by neurons in the rostra1 interstitial nucleus
of the medial longitudinal fasciculus (riMLF)t h a t project to the oculomo-
tor and trochlear nuclei in the cat. J . Comp. Neurol. (this issue).
Spencer, R.F., R.J. Wenthold, and R. Baker i 1989) Evidence for glycine as an
inhibitory neurotransmitter of vestibular, reticular, and prepositus
hypoglossi neurons that project to the cat abducens nucleus. J . Neurosci.
9:2718-2736.
162
Spencer. K.F.. S:F. Wang, and K. Baker (1992)The pathways and functions
of GABA in the oculomotor system. Progr. Brain Kes. 90.307-331.
Steiger, H.-J., and J.A. Buttner-Ennever (1979) Oculomotor nucleus affer-
ents in the monkey demonstrated with horseradish peroxidase. Brain
Kes. 160:1-15.
Thiimke. F., and H.C. Hopf i1992) Acquired monocular elevation paresis.
Brain 115:1901-1910.
Tomasulo. R.A.. J.J. Ramirez, and 0 . Steward (1993) Synaptic inhibition
regulates associative interactions between afferents during the induction
of long-term potentiation and depression. Proc. Natl. Acad. Sci. USA
90:11578-11582.
Tredici. G.. G. Pizzini, and S. Milanesi (1976) The ultrastructure of the
nucleus of the oculomotor nerve (somatic efferent portion) of the cat.
Anat. Embryol. 149.323-346.
Trojanowski, J.Q., and M.H. LaFontaine (1981)Neuroanatomical correlates
of selective downgaze paralysis. J . Neurol. Sci. 52191-101.
Uchino. Y., N. Hirai. and S. Watanabe (1978) Vestibulo-ocular reflex from
the posterior canal nerve to extraocular motoneurons in the cat. Exp.
Brain Kes. 32r377-388.
S.-F. WANG AND R.F. SPENCER
Uchizono, K. (1965)Characteristics of excitatory and inhibitory synapses in
the central nervous system. Nature (London) 207:642-643.
Vilis, T., K. Hepp, U. Schwarz, and V. Henn (1989) On the generation of
vertical and torsional rapid eye movements in the monkey. Exp. Brain
Kes. 77:l-11.
Walmsley, B., and K. Stuklis (1989) Effects of spatial and temporal disper-
sion of synaptic input on the time course of synaptic potentials. J.
Neurophysiol. 61:681-687.
Wang, S.-F., and K.F. Spencer (1995) Spatial organization of premotor
neurons related to vertical upward and downward saccadic eye move-
ments in the rostra1 interstitial nucleus of the medial longitudinal
fasciculus (riMLF)in the cat. J . Camp. Neurol. (thisissue).
Waxman, S.G., and G.D. Pappas (1979) Ultrastructure of synapses and
cellular relationships in the oculomotor nucleus of the rhesus monkey.
Cell Tiss. Res. 204:161-169.
White, G., W.B. Levy, and 0. Steward (1990) Spatial overlap between
populations of synapses determines the extent of their associative
interaction during the induction of long-term potentiation and depres-
sion. J. Neurophysiol. 6'4.1 186-1198.