Laboratory of Sensory Neuroscience, Institute of Neuroscience, Carleton University, Ottawa, Ontario K1S 5B6, Canada
![]() |
ABSTRACT |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Wu, Shu Hui. Physiological properties of neurons in the ventral nucleus of the lateral lemniscus of the rat: intrinsic membrane properties and synaptic responses. The physiological properties including current-voltage relationships, firing patterns, and synaptic responses of the neurons in the ventral nucleus of the lateral lemniscus (VNLL) were studied in brain slices taken through the young rat's (17-37 days old) auditory brain stem. Intracellular recordings were made from VNLL neurons, and synaptic potentials were elicited by electrical stimulation of the lateral lemniscus ventral to the VNLL. Current-voltage relations and firing patterns were tested by recording the electrical potentials produced by intracellular injection of positive and negative currents. There were two types of VNLL neurons (type I and II) that exhibited different current-voltage relationships. In response to negative current, both type I and II neurons produced a graded hyperpolarization. Type I neurons responded to positive current with a graded depolarization and multiple action potentials the number of which was related to the strength of the current injected. The current-voltage relations of type I neurons were nearly linear. Type II neurons responded to positive current with a limited depolarization and only one or a few action potentials. The current-voltage relations of type II neurons were nonlinear near the resting potential. The membrane properties of the type II VNLL neurons may play an important role for processing information about time of onset of a sound. Type I neurons showed three different firing patterns, i.e., regular, onset-pause and adaptation, in response to small positive current. The onset-pause and adaptation patterns could become sustained when a large current was injected. The regular, onset-pause, and adaptation patterns in type I neurons and the onset pattern in type II neurons resemble "chopper," "pauser," "primary-like," and "on" responses, respectively, as defined in in vivo VNLL studies. The results suggest that different responses to acoustic stimulation could be attributed to intrinsic membrane properties of VNLL neurons. Many VNLL neurons responded to stimulation of the lateral lemniscus with excitatory or inhibitory responses or both. Excitatory and inhibitory responses showed interaction, and the output of the synaptic integration depended on the relative strength of excitatory and inhibitory responses. Neurons with an onset-pause firing pattern were more likely to receive mixed excitatory and inhibitory inputs from the lower auditory brain stem.
![]() |
INTRODUCTION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The ventral nucleus of the lateral lemniscus
(VNLL) is the most ventral neuronal group of the nuclei of the lateral
lemniscus. In contrast to the dorsal nucleus of the lateral lemniscus
(DNLL), which receives inputs from the brain stem bilaterally and is
concerned primarily with binaural processing, the VNLL receives inputs
mainly from the contralateral ventral cochlear nucleus (VCN) with
smaller projections arising from the ipsilateral VCN and superior
olivary complex (SOC) (Browner and Webster 1975;
Covey and Casseday 1986
; Friauf and Ostwald
1988
; Glendenning et al. 1981
; Helfert et
al. 1991
; Huffman and Covey 1995
;
Schofield and Cant 1997
; Schwartz 1992
;
Spangler et al. 1985
; Warr 1982
;
Zook and Casseday 1985
). Physiological responses of VNLL
neurons to acoustic stimulation and functional organization of the VNLL
have been studied in cat (Aitkin et al. 1970
), and
extensively in echo-locating bat (Covey and Casseday 1986
,
1991
; Metzner and Radtke-Schuller 1987
;
Vater et al. 1997
). VNLL neurons mainly are influenced
by sound presented to contralateral ear and respond to tone bursts with
different temporal discharge patterns. The VNLL mainly is concerned
with monaural processing and may have an important role in acoustic temporal discrimination (Aitkin et al. 1970
;
Covey and Casseday 1991
; Guinan et al.
1972a
,b
; Metzner and Radtke-Schuller 1987
).
In the bat, the nuclei of the lateral lemniscus including the VNLL are
greatly expanded and highly developed. The bat VNLL can be divided into
two subdivisions, the columnar area (VNLLc) and the multipolar area
(VNLLm), based on the cytoarchitecture, synaptic arrangement, afferent
tonotopic projections and physiological properties (Covey and
Casseday 1986, 1991
; Huffman and Covey 1995
; Vater et al. 1997
). The connections of the VNLLc are
organized in sheets that are precisely related to the tonotopic
organization of its afferents from the anterior ventral cochlear
nucleus (AVCN) and its efferents from the inferior colliculus
(Covey and Casseday 1986
). Cells in the VNLLc are very
similar to spherical bushy cells in the AVCN and are contacted with
large calyx-like synaptic terminals in addition to conventional bouton
terminals. They are broadly tuned with no spontaneous activity and
respond with one spike per stimulus and with constant latencies to
stimulus onset. The VNLLc is thought to be specialized to encode the
onset of a sound (Covey and Casseday 1991
). Cells in the
VNLLm are multipolar in shape. They respond to tone bursts with various
temporal patterns, which are distinguished by the shape of the single
unit's poststimulus time histogram, i.e., tonic (discharge at a
constant high rate through the duration of the tone burst), chopper
(distinct and regularly spaced peaks of discharge throughout the
duration of the stimulus), primary-like (discharge throughout the
stimulus duration, but firing rate diminished after an initial
transient on response), or pauser (initial discharge followed by a
silent period and then firing at a diminished rate) but without the
single-spike constant-latency response pattern seen in the VNLLc. The
VNLLm may play a role in encoding ongoing properties of a sound
(Covey and Casseday 1991
).
In a variety of mammalian species including bat, cat, rat, mouse,
gerbil, mole, ferret, guinea pig, and opossum, the VNLL sends efferent
projections to the ipsilateral inferior colliculus (IC) (Adams
1979; Brunso-Bechtold et al. 1981
;
Coleman and Clerici 1987
; Druga and Syka
1984
; Frisina et al. 1998
; Kudo
1981
; Kudo et al. 1990
; Majorossy and
Kiss 1994
; Moore 1988
; Nordeen et al. 1983
; Ross et al. 1988
; Schofield and
Cant 1997
; Schweizer 1981
; Strutz
1980
; Whitley and Henkel 1984
; Willard
and Martin 1983
; Willard and Ryugo 1983
;
Zook and Casseday 1987
). Immunocytochemical studies have
shown that most of VNLL neurons are glycinergic (Vater et al.
1997
); half of this projection is also
-aminobutyric acid (GABA) immunoreactive (Saint Marie et al. 1997
). After
injection of tritiated glycine into the IC many VNLL neurons are
retrogradely labeled (Saint Marie and Baker 1990
). After
injection of horseradish peroxidase (HRP) or Fluorogold into the IC,
many of the retrogradely labeled neurons in VNLL show positive
immunostaining with GABA antibody
(González-Hernández et al. 1996
;
Zhang et al. 1998
). All these results suggest that the
VNLL is a major source of inhibitory input to the IC and may play a
significant role in inhibitory processing there.
Physiological studies on VNLL neurons in terrestrial mammals
(nonecho-locating species) are few and the function of the VNLL is
still largely unknown (Adams 1997; Aitkin et al.
1970
; Batra and Fitzatrick 1997
; Guinan
et al. 1972a
,b
). A better understanding of the function of the
VNLL can be gained by examining the membrane properties of individual
neurons and synaptic transmission in the VNLL. Because there have been
no previous intracellular studies of VNLL neurons, I am presenting here
basic information concerning intrinsic membrane properties and
discharge patterns of VNLL neurons in a rat brain slice preparation.
Synaptic responses of VNLL neurons also have been investigated.
Preliminary results of these studies have been presented previously in
the form of abstracts (Wu 1996
, 1997
).
![]() |
METHODS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Brain slices were obtained from young albino rats (Wistar,
Charles River, Quebec) between 17 and 37 days of age in this study. The
animals first were anesthetized with halothane and then were killed by
decapitation. The whole brain was removed and submerged in a warm
(30°C), oxygenated saline solution. The auditory midbrain was blocked
by making cuts rostral and caudal to the lateral lemniscus and was
mounted on the stage of a tissue slicer using a cyanoacrylate adhesive.
Coronal sections were taken at 400-µm thickness through the lateral
lemniscus with a Vibratome tissue slicer. Usually three slices
containing the VNLL were obtained from each animal. A slice that
contained the largest part of VNLL was chosen and placed in a small
recording chamber identical to that described in previous publications
(Oertel 1983; Wu and Kelly 1991
). The other two slices were kept in an oxygenated saline solution at room
temperature for later use. An oxygenated saline solution circulated
continuously through the recording chamber and around the brain slice
at the rate of 10-12 ml/min. The tissue slice was submerged fully in
saline and was held in place between two pieces of nylon mesh. The
volume of the recording chamber was ~0.6 ml.
The temperature of the saline in the chamber was monitored and maintained at ~33-34°C. The saline solution consisted of (in mM) 129 NaCl, 3 KCl, 1.2 KH2PO4, 2.4 CaCl2, 1.3 MgSO4, 20 NaHCO3, 3 N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid (HEPES), and 10 glucose in a double-distilled water, saturated with 95% O2-5% CO2. The pH was 7.4 after complete saturation with the O2-CO2 gas.
The brain slice was illuminated from below by light passed through a darkfield condenser. The structures of the auditory brain stem were identified with a Leitz dissecting microscope. The outlines of the VNLL, intermediate nucleus of the lateral lemniscus (INLL), and DNLL were clearly visible (Fig. 1A). A single recording electrode was inserted into the VNLL, and a stimulating electrode was placed on the lateral lemniscus ventral to the VNLL. The placement of the stimulating electrode is indicated in a cresyl-violet-stained frontal section shown in Fig. 1A.
|
Intracellular recordings were made with glass microelectrodes filled
with 4 M potassium acetate. The electrode impedances were between 120 and 160 M. Once the electrode reached the surface of the tissue, it
was advanced in 1-µm steps through the brain slice with a Burleigh
piezoelectric driver (Inchworm). Application of an oscillating current
(buzzing) was used to facilitate intracellular penetration. An Axoprobe
1-A amplifier was employed for all recordings. Experimental data were
stored with a Nicholet Benchtop Waveform Acquisition System 400 and
plotted later for analysis. The statistical analysis was conducted on
Microsoft Excel, running one-way ANOVAs at a criterion level of
= 0.05 on applicable data. The changes in membrane potential elicited by
intracellular injection of positive and negative currents were obtained
by measuring between the resting membrane potential and the plateau
depolarization and between the resting potential and the peak
hyperpolarization, respectively. The measured values were used to plot
the curve of current-voltage relationship. The slope of the curve was
calculated from a range of the hyperpolarization that was approximately
linear. The average data are presented as means ± SD.
Synaptic responses were elicited by electrical stimulation of the lateral lemniscus ventral to the VNLL (see Fig. 1A). The bipolar stimulating electrode was constructed from paired tungsten wires with a tip separation of ~50 µm. Electrical stimuli were square waves, 100 µs in duration, obtained from a Grass S-8800 stimulator and stimulus isolator. The rate of stimulation was one stimulus per second during the initial search for synaptic responses. When neural responses were encountered, the stimulus strength was adjusted within a range from the threshold to the suprathreshold level to produce graded postsynaptic responses.
![]() |
RESULTS |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
Integrity of the lateral lemniscus
The ascending neural pathway activated by stimulation of the
lateral lemniscus ventral to the VNLL was confirmed anatomically in
several experiments by injecting a small amount of biocytin (Vector,
2% in 2 M potassium acetate) into the slice at a position along the
lateral lemniscus ventral to the VNLL. A period of 6-8 h was allowed
for transport to take place. The tissue then was fixed by 4%
paraformaldehyde and imbedded in agar for cutting into 80-µm
sections. Finally the sections were reacted with diaminobenzidine (DAB)
and counterstained with cresyl violet (Wu and Kelly
1995). As can be seen in Fig. 1, C and D,
an injection of biocytin into the lateral lemniscus ventral to the VNLL
resulted in labeling of large-diameter fibers and large specialized
axon terminals, the calyces of Held in the medial part of the VNLL
(Fig. 1C), small-diameter fibers, and bouton terminals
throughout the VNLL (Fig. 1D). Some of thin fibers gave rise
to collaterals which turned at a right angle and ran perpendicularly
with respect to the fibers of the main stream of the lateral lemniscus.
These results demonstrate that some afferent pathway of the lateral lemniscus to the VNLL remains intact enough to transport biocytin in
the frontal slice preparation.
Membrane properties and firing patterns
Membrane properties of the VNLL neurons were investigated by
recording intracellular potentials in response to injection of current
into the cell through the recording electrode. The intracellular recordings were accepted based on several criteria: a resting potential
of at least 50 mV or more negative (mean =
57.4 ± 4.3 mV, n = 54), narrow and full-size action potentials,
and stability of the recording over a period of 0.5-4 h.
Current-voltage relations were examined in 54 VNLL neurons from 43 animals. The relations between current strength and magnitude of
voltage change provided a measure of the input membrane resistance
(mean = 51.6 ± 16.8 M
, n = 54). The
current-voltage (I-V) curve and temporal firing pattern were
used to characterize and distinguish cell types in the VNLL.
TYPE I CELL.
Figure 2 shows examples of the effect of
injecting current into four cells with type I characteristics.
Injection of negative current for 60 ms led to a hyperpolarization of
the cell membrane that was roughly proportional to the amount of
current injected (Fig. 2A, 1-3). The I-V
curves of these cells were essentially linear over the 0.1- to
0.8-nA current range (Fig. 2B, 1-3). With continued
application of current, the hyperpolarization was followed by a
"sag" of the membrane potential toward the resting level. A
depolarization that elicited two to three action potentials immediately
followed the end of the hyperpolarization (Fig. 2A1, bottom). The sag and afterdepolarization were more apparent with long-duration current pulses. Figure 2A4 shows an example of
a cell that was injected with 100-ms intracellular current. The sag and
afterdepolarization were seen clearly when the cell was hyperpolarized
with
0.3- and
0.5-nA currents. The afterdepolarization lasted for
30-35 ms and was large enough for generation of a train of action
potentials (Fig. 2A4, bottom 2 traces). The sag
and afterdepolarization probably reflect activation and deactivation, respectively, of an inward hyperpolarization-activated current (Ih). Activation of Ih
was found in 48.8% (21/43) of type I cells.
|
|
|
|
|
TYPE II CELL.
The other type of neuron in the VNLL, the type II cell, had a different
current-voltage relation and firing pattern from that of the type I
cells. Type II cells responded to positive current with only one action
potential and exhibited a nonlinear current-voltage relation as shown
in Fig. 6. Negative current (0.2 and
0.4 nA) led to a graded hyperpolarization of the cell membrane (Fig.
6, D and E). Both sag and afterdepolarization
with action potentials were found in 53.8% (7/13) of the type II
cells. Injection of suprathreshold positive current elicited one action
potential at the onset of the injection (Fig. 6, A-C). Most
type II cells responded with one action potential (2-3 spikes in some
cells) as current strength was increased
1 nA. Following the
spike(s), the membrane potential of the cell remained near rest and
quickly returned to the resting potential after the end of the current pulse. Equal amounts of depolarizing and hyperpolarizing current injection had a different effect on the change in membrane potential. Negative current resulted in a larger voltage shift than positive current. The current-voltage relation was nonlinear (Fig.
6F). Neurons with type II membrane characteristics were
found in 23.2% (13/56) of the cells in the sample. The mean resting
potential of the type II cell was
58.2 ± 4.6 mV
(n = 13) and the input resistance was 42.5 ± 12.6 M
(n = 13), not significantly different from type I
cells (P > 0.05). The type I and II cells were not segregated into separate regions within the VNLL. They were found throughout the VNLL.
|
Synaptic physiology
The synaptic physiology of the VNLL was investigated by examining synaptic potentials of VNLL neurons in response to electrical stimulation of the lateral lemniscus ventral to the VNLL. Figure 7 shows several examples of synaptic responses recorded from four VNLL neurons to stimulation of the lateral lemniscus. In Fig. 7A, the neuron responded to near-threshold stimulation with a very small excitatory postsynaptic potential (EPSP) that was followed by a prominent inhibitory postsynaptic potential (IPSP; bottom). At a slightly higher stimulus level, both EPSP and IPSP became larger (middle), and a single action potential could be evoked with sufficient stimulus strength (top). A similar response pattern with more predominant excitation was observed in some neurons. An example is shown in Fig. 7C. For this neuron, near-threshold stimulation elicited a small IPSP, and higher level of stimulation always evoked a suprathreshold response. In some neurons no suprathreshold response was observed although both EPSP and IPSP could be evoked. An example is shown in Fig. 7B. For this neuron, a lower level of stimulation (12.2 and 14 V) elicited EPSPs only. A higher stimulus level (16 V) evoked either EPSP or IPSP from trial to trial but did not elicit a suprathreshold response (action potential). In other cases, the threshold for the IPSP was lower than that for the EPSP, and inhibition dominated the synaptic response, as illustrated in Fig. 7D. This neuron responded to a low stimulus level (7 V) with an IPSP only. When the stimulus was increased to 19 V, a small EPSP was evoked. But the response could become an IPSP again at the same (19 V) or a much higher stimulus level (40 V). No suprathreshold response could be elicited at any stimulus level tested in this neuron.
|
In summary, many VNLL neurons, 35/43 (81.4%) of the type I and 11/13
(84.6%) of the type II neurons, responded to electrical stimulation of
the lateral lemniscus with EPSPs or IPSPs or both. The latencies of
most EPSPs and IPSPs were short (mean = 1.17 ms). Some type II
cells had EPSPs with shorter latency (0.6-0.8 ms), larger amplitude
and shorter duration, which resembled the characteristics of EPSPs
observed in bushy and octopus cells in the cochlear nucleus
(Golding et al. 1995; Oertel 1983
). Of 43 type I cells, I observed only EPSPs in 7 neurons (16.3%), only IPSPs
in 14 neurons (32.6%), and both EPSPs and IPSPs in 14 neurons (32.6%). Of 13 type II cells, I saw only EPSPs in 5 neurons (38.5%), only IPSPs in 2 neurons (15.4%), and both EPSPs and IPSPs in 4 neurons
(30.8%). The distribution of the EPSPs and IPSPs in type I and II VNLL
neurons is shown in Fig. 8.
|
A further analysis of the distribution of EPSPs and IPSPs among the three groups of neurons with regular, onset-pause, and adaptation firing patterns is shown in Fig. 9. Synaptic responses were observed in neurons of all three groups, but the responses were most common in the onset-pause (90.9%) compared with the adaptation (71.4%) and regular (85%) groups. Many more onset-pause neurons (63.6%) responded to stimulation of the lateral lemniscus with combined EPSPs and IPSPs, compared with the adaptation (14.3%) and regular (30%) groups.
|
![]() |
DISCUSSION |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|
The purpose of this study was to obtain basic information about the membrane characteristics and synaptic responses of VNLL neurons. I investigated the intrinsic membrane properties of VNLL neurons in response to intracellularly injected current and synaptic potentials in response to electrical stimulation of the lateral lemniscus. There were two types of VNLL neurons, type I and II, which exhibited different current-voltage relations. Three different firing patterns were generated in type I neurons. Excitatory and inhibitory synaptic potentials could be evoked by stimulation of the lateral lemniscus fibers.
Correlation of current-voltage relation with cell morphology
VNLL neurons sampled in this study showed two different responses
to intracellular injection of current. I named these type I and II
cells on the basis of their membrane properties. In type I cells,
injection of positive current produced a proportional depolarization of
the cell membrane and generated a sustained discharge of action
potentials, the number of which reflected the magnitude of the
depolarization. Both regular and irregular firing patterns were found
depending on the neuron from which recordings were made. For type I
cells, intracellular injection of positive or negative current led to
proportional depolarization or hyperpolarization of the cell membrane.
The magnitude of the initial depolarization and hyperpolarization was
an approximately linear function of current strength. In contrast, type
II cells responded to positive current injection with a limited
depolarization and only one or a few action potentials at the onset of
the current injection regardless of the intensity of the positive
current. These cells responded to negative current injection with
graded hyperpolarization. The current-voltage relation of the type II cells was nonlinear near the resting potential. The distinction between
type I and II cells in the VNLL closely resembles that previously
reported between stellate and bushy cells in the mouse VCN (Wu
and Oertel 1984) and that between cells in the lateral superior
olive (LSO) and medial nucleus of the trapezoid body (MNTB) (Wu
and Kelly 1991
). The nonlinear current-voltage relation of the
type II cells in the VNLL is also very much like that of the octopus
cells in the posteroventral cochlear nucleus (PVCN) (Golding et
al. 1995
) and principal cells in the medial superior olive
(MSO) (Smith 1995
).
The membrane properties of type II cells in the VNLL, bushy cells in
the VCN, octopus cells in the PVCN, and principal cells in the MNTB and
MSO are different from those of other types of neurons but similar to
each other. These cells all have nonlinear current-voltage
relationships. They respond to injection of positive current with only
one or a few action potentials. For VNLL type II neurons, the input
membrane resistance measured from the amplitude of voltage changes with
injected hyperpolarizing current was 42.5 ± 12.6 M (this
study), which is similar to that for bushy cells (30-50 M
)
(Oertel 1991
). However, when the cell is excited, the cell input resistance becomes lower. The consequence of the reduced input resistance is a limitation in generation of action potentials and
a faster repolarization in response to excitatory inputs. These
properties are very important for precisely encoding auditory temporal
information (Oertel 1991
, 1997
). The population of type II neurons in the rat VNLL may also precisely signal the time of onset
of a sound, like VNLLc neurons in the big brown bat (Covey 1993
). Voltage-clamp investigations have demonstrated that a
low-threshold (about
70 mV) K+ conductance is responsible
for the highly rectifying current-voltage relationship of bushy cells
in the VCN, and principal cells in the MNTB (Brew and Forsythe
1995
; Forsythe and Barnes-Davies 1993
; Manis and Marx 1991
). Whether or not type II VNLL
neurons have K+ channels similar to those in bushy cells in
VCN and principal cells in the MNTB requires further investigation.
In the VCN and SOC, the cell types based on membrane properties are
associated closely with the cell types for cytoarchitecture. Intracellular labeling studies have shown that cells with linear current-voltage relationships correspond to stellate cells in VCN and
principal cells in LSO as they appear in Golgi-impregnated material
(Wu and Fu 1998; Wu and Oertel 1984
).
These cells typically have multiple long dendritic branches extending
away from all sides of the neuron. In contrast, neurons with nonlinear
current-voltage relationship correspond to bushy cells in VCN
(Wu and Oertel 1984
) and principal cells in the MNTB
(Banks and Smith 1992
; Wu and Kelly
1991
). Bushy cells have fewer and shorter dendritic branches than stellate cells. Principal cells in the MNTB, like bushy cells, have compact or tufted dendrites. In this study, I did not attempt to
label the neurons from which the recordings were made. Nevertheless, the previous anatomic studies have shown that there are various morphological types of cells in the VNLL (Adams 1979
;
Covey and Casseday 1991
; Schofield and Cant
1997
; Willard and Ryugo 1983
; Zook and
Casseday 1982
). The morphology of oval or globular cells in
cat, guinea pig, and mouse VNLL and bat VNLLc is very much like that of
bushy cells in the VCN or principal cells in the MNTB (Adams
1979
; Covey and Casseday 1991
; Helfert
and Aschoff 1997
), whereas the anatomic features of other
types, mostly multipolar cells, in the VNLL are similar to stellate
cells in the VCN and principal cells in the LSO.
Another similarity of some neurons in the VNLL, bushy cells in the VCN,
and principal cells in the MNTB is that these neurons receive large
terminals from afferent fibers. Bushy cells are covered by end bulbs of
Held that arise from auditory nerve fibers (Lenn and Reese
1966; Lorente de Nó 1976
, 1981
;
Ryugo and Sento 1991
; Schwartz and Gulley
1978
). Principal cells of the MNTB are covered by calyces of
Held that come from the axons of the globular bushy cells in the
contralateral VCN (Friauf and Ostwald 1988
; Morest 1968a
,b
). Some VNLL neurons in cat and mouse and
VNLLc neurons in bat also receive large calyceal endings that resemble the end bulbs of Held in the VCN and MNTB (Adams 1979
,
1983
; Covey 1993
; Covey and Casseday
1986
; Vater and Feng 1990
; Willard and Ryugo 1983
; Zook and Casseday 1985
). In this
study, biocytin injection into the lateral lemniscus labeled thick
axons coursing along the fibers of the lateral lemniscus and
terminating with calyceal endings in the VNLL. Those morphological
similarities suggest that neurons that receive large synaptic terminals
in the VCN, MNTB, and VNLL may have common physiological properties.
Bushy cells in VCN and principal cells in MNTB, indeed, show very
similar intrinsic membrane properties. As shown in this study, type II neurons exhibited membrane properties, viz., nonlinear current-voltage relations, which resembled those for VCN bushy cells and MNTB principal
cells. One might expect that type II VNLL neurons would have bushy-like
morphological features. But further study using intracellular labeling
combined with physiological recording methods is required to test this hypothesis.
Firing characteristics
Type I VNLL neurons responded to depolarizing current injection
with sustained discharges. At low levels of current injection, three
different firing patterns, i.e., regular, onset-pause, and adaptation,
could be discerned. Type II VNLL neurons responded to positive current
injection with only one, or a few spikes at the onset of the injection.
These different firing patterns may correspond to different
poststimulus time histograms (PSTHs) recorded from VNLL neurons in vivo
in response to acoustic stimulation. For example, in the bat, neurons
in two distinct divisions, VNLLc and VNLLm, respond to tone or noise
bursts differently (Covey and Casseday 1991). In the
VNLLc area the majority of neurons respond to a tone burst with one
spike at short and constant latencies. This "onset" type of
neuron also is present in the VNLL of other mammals but appears to be
intermingled with other cell types (Vater et al. 1997
).
The response pattern of type II VNLL cells in this study resembles the
phasic type in the VNLLc of big brown bat (Covey and Casseday
1991
) and the VNLL of horseshoe bat (Metzner and
Radtke-Schuller 1987
), and the "onset" cell type in cat
VNLL (Adams 1997
; Guinan et al. 1972a
).
The phasic or onset response observed in in vivo studies could be
attributed to the type II intrinsic membrane characteristics. Type II
VNLL neurons in the rat, perhaps like VNLLc neurons in the bat
(Casseday and Covey 1995
), probably transmit information
about stimulus onset.
Previous physiological studies in vivo have shown that many VNLL
neurons respond to sounds with sustained discharges. In the early
physiological studies of cat VNLL, most neurons were found to have
sustained firing patterns, some with a silent period during the firing
(Aitkin et al. 1970), and some with primary-like or chopper patterns (Guinan et al. 1972a
). In a later
study, neurons in the VNLLm of the big brown bat were found to respond
to short tone bursts with four sustained firing patterns, i.e.,
chopper, tonic, primary-like, or pauser, which were classified
according to criteria established by Pfeiffer (1966)
(Covey and Casseday 1991
). Similar discharge types,
i.e., tonic and phasic-tonic patterns, were found in VNLL neurons in
the horseshoe bat (Metzner and Radtke-Schuller 1987
).
Neurons with regular, onset-pause, or adaptation firing patterns
demonstrated in the present study possibly correspond to cells with
chopper, pauser, and primary-like types, respectively. Although
synaptic events must drive VNLL neurons to generate these firing
patterns in vivo, the intrinsic properties of the cell membrane also
could contribute to the distinct discharge patterns. The firing
patterns of the VNLL neurons shown in the present study also depend on
the resting level of the cell membrane. For example, an onset-pause
type could be converted to a sustained type by holding the cell
membrane potential at a more depolarized level than the resting
potential. Similar results were obtained from an in vitro intracellular
study of the dorsal cochlear nucleus (DCN) (Manis 1990
).
In vivo studies of the DCN also have shown that response patterns can
be changed by hyperpolarizing the cell membrane in some DCN neurons
(Rhode and Smith 1986
; Rhode et al. 1983
). All these results support the idea that intrinsic
membrane properties and the resting membrane conductance are involved
in generation of different firing characteristics in auditory neurons (Manis 1990
) including VNLL neurons (this study).
Synaptic inputs
In this study stimulation of the lateral lemniscus elicited EPSPs
or IPSPs or both in many VNLL neurons. The lateral lemniscus is the
main route of ascending fibers to the VNLL from the trapezoid body and
lower brain stem (Glendenning et al. 1981;
Helfert and Aschoff 1997
). In this study many bouton
terminals of thin fibers are labeled after injection of biocytin into
the lateral lemniscus ventral to the VNLL. Some thicker fibers that may
originate from the octopus cells in the PVCN (Schofield and Cant
1997
) enter the ventral VNLL, run obliquely within the VNLL and
terminate with the calyces of Held in the ventral or middle region of
the VNLL (observation from this study). Therefore stimulation of the lateral lemniscus immediately ventral to the VNLL in the brain slice
probably activates many afferent inputs that make synapses onto VNLL neurons.
Which pathways contribute to these responses in the VNLL? Most of the
afferents to the VNLL originate in the contralateral VCN, with a very
small projection coming from the ipsilateral VCN (Adams and Warr
1976; Browner and Webster 1975
; Covey and Casseday 1986
; Friauf and Ostwald 1988
;
Glendenning et al. 1981
; Schofield and Cant
1997
; Suneja et al. 1995b
; Warr 1972
,
1982
; Zook and Casseday 1985
). Cells identified
as bushy, multipolar, and octopus cells in the VCN are the major
sources of the afferents to the VNLL (Friauf and Ostwald
1988
; Schofield and Cant 1997
; Thompson
1998
; Warr 1972
, 1982
). Large terminals, the
calyces of Held, found in the VNLL very likely originate from thick
axons of octopus cells in the PVCN (Adams 1997
;
Schofield and Cant 1997
; Thompson 1998
;
Vater and Feng 1990
). Calyceal endings are known to be
excitatory; neurons that receive calyceal terminals, such as bushy
cells in the VCN, principal cells in the MNTB and globular cells in the
VNLL, are excited by acoustical stimulation (Adams 1997
;
Rhode et al. 1983
; Smith et al. 1998
). A
recent electron microscopic (EM) study about synaptic organization of
VNLLc neurons in the big brown bat suggests that VNLLc neurons receive
excitatory inputs that may arise from large calyces derived from
neurons, probably octopus cells or large multipolar neurons, in the
PVCN (Vater et al. 1997
). I suggest that some excitatory
synaptic responses with shorter latency, larger amplitude, and shorter
duration observed in type II cells may represent excitatory inputs from
the octopus cells in the PVCN. But further study definitely is required
to clarify the characteristics of excitatory synaptic potentials of
type I and II cells in rat VNLL with pharmacological manipulation to
eliminate the IPSPs that can obscure the EPSPs in response to
stimulation of the lateral lemniscus.
In contrast to the calyceal endings from thick axons of octopus cells,
the medium-sized bead-like boutons probably originate from collaterals
of thinner axons derived from bushy and/or stellate (multipolar) cells
in the AVCN and PVCN, which pass through the VNLL toward the IC
(Covey 1993; Friauf and Ostwald 1988
;
Iwahori 1986
; Schofield and Cant 1997
;
Schwartz 1992
). The EM study about projections to the IC
from the AVCN has shown that axonal endings of CN neurons have small,
round synaptic vesicles and make asymmetric synaptic contacts on IC
neurons (Oliver 1987
). The results suggest that synaptic
inputs from the AVCN to the IC are excitatory. Because VNLL receives
axon collaterals of AVCN neurons the main axons of which ultimately
terminate in the IC, it is very likely that the AVCN provides
excitatory inputs to the VNLL as well.
Neurochemical studies further support the idea that projections from
the VCN to the VNLL are excitatory, probably glutamatergic. Suneja et al. (1995a) reported that the VNLL manifested
high-affinity uptake and release of D-[3H]
aspartate, which suggested the presence of synaptic endings that may
use glutamate or aspartate as an excitatory neurotransmitter. Furthermore ablation of the cochlear nucleus resulted in depression of
D-[3H] aspartate release in the VNLL,
indicating that glutamate or aspartate may be a transmitter for the
CN-VNLL synapses (Suneja et al. 1995b
).
The VNLL also receives minor inputs from the ipsilateral MNTB and
periolivary nuclei (lateral and ventral nuclei of the trapezoid body,
and ventral periolivary nucleus) (Elverland 1978;
Glendenning et al. 1981
; Huffman and Covey
1995
; Spangler et al. 1985
; Vater and
Feng 1990
; Warr and Beck 1996
). The projection
from the MNTB to the VNLL comes from axon collaterals of MNTB principal
neurons that give rise to efferents to the LSO. MNTB principal neurons have been shown to be immunoreactive for glycine (Helfert et al. 1989
). The projection from the MNTB to the LSO is inhibitory
and glycinergic (Moore and Caspary 1983
; Wu and
Kelly 1991
). The projection from axon collaterals of MNTB
neurons to the VNLL is probably inhibitory and glycinergic as well.
Immunocytochemical studies have shown that neurons in the periolivary
region, especially in the VNTB and LNTB, are GABAergic (Adams
and Mugnaini 1990
; González-Hernández et al.
1996
; Helfert et al. 1989
; Moore and Moore 1987
; Roberts and Ribak 1987
; Vater
et al. 1992
; Winer et al. 1995
). These neurons
are another possible source of inhibition to the VNLL. The presence of
glycine- and GABA-immunoreactive puncta and flattened synaptic vesicles
associated with inhibitory synapses in the VNLLc of the big brown bat
(Vater et al. 1997
), and the existence of glycine- and
GABA-immunoreactive perisomatic puncta in the VNLL of cat (Saint
Marie et al. 1997
) further support the concept that VNLL
neurons receive both glycinergic and GABAergic inputs.
Intracellular labeling of VNLL neurons reveals that axons of some VNLL
neurons give rise to collaterals some of which terminate within the
VNLL (Zhao and Wu 1998). Certainly these VNLL neurons can act as interneurons and may exert inhibitory influence on other
VNLL neurons. Therefore the IPSPs elicited by electrical stimulation of
the lateral lemniscus may originate from VNLL interneurons that also
were activated by electrical stimulation of the lateral lemniscus at
the same time that the recording was made. But the short-latency IPSPs
observed in some cases seem not to originate from the interneurons. In
addition, preliminary pharmacological data showed that the IPSPs
recorded from VNLL neurons still could be seen after blockade of the
EPSPs (Wu 1997
), indicating that in these neurons the
IPSPs were elicited by the afferent fibers of the lateral lemniscus
directly rather than elicited by the lateral lemniscus through
interneurons within the VNLL.
In the present study there were similar proportions of type I and II cells (81.4 and 84.6%) that responded to stimulation of the lateral lemniscus with either EPSP or IPSP or both. The results suggest that many type I and type II VNLL neurons may receive and integrate excitatory and inhibitory afferent inputs from the auditory lower brain stem. Although the response patterns seen in this study are not necessarily the patterns produced in vivo by acoustical stimulation, the results demonstrate the possible synaptic interaction in one neuron. The output of synaptic integration from one neuron can depend on relative strengths of excitatory and inhibitory inputs that impinge on it (Fig. 7).
In the type I neuron group, many onset-pause, regular, and adaptation neurons responded to ascending inputs of the lateral lemniscus with either a single type of synaptic response or a combination of EPSPs and IPSPs. There is so far no information available about what kind of synaptic inputs that project to the neurons with different firing patterns. The incidence of both excitatory and inhibitory synaptic responses recorded from the same neurons in the onset-pause group (63.6%) was much higher than in the adaptation (14.3%) and regular (30.0%) groups. This result suggests that neurons with an onset-pause firing pattern are more likely to receive mixed excitatory and inhibitory inputs from the lower brain stem.
In summary, the present study shows that there are different types of neurons in the VNLL in terms of intrinsic membrane properties and synaptic physiology. The results suggest that the VNLL is an important integrative nucleus in the auditory brain stem rather than a simple relay station connecting the lower brain stem with higher centers and that it plays multiple roles in auditory processing.
![]() |
ACKNOWLEDGMENTS |
---|
I thank Dr. J. B. Kelly for a critical reading of the manuscript and many helpful comments. I also thank B. van Adel for making Fig. 1.
This research was supported by the Natural Sciences and Engineering Research Council of Canada.
![]() |
FOOTNOTES |
---|
Address for reprint requests: S. H. Wu, Life Sciences Research Bldg., Institute of Neuroscience, Carleton University, 1125 Colonel By Dr., Ottawa, Ontario K1S 5B6, Canada.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Received 9 October 1998; accepted in final form 1 March 1999.
![]() |
REFERENCES |
---|
![]() ![]() ![]() ![]() ![]() ![]() ![]() |
---|