THE JOURNAL OF COMPARATivE NEUROLOGY 304103-122 (1991)

Topgraphyof Projections From the Auditory Cortex to the Inferior Colliculus in the Rat HORST HERBERT, ANDREAS ASCHOFF, AND JOACHIM OSTWALD Department of Animal Physiology, University of Tubingen, D-7400 Tubingen, Federal Republic of Germany

ABSTRACT We examined the organization of descending projections from auditory and adjacent cortical areas to the inferior colliculus (IC) in the rat by using the retrograde and anterograde transport of wheat germ agglutinin-horseradish peroxidase. Small tracer injections were placed into cytologically defined subnuclei of the IC. On the basis of the resulting pattern of retrogradely labeled neurons in the cortex, different cortical areas and fields were defined. Two secondary areas located ventrocaudally (Te2) and ventrally (Te3) to the primary auditory area (Tell were delineated. The primary auditory area was subdivided into a posterior (Tel.p), a medial (Tel.m), and an anterior (Te1.a) auditory field. In addition, we outlined an area located rostrally to the auditory areas comprising a part of the secondary somatosensory cortex, as well as a dorsal belt surrounding dorsally the auditory areas. The following basic patterns of corticocollicular projections are revealed: 1)layers 2 and 3 of the dorsal cortex of the IC (DC2, DC3) are differentially innervated by the primary auditory fields (Te1.p and Te1.a project bilaterally to DC2, while Te1.m projects bilaterally and in topographical order to DC3); cells in Tel.m, arranged in caudal to rostra1 sequence, project to corresponding loci in DC3 arranged from dorsolateral to ventromedial; 2) the fibrocellular capsule of the IC, comprising layer 1 of the dorsal and external cortex of the IC, receives input from the secondary auditory area Te2; 3) layers 2 and 3 of the external cortex of the IC are only weakly innervated by the primary and secondary auditory cortex; 4) the intercollicular zone receives its major input from the secondary auditory area Te3, the secondary somatosensory cortex, and the dorsal belt; and 5 ) finally, the central nucleus of the IC receives no input from the temporal cortex at all. Our results demonstrate that the corticocollicular projections are highly organized. These pathways may modulate auditory processing in different functional circuits of the inferior colliculus. Key words: descending auditory pathway, corticocollicular, cochleotopic, tonotopic, WGA-HRP

Even before the turn of the century, von Gudden, Flechsig, and von Monakow had reported the existence of fibers directly connecting the cortex with the midbrain. In 1900, Thompson reviewed their findings and documented his own research, reporting a projection from the temporal cortex to the posterior tubercle of the corpora quadrigemina in the primate. Since then, numerous investigators have confirmed the presence of such a pathway in a variety of mammalian species, employing degeneration techniques as well as retrograde and anterograde transport of fluorescent dyes, horseradish peroxidase, or tritiated amino acids (primate: Mettler, '35; cat: Massopust and Ordy, '62; Kusama et al., '66; van Noort, '69; Kelly and Wong, '81; Morest and Q

1991 WILEY-LISS, INC.

Oliver, '84; tree shrew: Casseday et al., '79; marsupial: Martin and Megirian, '72; Martin et al., '75; Willard and Martin, '84; mouse: Willard and Ryugo, '83; and rat: Krieg, '47; Beyerl, '78; Syka et al., '80; Druga and Syka, '84a,b; Land et al., '84; LeDoux et al., '85; Coleman and Clerici, '87; Druga et al.,'88; Games and Winer, '88; Roger and Amault, '89). These studies describe auditory cortical axons terminating in the shell nuclei of the IC and in the dorsomedial portion of the central nucleus. Interestingly, auditory cortical projections into the ventrolateral portion of the central nucleus of the IC were only reported in primates (Kuypers Accepted October 4,1990.

H. HERBERT ET AL.

104 and Lawrence, '67; Forbes and Moskowitz, '74; Fitzpatrick and Imig, '78). Despite the large number of anatomical studies, only a few investigators have demonstrated a differential pattern of projections from primary and secondary auditory areas to different inferior collicular subnuclei (cat: Diamond et al., '69; Rockel and Jones, '73a,b; Andersen et al., '80; tree shrew: Oliver and Hall, '78; rat: Faye-Lund, '85). In addition to input from the auditory cortex, the inferior colliculus is innervated by nonauditory cortical areas such as somatosensory or multisensory areas (primate: Kuypers and Lawrence, '67; cat: Cooper and Young, '76; marsupial: Martin and Megirian, '72; Martin et al., '75; Robards et al., '76; Robards, '79). Physiological studies have confirmed a pathway from the auditory cortex to the inferior colliculus. Furthermore, they demonstrate inhibitory, excitatory, and facilitatory influences of this descending pathway on acoustically driven collicular neurons (cat: Massopust and Ordy, '62; Watanabe et al., '66; Amato et al., '70; Mitani et al., '83; bat: Sun et al., '89; rat: Birt et al., '78; Syka and Popelar, '84; Syka et al., '88). The aim of our investigation was to examine the auditory corticocollicullar pathway in the rat in detail. By using the retrograde and anterograde WGA-HRP tracing technique, we sought to analyze this connection in both directions: retrogradely from the inferior colliculus; and anterogradely from the auditory cortex. With small, restricted injections of WGA-HRP into the different subnuclei of the inferior colliculus and into the auditory cortex, we could reveal a differential innervation pattern of primary andlor secondary auditory cortical neurons with respect to different inferior collicular subnuclei. A preliminary report of this study has been published (Herbert, '84).

MATERIALSANDMETHODS Experimentalanimalsandinjectionprocedure The experiments were performed in female SpragueDawley rats ranging in weight from 180 to 240 g. The animals were premedicated with atropine sulfate (0.8 mgkg s.c.) and anesthetized with pentobarbital (50 mgkg i.p.1. The rats were fixed in a stereotaxic device for tracer injections. The skull was exposed and a hole was drilled into the parietal or interparietal bone. Injections of the 1C and the AC were guided by stereotaxic coordinates from Paxinos and Watson ('86). Injections of the AC were made tangentially to the cortical surface to avoid damage of the

temporal areas. A 5% solution of WGA-HRP (Sigma) in physiological saline was pressure-injected using calibrated glass micropipettes (Drummond) bevelled to tip diameters of 20-30 pm. One to 3 nl of the tracer were injected into the IC and 2-8 nl into the AC over a period of 5 minutes. To minimize local diffusion and leakage of tracer into the tract, the pipette was left in place for several minutes before and after the injection (Mesulam, '82).

Tissue preparation After a survival time of 25-40 hours the animals were reanesthetized and perfused through the aorta with 50 ml of 0.9% saline at room temperature, followed by 500 ml of cold fixative containing 1.25% glutaraldehyde and 1% paraformaldehyde in 0.1 M phosphate buffer (PB) at pH 7.4 and followed by 300 ml of cold 10% sucrose in PB. Brains were soaked overnight at 4°C in 30% sucrose in PI3 for cryoprotection. Brains were cut on a freezing microtome in the coronal plane at 35 pm. Care was taken to cut all brains at the same angle. Sections were divided into four series. Two series were processed with tetramethylbenzidine (TMB; Mesulam, '821, mounted on gelatin-coated glass slides, and air dried. One TMB series was counterstained with neutral red, quickly dehydrated in a graded series of alcohols, and coverslipped with Entellan mounting medium. The second TMB series was coverslipped without counterstaining. The third series of sections was stained with thionin. The fourth series was discarded.

Evaluation of data The data of the retrograde tracing experiments are based on 38 cases. The locations of injection sites in the IC were determined with the aid of a camera lucida in the neutralred-counterstained sections. The locations of retrogradely labeled cortical neurons were mapped in two of four series with the aid of an X-Y plotter coupled to the microscope stage. From these plots a lateral view of the temporal cortical surface was reconstructed (see below). For the anterograde tracing experiments, the data from 24 animals were taken into account. Injection sites in the AC were plotted using the X-Y-plotter and then reconstructed. The anterograde axonal labeling in the IC was drawn with a camera lucida under polarized light. Fine punctate labeling in the neuropil was regarded as terminal labeling. Cytoarchitectural boundaries in the IC were defined by superimposing the drawings with the adjacent

Abbreviations

AAF AC AChE CG CnF contra cyox dB DLL EPSP IC ipsi IPSP MGB NADPH Par2 rhf

anterior auditory field auditory cortex acetylcholinesterase central gray matter cuneiform nucleus contralateral cytochrome oxidase dorsal belt dorsal nucleus of the lateral lemniscus excitatory postsynaptic potential inferior colliculus ipsilateral inhibitory postsynaptic potential medial geniculate body p-nicotinamide adenine dinucleotide phosphate secondary somatosensory cortex rhinal fissure

Sag

sc

2-DG WGA-HRP com CN DC1,2,3 EC1,2,3 12 Tel Te1.p Te1.m Te1.a Te2 Te3

sagulum nucleus superior colliculus 2-deoxyglucose wheat germ agglutinin-horseradish peroxidase Inferior Collicular Subnuclei commissure of the IC central nucleus dorsal cortex, layers 1 , 2 , 3 external cortex, layers 1 , 2 , 3 intercollicular zone Auditory Cortical Subdivisions primary auditory cortex posterior field of T e l medial field of T e l anterior field of T e l secondary auditory cortex, caudal area secondary auditory cortex, rostral area

AUDITORY CORTICOCOLLICULAR PROJECTIONS thionin-stained sections and using brain outlines and blood vessels as references.

Computer-aidedreconstruction The position of retrogradely labeled neurons in coronal cortical sections was reconstructed to a standardized lateral view of the rat cortex with the aid of a computer program. This standardized view was defined by drawing a lateral view of eight perfusion-fked brains with the aid of a camera lucida system. These eight plots were then superimposed and their outline and the position of their rhinal fissures was averaged. A coordinate system was formed by an X-axis connecting the most ventral points of the piriform cortex and the pyramidal tract in the pons. The corresponding Y-axispassed through the most caudal extent of the cortical surface. The position of the averaged rhinal fissure in this coordinate system served as the major reference line for the reconstructions. Brains were appropriately blocked and cut perpendicular to the baseline from back to front. Sections were numbered starting from the most caudal cortical end. The position of labeled neurons in each section was measured as the vertical distance to the rhinal fissure. For the lateral view, the X-coordinate was calculated from the number and thickness of the sections, and labeled neurons were plotted in relation to their distance from the averaged rhinal fissure.

Chemoarchitectureof the inferior mllidus In order to further define the subnuclear organization of the inferior colliculus,rats were perfused with 4% paraformaldehyde, and the brains cut in transverse planes at 40 pm on a freezing microtome. The sections were then stained histochemically for AChE (Mesulam, ’82), CyOx (WongRiley, ’791, and NADPH-diaphorase (Scherer-Singler et al., ’83).

RESULTS Chemoarchitectureof the IC Previous investigators have described the cytoarchitecture of the rat 1C on the basis of Golgi-, Nissl-, and myelin-stained preparations. The work of Faye-Lund and Osen (’85) has provided a generally accepted terminology for the subdivisions of the rat IC (see also Paxinos and Watson, ’86, plates 48-57): a CN; a DCl-3; and a laterally located EC1-3. The two cortices each consist of three layers, the outermost of which (layer 1)is common to the two cortices covering most of the IC (for discussion on the comparative anatomy of the mammalian IC see Morest and Oliver, ’84; Faye-Lund and Osen, ’85). Since our observations on the extent of the CN versus the DC differ from Faye-Lund and Osen (’85), we add some information here on the IC subdivisions based on different histochemical stains. This will be important in view of our connectional data on the corticocollicularpathway. Central nucleus. The NADPH staining proved to be most valuable for delineating the IC subnuclei. The CN, which is characterized by the complete lack of NADPH activity, clearly stands out against the dorsal and external cortices (Figs. lB, C, 2). In contrast, in the AChE-stained sections the CN exhibits the highest activity. Note the Complementary staining of the CN in the NADPH-diaphorase versus the AChE material (compare Fig. lB, C and lB’, C‘). Furthermore, in CyOx-stained sections, the CN also

105

stands out clearly from the DC and EC because of the dark staining of its neuropil (not illustrated). Dorsal cortex. In the NADPH-stained material, layers 1 and 2 of the DC exhibit a densely stained neuropil with a moderate number of stained cell bodies (Fig. 2). The darkly stained DC2 is most prominent in the caudal two-thirds of the IC (Fig. 1B-D). The DC1, which is continuous with EC1 and also termed “fibrocellular capsule,” wraps around the IC proper (Figs. 1B-D, 2). In the AChE preparations the DC2 is characterized by a complete lack of staining (Fig. 1B’-D‘). Layer 3 of the DC is outlined by evenly distributed NADPH-diaphorase-positiveneurons. These mostly multipolar neurons do not exhibit lamination or orientation (Fig. 2). AChE activity in DC3 is only moderate. The differential pattern of NADPH, AChE, and CyOx activity in the DC versus the CN reveals a boundary between the CN and layer 3 of the DC, which extends much further ventromedially than proposed by Faye-Lund and Osen (’85). External cortex. The three layers of the laterally located EC are clearly delineated by both histochemical stains. Layers 1 and 3 stand out because of the NADPHpositive neuropil (Figs. 1B-D, 21, while layer 2 of the EC is almost free of staining (Figs. lB, 2). Again, the AChE activity is complementary with patches of darkly stained neuropil in EC2 that is most prominent at the midlevel of the IC (Fig. lB’, C’).Layer 1is free of staining while layer 3 contains only moderate AChE activity. Intercollicular zone. In rat and mouse, the cell mass rostra1 to the central nucleus of the IC is interpreted as the rostromedial extension of the external cortex (Willard and Ryugo, ’83; Faye-Lund and Osen, ’85). We prefer to distinguish this region from the EC primarily based on our connectional data (see below), and will refer to it as the intercollicular zone (IZ; see Robards et al.,’76; Bjoerkeland and Boivie, ’84). In neither histochemical stain does the IZ exhibit a clear staining pattern. Both the NADPH-diaphorase and AChE histochemistry resulted in a patchy appearance of stained neuropil. However, the staining was again complementary with AChE activity in the center of the IZ, contrasted by NADPH activity in the outer portions (compare Fig. 1A and 1A‘).

Auditory cortical areas and fields The scheme of the auditory cortex and the surrounding cortical areas (Fig. 6) is based on the studies of Krieg (’46a), and was later confirmed by Zilles (’85) and Zilles and Wree (’85).Our delineation, however, is a result of the retrograde tracing experiments presented below. We have outlined two secondary auditory areas that are homologous to Krieg’s area 36 and 20 (Zilles’ Te2 and Te3), as well as a primary auditory area that is homologous to Krieg’s area 41 (Zilles’ Tell based cn the retrograde labeling. Furthermore, we have subdivided the primary auditory area into a posterior fTel.p), a middle (Tel.m), and an anterior field (Tel.a), adapting the nomenclature of Zilles. We have indicated a secondary somatosensory area rostrally, comprising part of Krieg’s area 40 (Zilles’ Par2). The dorsal belt, covering the primary auditory area, presumably comprises parts of Krieg’s areas 39 and 18a (see Fig. 14). In order to better comprehend the different cell pattern in the AC, we have superimposed our scheme of cortex organization with the reconstructions of the lateral aspect of the temporal hemispheres (Figs. 6-8, 11, 12).

Figure 1

AUDITORY CORTICOCOLLICULAR PROJECTIONS

107

Fig. 2. Photomicrograph of a NADPH-stained coronal section through the midlevel of the inferior colliculus demonstrating the different IC subnuclei. Scale = 0.5 mm.

however, was completely free of labeling, and hence was judged not to receive auditory cortical input. To demonstrate the total extent of the auditory efferent Retrograde tracing experiments: IC iqjections. To termination within the IC, we placed several large injecidentify the cells of origin of the corticocollicularprojection, tions of WGA-HRP into the auditory cortex and mapped the subnuclear distribution of the anterograde axonal labeling discrete injections of WGA-HRP were made into various IC in the IC. Presumed terminal labeling was found in layers subnuclei (Fig. 4) that had been previously shown to receive 1-3 of the dorsal and external cortex, as well as rostrally in descending cortical fibers (Fig. 3). The retrogradely labeled the intercollicular zone (Fig. 3). The central nucleus, neurons were located exclusively in layer V (Fig. 5). Our observations on the laminar distribution, size, and morphology of these corticocollicular neurons are in general agreement with the study of Games and Winer ('881, and thus will Fig. 1. Series of photomicrographs through four different levels of not be further discussed here. the inferior colliculus illustrating the chemoarchitecture in NADPHCentral nucleus. In two animals, WGA-HRP injections stained coronal sections (left column: A-D, rostrd to caudal) and AChE-stained coronal sections (right column: A'-D', rostral to caudal). were confined to the CN. As demonstrated by experiment IC 15 (Fig. 6), retrograde labeling in the auditory cortex was See text for descriptions of the IC subnuclei. Scale = 1 mm.

Tracingexperiments

H. HERBERT ET AL.

108

H Fig. 3. Line drawings of coronal sections through the inferior colliculus (A-H, rostral to caudal) illustrating the distribution of anterograde axonal labeling (shaded areas) following multiple WGAHRP injections covering most of the auditory cortex. Dots indicate presumed terminal labeling, lines indicate anterogradely labeled fibers. Note that differences in the density of labeling in the IC are not demonstrated.

insignificant.Very few weakly stained neurons were present in the ipsilateral AC, while numerous retrogradely labeled neurons were found in auditory brainstem nuclei as well as dense anterograde axonal labeling in the medial geniculate body. We conclude from this that the CN does not receive any descending projections from auditory cortical areas. In the following experiments, we therefore do not consider any

spread of tracer into the CN as relevant for the corticocollicular projection. Dorsal cortex. In experiment IC 10, a large WGA-HRP injection was placed into the inferior colliculus covering most of the DC and parts of the IZ. Retrogradely labeled neurons were present bilaterally in auditory and nonauditory cortical areas (Fig. 6). Numerous intensely labeled neurons were located in the auditory areas Tel, Te2, and Te3. In addition, few neurons were present in the Par2 and in the dB. Labeled neurons in the contralateral cortex were distributed in the same pattern as ipsilaterally, although the number was considerably smaller. In the following, we present experiments with injection sites confined to individual layers or subnuclei of the IC. The cases described are representative experiments demonstrating the specific connection pattern of the respective IC subnuclei. DClJECl. Injections into the superficial layers of the inferior colliculus, i.e., EC1 caudally and DC1 dorsally, consistently resulted in retrogradely labeled neurons in the ipsilateral secondary auditory area Te2. The number and extent of labeled neurons in Te2 largely depended on the size of the tracer deposit. Following a WGA-HRP injection covering most of EC1 (Fig. 8, experiment IC 12), numerous retrogradely labeled neurons were present almost exclusively in the secondary auditory area Te2. A smaller, more restricted tracer injection into the DC1 (Fig. 8, experiment IC 72) only resulted in a patch of weakly labeled neurons in Te2. DC2. Tracer injections placed into layer 2 of the DC resulted in a characteristic pattern of retrogradely labeled neurons in the auditory cortex (Fig. 7). In experiment IC 47, a tracer injection was placed into the midlevel of DC2, extending slightly ventrally into the dorsolateral portion of DC3 as well as rostrally into the IZ. DC1, of course, was always involved in these injections. The intensely labeled neurons in Te1.p and Te1.a were characteristic of this injection, while Te1.m contained only a patch of labeled neurons ventrocaudally. Furthermore, clusters of labeled neurons were found in Te2, Te3, and Par2. In the contralateral cortex, labeling was present in all homotypic loci except Par2. However, the staining intensity of the labeled neurons was weaker, and their total number much lower than on the ipsilateral side (Fig. 7, experiment IC 47). In experiment IC 46, a similar but considerably smaller injection was placed into DC2 (Fig. 4). As a result, an identical Dattern of retromade labeling occurred in the AC. although-the number of Labeled cells Gas much lower (Fig: 7, experiment IC 46). The most prominent labeling was present in Te1.p and Te1.a. In addition, a few poorly labeled neurons were found ventrally in Tel.m, as well as in Te2 and Te3. No labeled neurons were found in the contralateral AC. DC3. WGA-HRP injections into DC3 led to a pattern of retrograde labeling in the primary auditory fields that was complementary to the labeling found after injections into DC2. In experiment IC 18, a fairly large WGA-HRP injection was placed into the ventromedial portion of DC3 (Fig. 41, involving the superficial layers medially and the mediocaudal extent of the IZ. Numerous retrogradely labeled neurons were found bilaterally in the dorsorostral part of Tel.m, while Te1.p and Te1.a were almost free of labeling (Fig. 7, experiment IC 18). Moreover, labeled cells were

AUDITORY CORTICOCOLLICULAR PROJECTIONS

-

109

I___

Fig. 4. Bnghtfield photomicrographs of coronal sections through the midbrain showing the centers of WGA-HRP injection sites in four subnuclei of the inferior colliculus. The illustrated sites are located in layer 3 of the dorsal cortex (experiment IC 18),in layers 1and 2 of the

dorsal cortex (IC 46), in the external cortex (IC 341, and in the intercollicular zone (IC 59).See Figs. 6-8 for correspondingretrograde labelingin the temporal cortex Scales = 0 5 m m

present bilaterally in Te2 and Te3, as well as in the dB. The caudal portion of Te2 did not contain any retrogradely labeled cells. A much smaller injection into the ventromedial corner of DC3 resulted in a similar, even more distinct pattern of retrogradely labeled neurons dorsorostrally in Tel.m, in the secondary areas Te2 and Te3, and in the dB (Fig. 7, experiment IC 33). No labeled neurons were present contralaterally . External cortex. The origin of neocortical projections into the EC is demonstrated by experiment IC 34, in which the WGA-HRP injection was centered ventrolaterally in the EC proper (Fig. 4). The injection resulted in a moderate number of weakly labeled neurons that were diffusely distributed over ipsilateral primary and secondary auditory areas (Fig. 8, experiment IC 34). In a few cases, scattered neurons were also found in Par2 and the dB.

Intercollicular zone. Tracer injections placed into the IZ resulted in a rather unique pattern of retrogradely labeled neurons ipsilaterally in auditory and nonauditory cortical areas. In experiment IC 59, where the injection was clearly restricted to the IZ (Fig, 4), the largest numbers of retrogradely labeled cortical neurons were found in the secondary auditory area Te3, the secondary somatosensory area Par2, and the dB, wrapping around the primary auditory areas (Fig. 8, experiment IC 59). Labeled neurons were also present in the secondary auditory area Te2, and a few scattered neurons were found in the primary area Tel. Anterograde tracing experiments: AC injections. To further clarify the subnuclear termination of the corticocollicularprojection, we placed restricted WGA-HRP injections into the different auditory areas and fields (e.g.,Fig. 9) and plotted the distribution of anterograde axonal labeling in the IC (Figs. 11,12).

110

H. HERBERT ET AL.

Fig. 5. Left panel: Brightfield photomicrograph of a coronal section through the temporal cortex illustrating retrogradely labeled neurons of layer V in the secondary auditory area Te2 and the primary auditory field Te1.m following a WGA-HRP injection into layer 3 of the dorsal cortex (not illustrated). Arrowheads indicate the borders be-

tween auditory areas and fields. A and B: Enlargements of retrogradely labeled neurons in Te1.m and Te2, respectively. Note the different size and shape of primary (A) and secondary (B) auditory corticocollicular projection neurons. Scales = 250 pm.

Te1.p. Experiment AC 1 is an example of a WGA-HRP injection into the posterior primary field. The tracer deposit was located ventrally in Tel.p, extending into the ventrocaudal corner of Te1.m (Fig. 11).Intense anterograde labeling was found dorsally in DC2, starting in the most caudal sections and extending rostrally to the beginning of the IZ. The zone of sparse labeling dorsal to DC2 (Fig. 1OA) corresponds to DC1. The lateral portion of DC2 also exhibited some axonal labeling, though it was less intense. Some weak labeling was also present in the dorsolateral part of DC3 (see below: Te1.m injections). Te1.m. We present two cases with injection sites in the middle primary field differing in their rostrocaudal location. In experiment AC 2, we injected the ventrocaudal portion of Te1.m (Figs. 9, 11).Most prominent was the dense axonal labeling in DC3. Caudally it was present as a round dense

patch, while at the midlevel of DC3 the terminal labeling appeared as distinct bands oriented from dorsomedial to ventrolateral (Fig. 10B). In the rostral third of DC3, the axonal labeling became considerably weaker. The labeling extended into the ventral aspect of DC2, and some anterograde labeling was also present in the EC. The injection in experiment AC 3 was placed into the rostral portion of Te1.m just rostrally to AC 2 (Fig. 11). Again, dense axonal labeling was present in DC3, but it was located more medially in the ventromedial portion of DC3 (compare experiments AC 2 and AC 3 in Fig. 11). In addition, some weak labeling was found in EC. In short, injections placed from ventrocaudal to dorsorostral into Te1.m result in anterograde axonal labeling in DC3 moving from dorsolateral to ventromedial (compare experiments AC 1-3 in Fig. 11and AC 4 in Fig. 12).

111

AUDITORY CORTICOCOLLICULAR PROJECTIONS

-1 I

,,

J

_ - - --

,’dB

...

1

\

\

\

/

\

/

I I

Te1.m /

Par 2

Te3 .

I-

rhf

1

iDsi

I

Te1.a. The WGA-HRP deposit in experiment AC 4 was located in the dorsal part of Tel.a, presumably with some tracer spread into the dorsorostral edge of Te1.m (Fig. 12). The major terminal field was found in the lateral portion of DC2 reaching into EC2. This labeling was present from the very caudal portion of the DC to approximately two-thirds of the IC rostrally. The dorsal portion of DC2 contained only weak anterograde labeling. A second band of strong labeling was present in the ventromedial portion of DC3 and is attributed to the spread of tracer into the dorsorostral edge of Te1.m (compare with AC 3). Te2. The injection in experiment AC 5 (Fig. 12) was located in the ventrorostral portion of Te2. The anterograde axonal labeling was largely confined to the superficial layers wrapping around the IC. In DC1 the labeling was most dense and was found along the entire rostrocaudal extent of the IC. Weaker labeling was present laterally and caudally in EC1. Layers DC2 and DC3 were free of anterogradely labeled neuropil. Furthermore, anterograde labeling was found in the caudal portion of the commissure of the IC, extending slightly into the contralateral DC1 as well as ventromedially between DC3 and the aqueduct. Te3. WGA-HRP injections placed into Te3 are illustrated by experiment AC 6 (Fig. 12). The injection site was located rostrally in Te3, extending slightly into Para. Prominent axonal labeling in the IC was only found rostrally in the intercollicular zone. Labeling laterally in the EC abruptly became weaker, and in the caudal two-thirds of the EC only poorly labeled neuropil remained, A strong band of intense labeling remained medially. Further caudally only sparse labeling was present in superficial layers of the IC.

Summarized pattern of corticocollicular projections Based on the retrograde and anterograde tracing experiments we propose the following pattern of major projections from the cortex to the inferior colliculus (summarized in Fig. 13): 1. Te1.m primarily innervates DC3. This pathway is bilateral and topographically organized. Neurons located dorsorostrally in Te1.m project into the ventromedial portion of DC3, while neurons located ventrocaudally in Te1.m project into the dorsolateral portion of DC3 (Figs. 7, 11). 2. Te1.p and Tel.aproject bilaterally into DC2. The data from the anterograde tracing experiments further suggest that Te1.p primarily innervates the dorsal portion of DC2, while Te1.a primarily innervates the lateral portion of DC2 (Figs. 7, 11, 12). 3. The secondary auditory cortex Te2 primarily innervates the superficial layers of the IC caudally and laterally (ECl), dorsally (DCl), and medially, as well as the caudal portion of the commissure of the IC. The projection is largely ipsilateral. Only a weak projection extends through the commissure into the medial aspect of the contralateral DC1 (Figs. 8,12).

Fig. 6. Camera lucida drawings of coronal sections showing the centers of WGA-HRP injection sites in the inferior colliculus (experiments IC 15, IC 10) and the respective distribution of retrogradely labeled neurons in computer-assistedreconstructions of lateral views of the temporal cortical surface. Scales = 1 mm (also refer t o Figs. 7 and 8).

IC 47

IC 18

_----... ..

,

i

, /

/

,------.. .. \

\

/ /

I

I

, \

I

..

.

,

..

\ /

/

R

contra

IC 46

IC 33

Figure 7

AUDITORY CORTICOCOLLICULAR PROJECTIONS 4. The EC is not a major target of corticocollicular neurons since it receives only a weak and diffuse ipsilateral projection from primary and secondary auditory areas (Figs. 8, 11, 12). 5 . The IZ receives its main cortical input from the secondary auditory field Te3, as well as from nonauditory cortical areas including Par2 and dB. The projection is only ipsilateral (Figs. 8, 12). 6. The CN receives no temporal cortical input at all (Figs. 3, 6, 11, 12).

113

Superficial layers (DClIECl). Following WGA-HRP injections into the superficial layers of the IC, we found retrogradely labeled neurons exclusively in Te2 (Fig. 8, experiments IC 12 and IC 72). In turn, tracer injections into Te2 resulted as expected in anterograde labeling in the superficial layers of the IC with some labeling on the contralateral side (Fig. 12, experiment AC 5). This is in line with Faye-Lund's ('85) findings in the rat. She too reports anterograde axonal labeling posterior and dorsal in DC1 after injections into area 36. A bilateral projection from the secondary auditory area MI into the superficial layers of 1C is also reported for the DISCUSSION cat (Diamond et al., '69; Andersen et al., '80) and the tree shrew (Oliver and Hall, '78). In the latter, it was the The corticocollicularpathmy caudally located secondary area (area C) that innervated Our connectional data demonstrate a differential projec- the superficial layers of the IC. Hence, it seems to be a tion pattern of auditory and nonauditory cortical neurons general feature across species that a secondary auditory to different subnuclei of the IC in the rat. cortical area innervates a thin shell of neuropil wrapping Projections into the central nucleus. The present around most of the IC proper. tracing experiments revealed that in the rat the CN is not a DCZ. Injections involving DC2 consistently resulted in target of descending auditory or nonauditory cortical projec- retrogradely labeled neurons, in both the anterior (Te1.a) tions (Fig. 3 and experiment IC 15 in Fig. 6). This finding is and posterior fields (Te1.p) of the primary auditory cortex in line with most studies on the corticocollicular pathway (Fig. 7, experiments IC 46 and IC 47). However, following (for references see opening section). Only Beyerl ('78) tracer injections into either of these cortical fields, we reports retrogradely labeled neurons in the rat's AC follow- found that while Te1.p efferents terminate mainly dorsally ing HRP injections into the CN. However, considering the in DC2, Te1.a efferents terminate mainly laterally in DC2 amount of tracer injected (400 nl) and the size of the CN, (Figs. 11and 12, experiments AC 1and AC 4, respectively). his retrogradely labeled neurons are likely to be due to Because of the proximity of the lateral and dorsal parts of tracer spread into the DC. DC2, we assume that our WGA-HRP injections into the DC Auditory corticocollicular projections terminating in the mostly involved both portions, and hence led to retrograde CN are consistently reported for primates (Kuypers and labeling in both Te1.a and Te1.p. Faye-Lund ('85) reports Lawrence, '67; Forbes and Moskowitz, '74; Fitzpatrick and area 41 efferents that are distributed within DC2 and DC3. Imig, '78). Following 3H-leucineinjections into the primary She does not differentiate between anterior and posterior auditory cortex, bands of anterograde labeling oriented components of this pathway from the primary auditory from dorsomedial to ventrolateral were present in the cortex. Her injections may have exceeded the proposed ventrolateral central nucleus (corresponding to our CN). borders of the primary fields and hence obscured this These bands extend into the dorsomedial central nucleus, differential innervation. which corresponds to our DC. The intensity of labeling, In the cat it was demonstrated that the projection from however, was considerably weaker in the ventrolateral the auditory field AI and AAF into DC2 (as well as DC3, see compared to the dorsomedial portion (Forbes and Moskowitz, '74; Fitzpatrick and Imig, '78). In nonprimate mam- below) is topographically organized (Andersen et al., '80). For example, injections of 3H-leucine into high-frequency mals, the CN receives exclusively ascending projections from auditory brainstem nuclei and the contralateral IC loci in AI led to anterograde labeling laterally in DC2 (their (e.g., Rockel and Jones, '73a,b; Willard and Martin, '83; pericentral nucleus), while injections into low-frequency Morest and Oliver, '84; Oliver and Morest, '84; Coleman loci led to labeling medially in DC2. The topography in DC2 complements the tonotopic organization in this subnucleus and Clerici, '87). Projections into the dorsal cortex. Anatomical and with high frequencies represented laterally and low frequenphysiological studies agree that the DC is the major termi- cies medially (Merzenich and Reidt, '74). The frequency nation area of descending auditory cortical projections (for axis, however, is oriented antiparallel compared to the one references see opening section). However, only a few au- in CN and DC3. Comparable data about frequency representhors demonstrated that primary and secondary areas tations in different subnuclei of the rat's IC are lacking. DC3. The present study demonstrated that Te1.m innerdiffer in their connectivity with the IC. Primary auditory areas predominantly innervate the deeper layers of the DC, vates DC3 bilaterally and that this projection is topographiand it was shown that this pathway is topographically cally organized. Neurons located dorsorostrally in T e l .m organized (Diamond et al., '69; Oliver and Hall, '78, Casse- project to the ventromedial extent of DC3, while neurons day et al., '79; Andersen et al., '80; Willard and Ryugo, '83, located ventrocaudally in Te1.m project into the dorsolatera1 portion of DC3 (Fig. 7, experiments IC 18 and IC 33). Faye-Lund, '85). Following distinct injections of WGA-HRP into Te1.m that shifted from caudal to rostral, bands of anterograde axonal labeling in DC3 shift lateromedially (Fig. 11, experiments AC 1-AC 3).The orientation of these bands (Fig. 10B) is the same as the orientation of dendritic lamellae in the IC, as Fig. 7. Camera lucida drawings of coronal sections showing the demonstrated by Faye-Lund and Osen with the Golgi centers of WGA-HRP injection sites in the inferior colliculus (experitechnique ('85; see their Fig. 5B). Furthermore, these ments IC 47, IC 46, IC 18, IC 33) and the respective distribution of lamellae are in the same range of thickness as our bands of retrogradely labeled neurons in computer-assisted reconstructions of axonal labeling and are also separated by less dense zones. lateral views of the temporal cortical surface.

H. HERBERT ET AL.

114

IC 12

IC 34

IC 72

IC 59

Fig. 8. Camera lucida drawings of coronal sections showing the centers of WGA-HRP injection sites in the inferior colliculus (experiments IC 12, IC 72, IC 34, IC 59) and the respective distribution of retrogradely labeled neurons in computer-assisted reconstructions of lateral views of the temporal cortical surface.

It is likely, therefore, that the bands of axonal labeling represent corticocollicular terminals making synaptic contacts on dendritic lamellae formed by neurons in DC3. Diamond et al. ('69) were the first to report the topography of the corticocollicular pathway in the cat. Later, Andersen et al. ('80) injected 3H-leucine into frequencycharacterized loci in the primary auditory cortex AI or the AAF and found distinct bands of anterograde labeling

oriented from dorsomedial to ventrolateral in the DC (their dorsomedial portion of the central nucleus): injections into low-frequency loci in AI or AAF resulted in labeling dorsolateral in the DC, while injections into high-frequency loci resulted in labeling ventromedial in the DC. This cochleotopic termination of AI efferents within DC is in line with physiological studies. In the cat IC, there is a tonotopic gradient with low frequencies represented dorsolaterally

AUDITORY CORTICOCOLLICULAR PROJECTIONS

115

terminal fibers in the DC of the inferior colliculus may provide a morphologicalbasis for the integration of cortical and lower brainstem auditory information. Interestingly, the same phenomenon has been found in the central nucleus of the IC for ascending terminal fibers originating in different nuclei of the cat auditory brainstem (Henkel and Spangler, '83; Spangler et al., '85; Shneiderman and Henkel, '87). They concluded that the laminar pattern may be directly related to the tonotopic organization of the IC and allows for the convergence of auditory information from different levels of the neuraxis. Frequency representation in rat IC subnuclei has also been demonstrated with the 2-DG technique. Pure-tone stimulation led to bands of radioactive label in the IC oriented from dorsomedialto ventrolateral (Gonzales-Lima and Scheich, '84;Huang and Fex, '86; Ryan et al., '88). The isofrequency bands were not restricted to the CN, but extended into the DC with low frequencies represented dorsolaterally and high frequencies ventromedially. In conclusion, our study demonstrates that the location and orientation of the anterogradely labeled terminals in DC3 following Te1.m injections are similar to: 1) the dendritic laminae in IC; 2) the frequency-specificascending fibers; 3) physiologically determined isofrequency contours; and 4) pure-tone-elicited 2-DG bands. This suggests that in the rat, the projection from Te1.m into DC3 is also tonotopically organized. Electrophysiological mapping studies have revealed that in the rat primary auditory cortex, a low- to high-frequency gradient is present from caudal to rostral (Syka et al., '80; Sally and Kelly, '88). This further supports Fig. 9. Brightfield photomicrograph of a large WGA-HRP injection our assumption of a cochleotopic projection, since caudal site (experiment AC 2 ) in the medial primary auditory field (Te1.m). (low-frequency) and rostral (high-frequency) loci in Te.m Arrows indicate labeled fibers leaving the injection site through the project into dorsolateral (low-frequency)and ventromedial external capsule. See Fig. 11for the total extent of this injection site in (high-frequency)regions of DC3, respectively. Our data are Te1.m and the corresponding anterograde axonal labeling in the also in line with the topography found in the thalamocortiinferior colliculus. Scale = 0.5 mm. cal pathway in the rat: a lateromedial progression within the ventral nucleus of the MGB (presumably representing a low- to high-frequency gradient) induces a shift of correand high frequencies ventromedially, and the isofrequency contours extend continuously from the CN into the DC sponding loci upon Tel from ventrocaudal to dorsorostral (Sempleand Aitkin, '79). In the mouse, a complete represen- (Roger and Arnault, '89). Projections into the external cortex and intercollicular tation of all frequencies was present only if all collicular zone. The EC and the IZ are the only IC subnuclei that subnuclei, including CN and DC, were combined (Stiebler and Ehret, '85). These physiological studies demonstrate receive auditory as well as nonauditory cortical input. Auditory cortical efferents. Both the EC and I2 receive that the parameter "frequency" overrides cytoarchitecminor efferents from all auditory cortical areas. In addition, tural and anatomical boundaries. Comparative detailed data on the rat IC are not available. the IZ is prominently innervated by the secondary auditory Clopton and Winfield ('73) and Syka et al. ('80)found a area Te3 (Fig. 8, experiments IC 34 and IC 59).Anterograde frequency representation in the rat IC similar to the cat, tracing experiments resulted in weak anterograde axonal with low frequencies dorsolaterally and high frequencies labeling in layers 2 and 3 of the EC, and prominent labeling ventromedially. However, they did not differentiate be- in the IZ following Te3 injections (Figs. 11, 12). These findings are largely consistent with other studies in rats. tween a central nucleus and a dorsal cortex. Recent experiments in our laboratory revealed that injec- Faye-Lund's ('85) WGA-HRP injections into primary AC tions of the anterograde tracer Phaseolus vulgaris-leucoag- did not result in anterograde terminal labeling in the EC or glutinin into frequency-characterized loci in the cochlear IZ, while injections ventrorostrally into the auditory temponucleus result in discrete bands of labeled fibers in the IC ral field (her case 461 in Fig. 9, comparable to our Te3 oriented from dorsomedial to ventrolateral. These fibers injection) led to anterograde labelinglaterally in the EC and clearly extend from the CN into DC3 (Herbert, unpublished rostrally in the IZ. Coleman and Clerici ('87) also report an data). Thus, DC3 is provided with ascending auditory ipsilateral descending input to the EC from sparsely labeled information of discrete frequencies. Similar findings were area 41 neurons, while more labeled corticocollicular cells reported for the cat, in which efferents from the cochlear appeared in area 39. Furthermore, investigations in several other animals nucleus and from the dorsal nucleus of the lateral lemniscus also terminate in a banding pattern in the deeper layers including cat, marsupial, and primate confirm a minor of the dorsal cortex of the IC (Oliver, '84, '87; Shneiderman ipsilateral auditory input into the EC and IZ (Diamond et et al., '88). The interdigitation of descending and ascending al., '69; Rockel and Jones, '73b; Martin et al., '75; Cooper

H. HERBERT ET AL.

116

Fig. 10. Polarization photomicrographs of two coronal sections through the inferior colliculus (dorsal is to the top, medial to the right) illustrating the differential anterograde axonal labeling in the IC following WGA-HRP injections into (A) the posterior primary auditory field (Tel.p, experiment AC 11, and (B) the medial primary auditory field (Tel.m, experiment AC 2). A The Te1.p injection resulted primarily in anterograde labeling in the medial portion of layer 2 of the dorsal cortex while the labeling in the lateral portion of DC2 is much

and Young, '76; Fitzpatrick and Imig, '78; Andersen et al., '80; Land et al., '84). Nonauditory cortical egerents. We demonstrate that the IZ, and to a lesser extent the EC, are innervated by presumed multisensory (dB) and somatosensory (Par21 cortical areas (Fig. 8, experiment IC 59). Faye-Lund's ('85) injections into area 22,,1, corresponding in location to our dorsal belt, also resulted in anterograde axonal labeling ipsilaterally in EC2 and EC3, but only in poor labeling rostrally in the IZ. Multimodal inputs from somatosensory, and even motor and visual, cortices into the IZ and the EC are also reported for other animals, including marsupial, cat, and primate (Kuypers and Lawrence, '67; Martin and Megirian, '72; Martin et al., '75; Cooper and Young, '76; Robards et al.,'76; Robards, '79). Hence, the EC and the IZ differ considerably from other IC subnuclei in receiving only poor auditory but strong multisensory input from neocortex.

Cortical subdivisions The present study suggests a parcellation of the rat's temporal cortex that is based on the corticocollicular projection pattern. In some respects, our scheme of cortex is different from parcellations, which are based on histological 'Although working with rats, Faye-Lund has adopted the nomenclature for the mouse temporal cortex as published by Caviness, '75 and Caviness and Frost, '80 (illustrated in Fig. 14). In the following we will indicate the mouse nomenclatureby the index: cau since it differs in some respects from the rat nomenclatureby Krieg 1'46a).

weaker. The zone of sparse labeling dorsal to DC2 corresponds to layer 1. B: The Te1.m injection resulted in dense terminal labeling in layer 3 of the dorsal cortex. Note the two bands of labeling in DC3, separated by a zone of less dense labeling. A third band of anterograde axonal labeling is visible ventromedially between DC3 and the adjacent central gray matter. See Fig. 11for location of the injection sites and the total extent of terminal labeling in the IC subnuclei. Scale = 0.5 mm.

criteria as well as on various connectional studies (summarized in Fig. 14). Primary auditory area. The primary auditory area is located centrally in the temporal cortex and corresponds to Krieg's area 41 (Zilles' Tel). We have subdivided T e l into three fields based on the corticocollicular connections (Fig. 14): Te1.p and Te1.a innervate layer 2 of the dorsal cortex of IC, while Te1.m topographically innervates layer 3 of the dorsal cortex of IC. Faye-Lund ('85) and Willard and Ryugo ('83) also demonstrate a strong corticofugal projection from area 41 to the deep layers of the DC. However, they do not report different projections of discrete parts of area 41. Their fairly large lesions or injections might have obscured a more differential pattern. Several other investigators have delineated or further subdivided the primary auditory cortex based on cyto- and myeloarchitecture (Krieg, '46b; Ryugo, '76; Patterson, '77; Willard and Ryugo, '83; Zilles, '85; Zilles and Wree, '85; Schober, '86), on auditory thalamocortical and corticothalamic projections (Ryugo, '76; Patterson, '77; Willard and Ryugo, '83; Faye-Lund, '85; Scheel, '88; Roger and Amault, '89), and on callosal termination patterns (Ryugo, '76; Cipolloni and Peters, '79; Vaughan, '83; Riittgers et al., '90). However, none of the reported subdivisions of the primary auditory cortex corresponds completely with our parcellation (compare the temporal cortices in Fig. 14). The different parcellations of T e l are based on the thalamocortical, callosal, or corticocollicular connectivity, and these pathways either terminate or originate in separate cortical

AUDITORY CORTICOCOLLICULAR PROJECTIONS

117

f

U

-

Fig. 11. Lateral views of the left temporal cortex with reconstructed injection sites in three different auditory areas and fields (top row; experiments AC 1-AC 31, and below, camera lucida drawings of the corresponding anterograde labeling in subnuclei of the inferior collicu-

lus (rostral to caudal from top to bottom). Fine dots depict axonal and terminal labeling; thin lines depict anterogradely labeled fibers. Scales = 1mm.

layers. Hence, it is feasible that individual functional circuits occupy different, layer-specific fields of Tel. It would be interesting to know whether any of the proposed parcellations of the rat's primary auditory cortex may also be reflected in physiological differences. Secondary auditory areas. The secondary auditory cortex includes Krieg's areas 36 and 20 (corresponding to Zilles' Te2 and Te3, respectively). In the present study, we demonstrate that both Te2 and Te3 have specific corticocollicular projections. Efferents originating in Te2 primarily terminate in the superficial layers of the IC, including DC1 and EC1. This is in agreement with the findings of FayeLund ('85).In contrast, efferents originatingin Te3 predominantly innervate the intercollicular zone of the IC.

Previously, two secondary areas were also verified on histological grounds (Krieg, '46b; Ryugo, '76; Zilles et al., '80; Zilles, '85; Zilles and Wree, '85; Schober, '861, as well as on thalamocortical (Ryugo, '76; Patterson, '77; Faye-Lund, '85), callosal (Ryugo, '76; Vaughan, '83; Cipolloni and Peters, '79; Ruttgers et al., '901, and intracortical connections (Miller and Vogt, '84). Most authors agree on the relative location of the two secondary auditory areas (illustrated in Fig. 14), but disagree on the absolute extent of or certain subdivisionswithin, secondary areas (e.g., Miller andVogt, '84; Riittgers et al., '90). NonauditoryJintegrativecortex areas. In our study we found that neurons in a cortical field surround Tel and project into the multisensory intercollicular zone of the IC

H. HERBERT ET AL.

118

Fig. 12. Lateral views of the left temporal cortex with injection sites (experiments AC 4-AC 6), and corresponding anterograde labeling in the inferior colliculus (for details see legend of Fig. 11).

(Fig. 14). This dorsal belt (dB) presumably comprises parts of Krieg’s areas 18a and area 39, the latter of which he calls “the great sensory correlation area.” Another cluster of cortical neurons projecting into the IZ is located within the secondary somatosensory area Par2. Faye-Lund (’85)found neurons innervating the EC also surround area 41 and termed it, according to Caviness (’75), area 22,”. So far, however, in the rat there is no area defined that correlates with area 22,”. Zilles and Wree (’85)emphasize an inhomogenity in Nissl- and myelin-stained preparations resulting in a “ring-like’’ subfield surrounding the center of Tel (see Fig. 14). This “ring-like” subfield might represent a comparable area in the rat. Several anatomical studies point out the multisensory features of the cytologically characterized areas 40 and 39

and a not yet clearly defined region dorsal to the primary auditory cortex. These parts of the cortex have afferent and efferent connections with various cortical areas and brain nuclei of different sensory modalities (Krieg, ’46a,b; ’47; Jones and Leavitt, ’74; Ryugo, ’76; Patterson, ’77; Caviness and Frost, ’80; Miller and Vogt, ’84; Faye-Lund, ’85). Furthermore, electrophysiological studies support the idea of a multisensory area dorsal to the primary AC (LeMessurier, ’48;Azizi et al., ’85;Sally and Kelly, ’88). Taken together, these studies demonstrate that in the rat there is a multisensory cortical region (corresponding to our dB) that is located between the auditory and visual cortex and the auditory and somatosensory cortex. In the cat a comparable region was defined and termed “association cortex” (for a review, see Irvine and Phillips, ’82).

AUDITORY CORTICOCOLLICULAR PROJECTIONS

cortex

;

Te1.m

Y

-0

J

m

.-E

h

.-0

i

3

m v

$

Te1.a

+

inf. coil.

r

DC3 DC 2

Te1.p

-

[:

EC 2,3

/

DC 1, EC 1

-

.-

0 J

m

c'

0

C

Fig. 13. Summary diagram illustrating the pattern of the major projections from auditory and nonauditory cortical areas into the subnuclei of the inferior colliculus (inf. coll.), as revealed in the present study. Thick lines indicate strong bilateral auditory projections; thin lines indicate weaker ipsilateral auditory projections; and dashed lines indicate projections originating in auditory and nonauditory areas.

Physiologyof the CortimKcularpathway Early physiological studies of the auditory corticocollicular pathway showed that electrical stimulation of primary auditory cortex suppresses the amplitude of acoustically evoked potentials in the IC (Massopust and Ordy, '62; Amato et al., '70). A generally inhibitory function of this auditory corticofugal projection was concluded. Watanabe et al. ('66) first demonstrated that AC stimulation could also result in facilitation of spike responses in the IC. Recent single-unit recordings revealed a variety of differential inhibitory and/or facilitatory effects of AC stimulation on acoustically driven IC neurons (Syka and Popelar, '84; Syka et al., '88; Sun et al., '89). Three general response types could be distinguished: 1)excitation, i.e., an increase in tone-evoked spike rate; 2) inhibition, i.e., a decrease in spike rate of spontaneously active as well as acoustically driven IC neurons; and 3) an initial excitation followed by a long-lasting(up to 350 ms) inhibition. These response types were found in both the DC and the CN. This is in line with anatomical results demonstrating that the CN is heavily innervated by IC subnuclei that receive major projections from AC (Rockel and Jones, '73b; Coleman and Clerici, '87). In an intracellular recording study on antidromically identified colliculogeniculate neurons, Mitani et al. ('83) found three types of responses after electrical stimulation of the primary auditory cortex: EPSPs, IPSPs, or EPSPIPSP sequences. Short latency responses (I 1.4ms), indicating a monosynaptic connection, were only found in the

119 dorsal cortex of the IC. This is to be expected from anatomical data. Interestingly, only the EPSPs were considered monosynaptic, while the IPSPs, in the CN as well as in the DC, with latencies 2 2.0 ms are likely to be polysynaptic. Mitani et al. ('83) assumed that corticocollicular efferents excite inhibitory interneurons, which in turn innervate colliculogeniculate neurons. This is in line with ultrastructural findings in the cat. Terminals of corticocollicular fibers in the dorsal cortex have spherical synaptic vesicles and end in asymmetrical synaptic contacts on dendritic spines (Rockel and Jones, '73c). These "type I" synapses are generally associated with excitatory functions (Irvine, '86).

Functionalconsiderations The auditory corticocollicular projections may influence processingof acoustic signals in different functional circuits of both the ascending and descending auditory pathways. The descending projection from the primary auditory cortex to the dorsal cortex of the IC is able to modulate the activity of ascending auditory information in the central nucleus of the IC. The central nucleus is heavily innervated by the DC (Rockel and Jones, '73b; Coleman and Clerici, '87), which in turn receives prominent projections from primary AC. Neuronal activity of colliculogeniculate neurons is influenced by cortically driven DC neurons (Mitani et al., '83). Since in cat and rat the projection from primary AC to DC3 is cochleotopicallyorganized, this pathway could perform frequency-specific modulations and play a role in pattern recognition mechanisms. The central nucleus of the IC contains a large number of neurons that project to the cochlear nuclei (Hashikawa, '83; Herbert, unpublished data). Again, these colliculocochlear neurons may be innervated by cortically driven cells in the dorsal cortex, and hence may transmit cortical information to the cochlear nucleus. In the cat, different parts of the IC innervate different cell populations in the dorsal cochlear nucleus. Dorsally located parts of the IC (correspondingto our DC) seem to terminate in layers containing mostly interneurons, while the more substantial projection originating in ventral parts of the IC (corresponding to our CN) predominantly innervates neurons that project to higher auditory levels (Conlee and Kane, '82). The colliculocochlear projection may thus provide a feedback loop for the analysis and modulation of incoming acoustic signals. Minor descending projections from AC terminate in EC3 where the majority of the colliculoolivary neurons are located (Faye-Lund, '86). The EC receives a major input from ipsi- and contralateral CN (Coleman and Clerici, '87; Herbert, unpublished data). Since little is known about the intranuclear connectivity of the IC, we cannot rule out that colliculoolivary neurons in the EC receive descending auditory information from cortically driven IC interneurons or via cortically modulated neurons from the contralateral CN. It is assumed that colliculoolivary fibers terminate predominantly on large olivocochlear neurons (Faye-Lund, '86) that innervate the outer hair cells in the cochlea (Warr et al., '86). This medial olivocochlear system is thought to suppress auditory nerve responses, and may consequently serve as a cortically driven control system to filter distracting acoustic signals, thereby enhancing selective attention to an auditory stimulus (reviewedin Wiederhold, '86). Secondary auditory cortical areas innervate a thin shell of neuropil wrapping around most of the IC proper (for references see above). In the rat, neurons in this shell

H. HERBERT ET AL.

120 1

present study

I

Patterson, '77

L

I

1

Krieg, '46a

Miller and Vogt, '84

1

Cipoiioni and Peters, '79

Zilles and Wree, '85

I

1

J

I

I

I I

Caviness and Frost, '80

I

Fig. 14. Comparison of auditory areas in rat and mouse as defined by different authors (originals are adapted in size and orientation). The subdivisions of the temporal cortex are based on cyto- and myeloarchitectural studies (Krieg, '46A; Caviness, '75; Zilles, '85; Zilles and Wree,

'85), as well as connectional studies, including thalamocortical projections (Ryugo, '76; Patterson, '77; Caviness and Frost, '801, intracortical connections (Miller and Vogt, '84),callosal connections (Cipolloni and Peters, '791, and corticocollicular projections (present study).

project to the suprageniculate, medial, and posterior intralaminar divisions of the MGB, and the peripeduncular area (Arnault and Roger, '87; LeDoux et al., '85). These thalamic areas, in turn, innervate the lateral nucleus of the amygdala (LeDoux et al., '90af, which is discussed as a sensory interface mediating the conditioning of emotional responses to acoustic stimuli (LeDoux et al., '90b). Besides projections from the secondary auditory area Te3 into the intercollicular zone, we found nonauditory projections originating in the secondary somatosensory cortex and the dorsal belt. There is a large body of literature dealing with the multisensory connections of the IZ (e.g., Bjorkeland and Boivie, '84; Itoh et al., '84; Robards et al., '76; Wiberg and Blomqvist, '84; discussed in Faye-Lund, '85). The IZ is assumed to integrate auditory and somatosensory information and may play a role in audiomotor functions, orienting an animal to a sound source, for example (Aitkin et al., '78, '81). A modulatory function of the auditory corticocollicular projection became obvious in a conditioning experiment in which rats had to respond to an auditory stimulus (Birt et al., '78). Bilateral lesions of the AC had a significant impact

on the ability to learn this task. On the neuronal level, the change of background firing rates of IC neurons during the conditioning process was different in lesioned and nonlesioned animals. The authors concluded that during learning of a discrimination task, a nonspecific activation in the IC is selectively antagonized by corticocollicular efferents. Thus, it becomes apparent that the corticocollicular projections fulfill a modulatory role in a variety of fimctional circuits in the IC. Even so, we still need to improve our knowledge about the neuronal elements of the involved microcircuits to better understand the key function of the IC as a "distributor" of cortical information to different levels of the neuraxis.

ACKNOWLEDGMENTS The authors thank Helga Zillus for technical assistance, Karin Ruttgers for access to histological material,, Dr. Dieter Menne for writing the computer program, and Krista Nadakavukaren for correcting the English. The most helpful comments of Dr. Ecki Friauf on the manu-

AUDITORY CORTICOCOLLICULAR PROJECTIONS script are also appreciated. This work was supported by the Deutsche Forschungsgemeinschaft SFB 307.

Aitkin, L.M., H. Dickhaus, W. Schult, and M. Zimmermann (1978) The external nucleus of the inferior colliculus: Auditory and spinal somatosenwry afferents and their interactions. J. Neurophysiol. 412337-847. Aitkin, L.M., C.E. Kenyon, and P. Philpott (1981) The representation of the auditory and somatosensory systems in the external nucleus of the cat inferior colliculus. J. Comp. Neurol. 1 9 6 2 5 4 0 . Amato, G., V. LaGrutta, and F. Enia (1970) The control of acoustic input in the medial geniculate body and inferior colliculi by auditory cortex. Experientia 26:55-56. Andersen, R.A., R.L. Snyder, and M.M. Merzenich (1980) The topographic organization of corticocollicular projections from physiologically identified loci in the AI, MI, and anterior auditory cortical fields of the cat. J. Comp. Neurol. 191:479-494. Arnault, P., and M. Roger (1987) The connections of the peripeduncular area studied by retrograde and anterograde transport in the rat. J. Comp. Neurol. 258:463-476. Azizi, S.A., R.A. Burne, and D.J. Woodward (1985) The auditory corticopontocerebellar projection in the rat: Inputs to the paraflocculus and midvermis. An anatomical and physiological study. Exp. Brain Res. 59r36-49. Beyerl, B.D. (1978)Afferent projections to the central nucleus of the inferior colliculus in the rat. Brain Res. 145t209-223. Birt, D., R. Nienhuis, and J. Olds (1978) Effects of bilateral auditory cortex ablation on behavior and unit activity in rat inferior colliculus during differential conditioning. J. Neurophysiol. 41t705-715. Bjoerkeland, M., and J. Boivie (1984) An anatomical study of the projections from the dorsal column nuclei to the midbrain in cat. Anat. Embryol. 170:29-43. Casseday, J.H., D.R. Jones, and I.T. Diamond (1979)Projections from cortex to tectum in the tree shrew, Tupaiaglis. J. Comp. Neurol. 185.253-292. Caviness, V.S. (1975)Architectonic map of neocortex of the normal mouse. J. Comp. Neurol. 164247-264. Caviness, V.S., and D.O. Frost (1980) Tangential organization of thalamic projections to the neocortex in the mouse. J. Comp. Neurol. 194:335367. Cipolloni, P.B., and A. Peters (1979) The bilaminar and banded distribution of the callosal terminals in the posterior neocortex of the rat. Brain Res. 176:3347. Clopton, B.M., and J.A. Winfield (1973) Tonotopic organization in the inferior colliculus of the rat. Brain Res. 56:355-358. Coleman, J.R., and W.J. Clerici (1987) Source of projections to subdivisions of the inferior colliculus in the rat. J. Comp. Neurol. 262.215-226. Conlee, J.W., and E.S. Kane (1982) Descending projections from the inferior colliculus to the dorsal cochlear nucleus in the cat: An autoradiographic study. Neuroscience 7:161-178. Cooper, M.H., and P.A. Young (1976) Cortical projections to the inferior colliculus of the cat. Exp. Neurol. 51:488-502. Diamond, I.T., E.G. Jones, and T.P.S. Powell (1969) The projection of the auditory cortex upon the diencephalon and brainstem in the cat. Brain Res. 15:305-340. Druga, R., and J. Syka (1984a) Neocortical projections to the inferior colliculus in the rat. An experimental study using anterograde degeneration techniques. Physiol. Bohem. 332251-253. Druga, R., and J. Syka (1984b) Ascending and descending projections to the inferior colliculus in the rat. Physiol. Bohem. 33t31-42. Druga, R., J.Syka, and G. Rajkowska-Markow (1988) Localization of cortical neurons projecting to the inferior colliculus in the rat and guinea pig. In J. Syka and R.B. Masterton (eds): Auditory Pathway. Structure and Function. New York and London: Plenum Press, pp. 293-298. Faye-Lund, H. (1985) The neocortical projection to the inferior colliculus in the albino rat. Anat. Embryol. 173:53-70. Faye-Lund, H. (1986) Projections from the inferior colliculus to the superior olivary complex in the albino rat. Anat. Embryol. 17535-52. Faye-Lund, H., and Osen, K.K. (1985) Anatomy of the inferior colliculus in the rat. Anat. Embryol. 171:l-20. Fitzpatrick, K.A., and T.J. Imig (1978) Projections of auditory cortex upon

121 the thalamus and midbrain in the owl monkey. J. Comp. Neurol. 177:537-556. Forbes, B.F., and N. Moskowitz (1974) Projections of auditory responsive cortexin the squirrel monkey. Brain Res. 67939-254. Games, K.D., and J.A. Winer (1988) Layer V in rat auditory cortex: Projections to the inferior colliculus and contralateral cortex. Hearing Res. 34:l-26. Gonzales-Lima, F., and H. Scheich (1984) Functional activation in the auditory system of the rat produced by arousing reticular stimulation: A 2-deoxyglucosestudy. Brain Res. 299t201-214. Hashikawa, T. (1983) The inferior colliculopontine neurons of the cat in relation to other collicular descending neurons. J. Comp. Neurol. 219t241-249. Henkel, C.K., and K.M. Spangler (1983) Organization of the efferent projections of the medial superior olivary nucleus in the cat as revealed by HRP and autoradiographic tracing methods. J. Comp. Neurol. 221:416-428. Herbert, H. (1984) Descending auditory system: Projections from auditory cortex to inferior colliculus in the rat. Neurosci. Lett. Suppl. 18:244. Huang, C., and J. Fex (1986)Tonotopic organization in the inferior colliculus of the rat demonstrated with the 2-deoxyglucose method. Exp. Brain Res. 61:506-512. Itoh, K., T. Kaneko, M. Kudo, and N. Mizuno (1984) The intercollicular region in the c a t A possible relay i n the parallel somatosensorypathways from the dorsal column nuclei to the posterior complex of the thalamus. Brain Res. 308:166-171. Irvine, D.R.F. (1986) The auditory brainstem. A review of the structure and function of auditory brainstem processing mechanisms. I n D. Ottoson (ed): Progress in Sensory Physiology 7. Berlin: Springer-Verlag, pp 1-279. Irvine, D.R.F., and D.P. Phillips (1982) Polysensory association areas of the cerebral cortex. Organization of acoustic input in the cat. In C.N. Woolsey (ed):Cortical Sensory Organization, Vol. 3, Multiple Auditory Areas. Clifton, New Jersey: Humana Press, pp. 111-156. Jones, E.G., and R.Y. Leavitt (1974) Retrograde axonal transport and the demonstration of non-specific projections to the cerebral cortex and striatum from thalamic intralaminar nuclei in the rat, cat and monkey. J. Comp. Neurol. 154:349-378. Kelly, J.P., and D. Wong (1981) Laminar connections of the cat's auditory cortex. Brain Res. 212:l-15. Krieg, W.J.S. (1946a)Connections of the cerebral cortex. I. The albino rat. A) Topography of the cortical areas. J. Comp. Neurol. 84.221-275. Krieg, W.J.S. (1946b)Connections of the cerebral cortex. I. The albino rat. B) Structure of the cortical areas. J. Comp. Neurol. 84.277-323. Krieg, W.J.S. (1947) Connections of the cerebral cortex. I. The albino rat. C) Extrinsic connections. J. Comp. Neurol. 86.967-394. Kusama, T., K. Otani, and E. Kawana (1966) Projections of the motor, somatic sensory, auditory, and visual cortices in cats. In T. Tokizane and J.P. Schade (eds): Progress in Brain Research, Vol. 21A, Correlative Neurosciences, Part A, Fundamental Mechanisms. Amsterdam: Elsevier, pp. 292-322. Kuypers, H.G.J.M., and D.G. Lawrence (1967)Cortical projections to the red nucleus and the brain stem in the rhesus monkey. Brain Res. 4:151-188. Land, P.W., L.L. Rose, A.R. Harvey, and S.A. Liverman (1984) Neonatal auditory cortex lesions result in aberrant crossed cortico-tectal and cortico-thalamic projections in rats. Dev. Brain Res. 12:126-130. LeDow, J.E., D.A. Ruggiero, and D J . Reis (1985) Projections to the subcortical forebrain from anatomically defined regions of the medial geniculate body in the rat. J. Comp. Neurol. 242:182-213. LeDoux, J.E., C. Farb, and D.A. Ruggiero (1990a) Topographic organization of neurons in the acoustic thalamus that project to the amygdala. J. Neurosci. 10:1043-1054. LeDow, J.E., P. Ciccetti, A. Xagoraris, and L.M. Romansky (1990b) The lateral amygdaloid nucleus: Sensory interface of the amygdala in fear conditioning. J. Neurosci. 10: 1062-1069. LeMessurier, D.H. (1948) Auditory and visual areas of the cerebral cortex of therat. Fed. Proc. 7:70. Martin, G.F., and D. Megirian (1972) Corticobulbar projections of the marsupial phalanger (Trichosurus uulpecula). 11. Projections to the mesencephalon. J. Comp. Neurol. 144:165-192. Martin, G.F., J.C. Bresnahan, C.K. Henkel, and D. Megirian (1975)Corticobulbar fibers in the North American opossum (Didelphis marsupidis

H. HERBERT ET AL.

122

Schober, W. (1986) The rat cortex in stereotaxic coordinates. J. Hirnforsch. 27:121-143. Semple, M.N., and L.M. Aitkin (1979) Representation of sound frequency and laterality by units in central nucleus of cat inferior colliculus. J. Neurophysiol. 42:1626-1639. Shneiderman, A,, and C.K. Henkel(1987) Banding of lateral superior olivary nucleus afferents in the inferior colliculus: A possible substrate for sensory integration. J. Comp. Neurol. 266519-534. Shneiderman, A,, D.L. Oliver, and C.K. Henkel (1988) Connections of the dorsal nucleus of the lateral lemniscus: An inhibitory parallel pathway in the ascending auditory system? J. Comp. Neurol. 276:188-208. Spangler, K.M., W.B. Warr, and C.K. Henkel (1985) The projection of 202. principal cells of the medial nucleus of the trapezoid body in the cat. J. Mitani, A,, M. Shimokouchi, and S. Nomura (1983) Effects of stimulation of Comp. Neurol. 238249-262. the primary auditory cortex upon colliculo-geniculate neurons in the Stiebler, I., and Ehret, G. (1985) Inferior colliculus of the house mouse. t. A inferior colliculus of the cat. Neurosci. Lett. 42:185-189. quantitative study of tonotopic organization, frequency representation, Morest, D.K., and D.L. Oliver (1984) The neuronal architecture of the and tone-threshold distribution. J. Comp. Neurol. 238:65-76. inferior colliculus in the cat: Defining the functional anatomy of the Sun, X., P.H.-S. Jen, D. Sun, and S. Zhang (1989) Corticofugal influences on auditory midbrain. J. Comp. Neurol. 222:209-236. the responses of bat inferior collicular neurons to sound stimulation. Noort, J.V. (1969) The structure and connections of the inferior colliculus. Brain Res. 495:l-8. An investigation of the lower auditory system. Proefschrift. Leiden: Van Syka, J., and J. Popelar (1984) Inferior colliculus in the rat: Neuronal Gorcurn and Cornp., pp. 1-116. responses to stimulation of auditory cortex. Neurosci. Lett. 51935-240. Oliver, D.L. (1984) Dorsal cochlear nucleus projections to the inferior Syka, J., R. Druga, J. Popelar, and 8. Kalinova (1980) Functional organizacolliculus in the cat: A light and electron microscopic study. J. Comp. tion of the inferior colliculus. In J. Syka and L.M. Aitkin (eds): Neuronal Neurol. 224:155-172. Mechanisms of Hearing. New York and London: Plenum Press, pp. Oliver, D.L. (1987) Projections to the inferior colliculus from the anteroven137-153. tral cochlear nucleus in the cat: Possible substrates for binaural interaction. J. Comp. Neurol. 264:24-46. Syka, J., J. Popelar, R. Druga, and A. Vlkova (1988) Descending central auditory pathway-structure and function. In J. Syka and R.B. MasterOliver, D.L., and W.C. Hall (1978) The medial geniculate body of the tree ton (eds): Auditory Pathway. Structure and Function. New York and shrew, Tupaiaglis. 11. Connections with the neocortex. J. Comp. Neurol. London: Plenum Press, pp. 279-292. 182:459-494. Thompson, W.H. (1900) Degenerations resulting from lesions of the cortex Oliver, D.L., and D.K. Morest (1984) The central nucleus of the inferior colliculus in the cat. J. Comp. Neurol. 222.237-264, of the temporal lobe. J. Anat. Physiol. 35147-165. Patterson, H.A. (1977) An anterograde degeneration and retrograde axonal Vaughan, D.W. (1983) Thalamic and callosal connections of the rat auditory transport study of the cortical projections of the rat medial geniculate cortex. Brain Res. 260:181-189. body. Thesis, Boston University. Warr, W.C., J.J. Guinan, and J.S. White (1986) Organization of the efferent Paxinos, G., and C. Watson (1986) The Rat Brain in Stereotaxic Coordinates. fibers: The lateral and medial olivocochlear systems. In R.A. Altschuler, San Diego: Academic Press. D.W. Hoffman, and R.P. Bobbin (eds): Neurobiology of Hearing: The Robards, M.J. (1979) Somatic neurons in the hrainstem and neocortex Cochlea. New York: Raven Press, pp. 333-348. projecting to the external nucleus of the inferior colliculus: An anatomiWatanabe, T., K. Yanagiasawa, J. Kanazaki, and Y. Katsuki (1966) Cortical cal study in the opossum. J. Comp. Neurol. 184:547-566. efferent flow influencing unit responses of medial geniculate body to Robards, M.J., D.W. Watkins, and R.B. Masterton (1976) An anatomical sound stimulation. Exp. Brain Res. 2:302-317. study of some somesthetic d e r e n t s to the intercollicular terminal zone Wiederhold, M.L. (1986) Physiology of the olivocochlear system. In R.A. of the midbrain of the opossum. J. Comp. Neurol. 170:499-524. Altschuler, D.W. Hoffman, and R.P. Bobbin (eds): Neurobiology of Rockel, A.J., and E.G. Jones (1973a) The neuronal organization of the Hearing: The Cochlea. New York Raven Press, pp. 349-370. inferior colliculus of the adult cat. I. The central nucleus. J. Comp. Wiherg, M., and A. Blomqvist (1984) The projection to the mesencephalon Neurol. 147:11-60. from the dorsal column nuclei. An anatomical study in the cat. Brain Rockel, A.J., and E.G. Jones (1973b) The neuronal organization of the Res .311:225-244. inferior colliculus of the adult cat. 11. The pericentral nucleus. J. Comp. Willard, F.H., and D.K. Ryugo (1983) Anatomy of the central auditory Neurol. 149:301-334. system. In J.F. Willot (ed): The Auditory Psychobiology of the Mouse. Rockel, A.J., and E.G. Jones (1973~)Observations on the fine structure of Springfield: Thomas, pp. 201-304. the central nucleus of the inferior colliculus of the cat. J. Comp. Neurol. Willard, F.H., and G.F. Martin (1983) The auditory brainstem nuclei and 14751-92. some of their projections to the inferior colliculus in the North American Roger, M., and P. Arnault (1989) Anatomical study of the connections of the opossum. Neuroscience 10: 1203-1232. primary auditory area in the rat. J. Comp. Neurol. 287:339-356. Willard, F.H., and G.F. Martin (1984) Collateral innervation of the inferior Riittgers, K., A. Aschoff, and E. Friauf (1990) Commissural connections colliculus in the North American opossum: A study using fluorescent between the auditory corticos of the rat. Brain Res. 509:71-79. markers in a double labeling paradigm. Brain Res. 303rl71-182. Ryan, A.F., Z. Furlow, N.K. Woolf, and E.M. Keithley (1988) The spatial WongRiley, M.T.T. (1979) Changes in the visual system of monocularly representation of frequency in the rat dorsal cochlear nucleus and sutured or enucleated rats demonstrated with cytochrome oxidase inferior colliculus. Hearing Res. 36:181-190. histochemistry. Brain Res. 171.11-28. Ryugo, D.K. (1976) An attempt towards an integration of structure and function in the auditory system. Thesis, University of California, Irvine. Zilles, K. (1985) The Cortex of the Rat. A Stereotaxic Atlas. Berlin: Springer-Verlag. Sally, S.L., and J.B. Kelly (1988) Organization of auditory cortex in the albino rat: Sound frequency.J. Neurophysiol. 59: 1627-1638. Zilles, K., and A. Wree (1985) Cortex: Areal and laminar structure In G . Paxinos (4):The Rat Nervous System. Vol. 1: Forebrain and Midbrain. Scheel, M. (1988) Topographic organization of the auditory thalamocortical Sydney: Academic Press, pp. 375-415. system in the albino rat. Anat. Embryol. 179:181-190. K., B. Zilles, and A. Schleicher (1980) A quantitative approach to Scherer-Sineler. U.. S.R. Vincent. H. Kimura. and E.G. McGeer (1983) . ~ Zilles, _ Demonstration 'of a unique population of neurons with NADPHcytoarchitectonics. VI. The areal pattern of the cortex of the albino rat. diaphorase histochemistry. J. Neurosci. Methods 9:229-234. Anat. Embryol. 159:335-360.

uirginiana) with notes on the Tasmanian brush-tailed opossum (TrichoSUPUS uulpecula) and other marsupials. 3. Anat. 120:439-484. Mmsopust, L.C., and J.M. Ordy (1962) Auditory organization of the inferior colliculi in the cat. Exp. Neurol. 6:456-477. Merzenich, M.M., and M.D. Reidt (1974) Representation of the cochlea within the inferior colliculus in the cat. Brain Res. 77r397-415. Mesulam, M.M. (1982) Tracing Neural Connections With Horseradish Peroxidase. Chichester: John Wiley & Sons. Mettler, F.A. (1935) Corticofugal fiber connections of the cortex of Macaca mulatta. The temporal region. J. Comp. Neurol. 6325-47. Miller, M.W., and B.A. Vogt (1984) Direct connections of rat visual cortex with sensory, motor and association cortices. J. Comp. Neurol. 226: 1 8 6

~~