Hearing Research, 17 (1985) 87-93 Elsevier 87 HRR 00570 Low frequency neurons in the lateral central nucleus of the cat inferior colliculus receive their input predominantly from the medial superior olive Lindsay Aitkin and Dianne Schuck Department of Physiology, Monash University, Clayton, 3I68, Australia (Received 12 October 1984; accepted 4 December 1984) When iontophoretic injections of horseradish peroxidase were made in the vicinity of inferior colliculus units in the cat responding to low sound frequencies, retrograde labelling occurred in the ipsilateral medial superior olive. Both bipolar and multipolar cells of the medial superior olive participated in this projection, and the focus of labelling shifted topographically within the olive as the injection site best frequency varied. These observations confirm previous anatomical findings and link them to the tonotopic organization of the central nucleus of the inferior colliculus. However, the fact that the medial superior olive alone provides between 50 and 98% of labelled cells in the brain stem projecting to this region of the central nucleus is an unexpected observation. This study gives further support to an hypothesis of ‘core zones’ within the central nucleus that receive preferentially input from specific brain stem auditory nuclei. low frequency, central nucleus of inferior colliculus, medial superior olive, core zone, physiology and tracing Introduction The inferior colliculus is a major nexus of the auditory pathway, being an obligatory relay for most, if not all, afferents originating from the cochlear nuclear and superior olivary complexes [5,9]. There is some evidence that afferents from these complexes terminate in differing, although overlapping, parts of the central nucleus of the inferior colliculus [6,13,15,16]. These subregions of the central nucleus appear to share, in the cat. a comrrKm tonotopic organization [12,18]. In a re- cent Golgi analysis of the central nucleus of the cat, Oliver and Morest [14] have argued for subdi- vision of the central nucleus on the basis of den- dritic morphology and orientation. Their schema is compatible with a concept of segregation of affer- ent terminations. It is known that low frequency sounds (< 3 kHz) are represented in lateral, rostra1 and dorsal parts of the central nucleus of the inferior collicu- lus of the cat [4,12,15,18]. It is also known that the medial superior olivary nucleus projects to this part of the inferior colliculus [6,7,10]. In a recent report from this laboratory, the sensitivities of low best frequency inferior collicular neurons of the cat were examined in relation to sound source location in anechoic space [4]. In the course of some of these experiments electrophysiological data were obtained using micropipettes containing a solution of horseradish peroxidase in physiologi- cal saline. Following study of low frequency units in the inferior colliculus, the retrograde tracer was deposited by iontophoresis in the vicinity of selected units in the central nucleus. The results of these combined recording and tracing experiments, described in this report, con- firm that the medial superior olive projects ipsi- laterally to regions of the inferior colliculus con- taining low best frequencies and suggest, addition- ally, that the medial superior olive provides the majority of afferents to this part of the inferior colliculus. Some of these observations have been published elsewhere in abstract form [2]. Methods This study required satisfaction of the following criteria: (1) the isolation and characterization of single units in the inferior colliculus, (2) the loca- 0378-5955/85/$03.30 0 1985 Elsevier Science Publishers B.V. (Biomedical Division)
Transcript
Hearing Research, 17 (1985) 87-93
Elsevier
87
HRR 00570
Low frequency neurons in the lateral central nucleus of the cat inferior colliculus receive their input predominantly from the medial superior olive
Lindsay Aitkin and Dianne Schuck Department of Physiology, Monash University, Clayton, 3I68, Australia
(Received 12 October 1984; accepted 4 December 1984)
When iontophoretic injections of horseradish peroxidase were made in the vicinity of inferior colliculus units in the cat responding
to low sound frequencies, retrograde labelling occurred in the ipsilateral medial superior olive. Both bipolar and multipolar cells of the
medial superior olive participated in this projection, and the focus of labelling shifted topographically within the olive as the injection
site best frequency varied.
These observations confirm previous anatomical findings and link them to the tonotopic organization of the central nucleus of the
inferior colliculus. However, the fact that the medial superior olive alone provides between 50 and 98% of labelled cells in the brain
stem projecting to this region of the central nucleus is an unexpected observation. This study gives further support to an hypothesis of
‘core zones’ within the central nucleus that receive preferentially input from specific brain stem auditory nuclei.
low frequency, central nucleus of inferior colliculus, medial superior olive, core zone, physiology and tracing
Introduction
The inferior colliculus is a major nexus of the auditory pathway, being an obligatory relay for most, if not all, afferents originating from the cochlear nuclear and superior olivary complexes
[5,9]. There is some evidence that afferents from these complexes terminate in differing, although overlapping, parts of the central nucleus of the inferior colliculus [6,13,15,16]. These subregions of
the central nucleus appear to share, in the cat. a comrrKm tonotopic organization [12,18]. In a re- cent Golgi analysis of the central nucleus of the cat, Oliver and Morest [14] have argued for subdi- vision of the central nucleus on the basis of den- dritic morphology and orientation. Their schema is compatible with a concept of segregation of affer- ent terminations.
It is known that low frequency sounds (< 3
kHz) are represented in lateral, rostra1 and dorsal parts of the central nucleus of the inferior collicu- lus of the cat [4,12,15,18]. It is also known that the medial superior olivary nucleus projects to this part of the inferior colliculus [6,7,10]. In a recent report from this laboratory, the sensitivities of low
best frequency inferior collicular neurons of the cat were examined in relation to sound source location in anechoic space [4]. In the course of some of these experiments electrophysiological
data were obtained using micropipettes containing a solution of horseradish peroxidase in physiologi- cal saline. Following study of low frequency units in the inferior colliculus, the retrograde tracer was
deposited by iontophoresis in the vicinity of selected units in the central nucleus.
The results of these combined recording and tracing experiments, described in this report, con- firm that the medial superior olive projects ipsi- laterally to regions of the inferior colliculus con- taining low best frequencies and suggest, addition- ally, that the medial superior olive provides the majority of afferents to this part of the inferior
colliculus. Some of these observations have been published elsewhere in abstract form [2].
Methods
This study required satisfaction of the following criteria: (1) the isolation and characterization of single units in the inferior colliculus, (2) the loca-
tion of the electrode in a region of low best frequency, (3) adequate iontophoresis of horsera- dish peroxidase (HRP) from the electrode, and (4) maintenance of the animal under anesthetic for at least 24 h to allow retrograde transport. These criteria were met in 4 out of 10 cats in which their realization was attempted. In the unsuccessful ex- periments, HRP was apparently not extruded from the electrode in two animals, two other cats died prematurely, and the tracer was deposited in high best frequency regions in the remaining two animals.
Cats were anesthetized with sodium pentobar- bital (40 mg/kg). Details of surgical techniques and the expe~mental ~rangement were given in a recent report from this laboratory j3] and need only be briefly outlined here. For each cat the skull was cemented with dental acrylic to a stain- less steel bar that was attached to a holding cradle containing a feedback electric blanket (38’C). The cat, suspended in the cradle, was relocated in an anechoic sound-attenuating room.
Pure tone bursts (200 ms duration, 10 ms rise and fall times, rate l/s) were presented from a Magnavox 3T speaker that could be moved around the cat’s head along a semicircular hoop [3}. In these experiments the hoop was maintained in the azimuthal (horizontal) plane, shared by the cat’s eyes and meatal openings. Movements of the speaker were made by a remotely controlled step- ping motor. For each unit, the speaker was located at the ~ontralateral azimuth at which strongest evoked activity occurred. The sound intensity was decreased and the sound frequency at which the threshold was lowest was determined (best fre- quency).
E~e~trades and ia~ta~hor~sis Fine mi~ropipettes were pulled from capillary
glass (Clark Electromedical GC 150F-15), blunted under a microscope to tip diameters of 5-10 pm and backfilled with HRP (Sigma Type VI) as a 10% solution in 0.9% sterile saline. This solution was used in order to bias the proportion of posi- tively charged ion species in favor of HRP yet to maintain electrode impedance at an adequately low level (lo-20 MQ).
The inferior colliculus was exposed by ablation of the overlying occipital cortex, and the micro-
pipette was positioned laterally to penetrate the inferior colliculus from dorsal to ventral. Unit recording was provided through a high input im- pedance FET recording stage; while background noise levels were higher than those associated with tungsten electrodes, spike isolation was usually excellent. Following physiological characterization of the units isolated, HRP iontophoresis was made along a portion of the recording track co~~taining units the best frequencies of which varied within less than 1 kHz, and lay below 2 kHz.
The enzyme was extruded by passing 100 ms wide square pulses of positive current (usually 1 PA) through the pipette at 5 Hz for 5 min. De- posits were placed every 200 pm along a track length of 0.6-1.4 mm, and 5-10 rnin were allowed to elapse between successive deposits in order that the tip of the micropipette be replenished with HRP after each injection.
The cats were maintained under pentobarbital anesthesia for approximately 24 h after ionto- phoresis and then perfused intracardially with 1.5 1 of warm heparinized 0.9% saline (38°C) followed by 1.5 1 of 0.1 M phosphate buffered (pH 7.2) 2% ~ut~aldehyde (4°C). A block of tissue, extending from the posterior thalamus to a plane just caudal to the cochlear nuclei, was fixed for a further 24-48 h in the cold buffered glutaraldehyde and transferred to cold 30% phosphate buffered sucrose until it sank.
Orientation marks were made on each block and sections were cut at 50 ,um on a Leitz freezing microtome at a plane parallel to the posterior faces of the inferior colliculi (i.e. about 1.5’ to the fron- tal plane). Alternate sections were reacted with tetr~ethylben~d~e using the cold reaction tech- nique of Lane ill]. These sections were not coun- terstained in order to avoid a loss in the sensitivity provided by the cold reaction method and thus preserve the maximum detail of labelled material. Adjacent sections were stained with cresyl violet or thionin as cytoarc~t~tural controls. The counts to be presented thus represent approximately half of the total number of projecting neurons.
Analysis of material Labelled neurons were of two types: aggrega-
89
tions of granules, surrounded by a faint membrane and containing a granule-free nucleus, or, more commonly, profiles in which all features were sharply delineated and in which granular material sometimes spread for a considerable distance along
the proximal dendrites (e.g. Fig. 3).
B 83-23
All lower auditory nuclei were scanned at high power for labelled cells including the nuclei of the
lateral lemniscus and the superior olivary and cochlear nuclear complexes of both sides. The locations of all labelled neurons were plotted onto outline drawings at 7X x using a Leitz micropro-
mar; drawings of individual cells were made at 300 x using this device and selected photographs taken at a similar magnification with a Leitz Or- tholux microscope.
C 83 - 17 0 83-12
Results
Electrode tracks, recording details and HRP deposit sites are shown in Fig. 1. The four brains
are arranged in order of mean best frequency at the injection site, from 1.4 kHz for 83-10 (Fig. 1A) to 0.5 kHz for 83-12 (Fig. 1D). The tracks in 1A and 1B are caudally-located relative to those in Figs. 1C and 1D; those in Figs. lA, B and D are located in the lateral one-third of the central nucleus, while that in Fig. 1C is located in the rostra1 tip of the central nucleus.
Fig. 1. Outline drawings of frontal sections through the centers of horseradish peroxidase injection sites in 4 cats (83-10, etc.). The trajectory of the electrode penetrations is given by the dotted lines, the centers of the injections by the black columns and the zone of coloration around the center is hatched. Numbers shown to the left of each outline are the best frequen- cies in kHz of single units. ICC, ICX, central and external nuclei of the inferior colliculus; LLD. dorsal nucleus of lateral lemniscus.
The best frequencies of all units at the injection sites were less than 1.7 kHz and varied in their free-field spatial properties [3,4]. Those studied in cats 83-10 (Fig. IA) and 83-17 (Fig. 1C) were all azimuth selective, each responding maximally when the loudspeaker was located at a particular azimuthal location. In contrast, the three units so studied in cat 83-12 (Fig. 1D) were omnidirec- tional, while a mixture of omnidirectional and azimuth selective units was found in cat 83-23 (Fig. 1B).
Horseradish peroxidase injection sites consisted of cylindrical columns of dense reaction product
1-2 mm in length and about 0.2-0.3 mm in diam- eter (dark zones in Fig. ‘l), surrounded by an elliptical zone of lighter Iabelling approximately 1 mm in diameter (hatched zones in Fig. 1).
brain stem nuclei and the ~ontralateral cochlear nuclear complex are shQwn for the four cats in
Fig. 2 in the order in which they are presented in Fig. 1. It is a striking fact that in all cats, irrespec- tive of injection site best frequency or absolute number of labelled cells, the ipsilateral medial
superior olive (MSO) contained the majority of labelled cells, varying from 50% in cat 83-10 to 98% in cat 83-12. Small contingents of well labelled cells were also observed in the dorsal nucleus of the lateral lemniscus, the contralateral dorsal cochlear nucleus (fusiform cells) and the lateral superior olive.
There were no gross differences between the distributions of labelled cells in experiments where
recorded units were azimuth selective or omnidirectional. However, the sample size is clearly too small to decide this question unequivocally.
Bruin stem locations of retrograde/y-labelied neurons CeN types and their location within MS0
The numbers of labelled cells observed in 7 Labelled neurons in MS0 the dendrites of which
Fig. 2. Histograms for 4 cats (83-10, etc.) in which the numbers of labelled cells found in each of 8 brain stem nuclei are shown. C, I. contralateral, ipsilateral; LLD, LLV, dorsal and ventral nuclei of lateral lemniscus; LSO, MSO, lateral and medial superior olives; CNC, co&ear nuclear complex. Center frequency at injection site shown in kHz, n = total number of labelled cells in each cat; 50% etc., percentage of total found in ipsilateral MSO.
were adequately filled with reaction product fell into two groups (Fig. 3). For one group, bipolar cells, two stem dendrites were directed at an ap- pro~mately 45 o angle to the intersection of dorsoventral and horizontal axes. The cell bodies of these neurons were elliptical in shape with soma diameters in the short axis of lo-15 pm and for the long axis 30-35 pm (Figs. 3A-C). For the second group, multipolar cells, four or more stem dendrites could be recognized, although only two are visible in the planes of focus of Figs. 3D-F. These dendrites arose from a roughly circular soma of diameter 18-20 pm. While the dendrites of bipolar cells in a given section tended to be aligned in a single plane, those of multipolar cells spread in all directions, including that orthogonal to the bipolar cell plane, and in some cases could be followed for up to 500 pm.
In the present experiments labelled neurons in MS0 clustered together in the frontal plane but were spread out along the rostrocaudal extent of the nucleus. The location in the frontal plane of
the labelled cell clusters depended on the best frequency at the injection site. The lowest center frequency (0.5 kHz, 83-12) provided labelling which was concentrated dorsally and medially in MS0 (Figs. 4E-H); the cell clusters following a 1.4 kHz injection were displaced ventrally and laterally (Figs. 4A, B) and, at caudal levels, were located in the ventrolateral half of MS0 (Figs. 4C, D). Labelled cells resulting from the two inter- mediate frequency injections (1.1 kHz, cat 83-17; 1.2 kHz, cat 83-23) lay between these distributions.
Discussion
We have shown that iontophoretic injections of horseradish peroxidase, made in the vicinity of low best frequency units in the central nucleus of the inferior colliculus of the cat, provide retrograde labelling to neurons in the ipsilateral MSO. The present findings thus form a bridge between the physiological studies detailing the tonotopic organization of the central nucleus [12,18], and
91
Fig. 3. Examples of labelled cells observed in ipsilateral MSO. Note the bipolar cells in A (lower), B and C: a triangular soma with two
dendrites emerging from the apex is seen in A (upper). Multipolar cells are shown in D-F; the dark hand in D. upper right. is an
artefact. Calibration applies to all frames.
anatomical observations about connections be- tween the MS0 and the inferior colliculus
[1,6,8.10]. These data also lend support to the concept of a
partial segregation of ascending afferent terminals
within the central nucleus, originally proposed by Roth and his colleagues [15] on the basis of micro- electrode mapping followed by relatively large horseradish peroxidase injections. Thus, we have shown not only that the ipsilateral MS0 projects to the low best frequency part of the inferior colliculus sampled in our study, but that it pro-
vides between 50 and 98% of the total population of brain stem neurons, labelled in these experi- ments, projecting to this part of the central nucleus. Some other brain stem nuclei containing low best
frequencies (e.g. the anteroventral cochlear nucleus) were not labelled in these experiments, although their projections to the inferior colliculus are well documented [1,15].
The strength of these conclusions hinges on the assumption that the tracer has labelled all cell bodies projecting terminals to the injection sites in these experiments. While it is difficult to prove
92
L I’
83-10 14 kHz
Y 83-12 \ fbr; LHz
// CAUDAL ” ” %’
Fig. 4. Outline drawings of the ipsilateral MS0 seen in frontal sections for two cats (83-10 and 83-12) with dorsal (DO) up and lateral (L) to the right. Each outline pools labelled cells located in 2-5 sections each separated by a Nissl-stained section; the distance in mm over which pooling has been made is given on each section; these distances are expressed from the rostra1 limit of MS0 (0 mm). Injection site center frequency shown under experiment number
this assumption, certain observations lend support to it. First, clear labelling was always observed in nuclei other than the MSO, and was found in 3 cats in the contralateral dorsal cochlear nucleus - in terms of retrograde transport distance, well beyond the ipsilateral MSO. Secondly, the pre- dominance of labelled cells in MS0 relative to other nuclei occurred irrespective of the absolute numbers of labelled cells - compare experiment 83-17 with experiments 83-10 and 83-12 (Fig. 2). Finally, the focus of labelling in MS0 shifted topographically with shifts in the best frequency at the injection site, indicating that the projections from MS0 revealed using this method were orderly and not random.
These findings suggest the possibility of ‘core zones’ in which terminals from specific brain stem nuclei are concentrated in the central nucleus. Given the lateral locations of three of the horseradish peroxidase injection sites in the pre- sent study, it is possible that the core zone for the MS0 is coextensive with the region termed ‘pars lateralis’ by Oliver and Morest [14]. Similar ideas of differential ter~nations for ascending afferents within the central nucleus have been presented for the tree shrew (DR. Jones, as cited in [14]) and the mouse [16].
Finally, we have shown that both bipolar and multipolar cells of MS0 project to the central nucleus of the inferior colhcuhrs of the cat, thus
confirming the observations of Adams [l] made on
the basis of large horseradish peroxidase deposits in the central nucleus. These two classes correlate welf with the bipolar and multipolar cells re- cognized by Schwartz [17] in Golgi material of the
MSO.
Acknowledgements
The authors would like to thank Dr. S.C. Phii- lips, who participated in some of the experiments. Dr. D.R.F. Irvine for his critical comments on an earlier draft of this paper, Lynne Hepburn, who typed the manuscript and Jill Poynton, illustrator.
This study was supported by a grant from the Australian Research Grants Scheme.
References
I Adams, J.C. (1979): Ascending projections to the inferior
coliiculus. J. Comp. Neurol. 183, 519-538.
Aitkin. L.M. (1983): The pathway for low frequency sound
localisation in the cat hindbrain. In: Mechanisms of Hear-
mg. p. 100. Editors: W.R. Webster and L.M. Aitkin. Monash
University Press. Clayton.
Aitkin. L.M., Gates, G.R. and Phillips, SC. (1984): Re-
sponses of neurons in inferior colliculus to variations in
