Auditory nerve fibers project into the CN following a stereotyped plan (Figs. 6.1 and 6.2; Ramón y Cajal, 1909; Lorente de Nó, 1933a). On crossing the Schwann–glia border, fibers ascend dorsally into the nucleus forming the root branch of the nerve. After traversing for some distance, the root branch bifurcates, giving rise to an ascending branch and a descending branch (collectively called “parent” branches). The ascending branch projects into the anterior region of the AVCN, where one of its most pronounced terminals is a large, axosomatic ending known as the endbulb of Held (Held, 1893). The descending branch is directed posteriorly through the PVCN and usually (85 % of cases in cats) continues onward and upward into the DCN (Fekete et al., 1984). Nodes of Ranvier can be observed at regular intervals along the parent branches, and the mean lengths of ascending and descending branches are similar irrespective of CF or SR (Fekete et al., 1984).

Each parent branch maintains a relatively straight trajectory and can give rise to short collaterals and multiple endings as they traverse through the nucleus. In the cat, the ascending branch produces an average of nine primary collaterals, and the descending branch gives rise to an average of 11 primary collaterals (Fekete et al., 1984). These collaterals are thinner, divide several times, exhibit en passant and terminal swellings, and generally do not extend far from the parent branch. En passant swellings do not consistently display synapses, but terminal swellings always do (Fekete et al., 1984; Ryugo & Sento, 1991). Collaterals are scattered along the length of each parent branch, resulting in a patchy terminal field for any given auditory nerve fiber. In the cat, approximately 15 % of terminal swellings reside in the DCN, with the remainder scattered throughout the VCN (Ryugo & May, 1993). Fibers sharing similar CFs (see Sect. 6.3.2.2) occupy the same region, thus filling in the patches to form a solid terminal field along the length of the projection (Fekete et al., 1984; Ryugo & May, 1993). All terminals, however, remain within the magnocellular core of the VCN and the deep layer (III) of the DCN (Fekete et al., 1984; Liberman, 1991).

Topographic maps are one of the fundamental principles of brain organization, defined by an orderly representation of sensory and motor systems (Kaas, 1997). In the auditory system, frequency maps are conserved throughout the ascending pathways, replicating the place-frequency map established at the cochlear sensory epithelium (Clopton et al., 1974). Tonotopy is thus imprinted upon the CN via the precise distribution of auditory nerve fibers relative to CF (Fekete et al., 1984; Ryugo & May, 1993). This organization can also be referred to as cochleotopy, which describes the organization of projections with respect to their position of origin on the cochlear spiral. Because of the orderly organization of the cochlear place-frequency map (Liberman, 1982b), the concepts of tonotopy and cochleotopy are often interchangeable. This correspondence is useful for interpreting anatomical studies from which physiological data may not be available.

An electrophysiological study first established that a complete representation of audible frequencies could be found within the CN (Rose et al., 1959). Electrode penetrations along a dorsal to ventral axis revealed frequency tunings that progressed from high to low CFs. The range of CFs encountered on a particular electrode track varied with location, but similar trends were observed in both dorsal and ventral divisions. Subsequent physiological experiments have expanded upon this view, establishing that frequency tuning is highly organized within the DCN, AVCN, and PVCN (Bourk et al., 1981; Spirou et al., 1993). Spatial reconstructions of electrode penetrations in these studies were sufficiently detailed to support the conclusion that frequency representations were conceptually organized into isofrequency laminae.

Exploration of the cochleotopic projection pattern of the auditory nerve into the CN began with the histologically stained material of early neuroanatomists (Ramón y Cajal, 1909; Poljak, 1927; Lorente de Nó, 1933a). Initial experimental studies verified a rough correspondence between a lesion point in the cochlea and the position of the silver grains attracted by degenerating axons to identify the bifurcation of the root branch and the distribution of auditory nerve fiber projections (Sando, 1965; Osen, 1970; Webster, 1971). Apical parts of the cochlea, representing low CFs, projected to ventral regions of the CN, whereas basal, high-CF portions projected to dorsal areas. More direct results were obtained following small extracellular tracer deposits made in the cochlea that filled the axons and their terminations in the CN (Leake & Snyder, 1989; Brown & Ledwith, 1990). Another study involved small injections into defined frequency locations of the CN that backfilled auditory nerve fibers within the nucleus as well as peripheral terminations in the cochlea (Muniak et al., 2013). When analyzed in the coronal plane, these injections produced “slabs” of labeled fibers that coursed throughout the rostro-caudal extent of the nucleus. The dorsoventral position of these slabs was cochleotopic, giving an anatomical substrate to the isofrequency laminae observed electrophysiologically.

Further refinement of these findings was obtained through the reconstruction of physiologically characterized and intracellularly labeled auditory nerve fibers in the cat (Fekete et al., 1984; Liberman, 1991; Ryugo & May, 1993). Single-fiber studies not only provide excellent resolution for analyzing the fine details of primary projections but also yield unambiguous structure–function relationships. These results provided unequivocal evidence that the position of bifurcation of the root branch and the spatial distribution of the central axon is strongly correlated with the CF of the fiber (Fig. 6.2). Fibers with low CFs bifurcate almost immediately on entering the CN, with the ascending and descending branches distributed ventrally in the VCN and DCN. Fibers with progressively higher CFs bifurcate and distribute at progressively more dorsal locations.

One consequence of this organization is that the overall length of each auditory nerve fiber remains approximately constant across CFs (Arnesen & Osen, 1978; Fekete et al., 1984). Although apical fibers must traverse a greater distance from their entrance in the spiral lamina to the Schwann–glia border, this length is offset by a shorter distance to the point of bifurcation. In the cat, the total length of the ascending and descending branches after the point of bifurcation is 6–7 mm; the length of the intracochlear axon from the bifurcation to the habenula perforata is also about 6–7 mm. Within the DCN, auditory nerve terminal fields form an anisotropic spatial envelope, with the long axis oriented perpendicular to the ependymal surface and short axis confined to an isofrequency lamina (Ryugo & May, 1993). The orientation of the terminal field shifts systematically with CF, corresponding to the gradual curvature of the DCN.

Owing to technical constraints, most studies of fiber projections have historically analyzed and/or presented data using 2D plots, typically along a parasagittal plane to maximize the information yield across all subdivisions (Fekete et al., 1984; Ryugo & May, 1993). A limitation of such analyses is the difficulty in capturing the undulating variations of isofrequency laminae within all three dimensions. Recently, a quantitative 3D model of frequency representation in the CN was developed for the mouse (Fig. 6.3) that confirmed and extended the 2D projection pattern previously described (Fekete et al., 1984; Ryugo & May, 1993; Muniak et al., 2013). Visualizations of the helical twist of the nerve root and the angular projection of fibers transitioning from PVCN to DCN were also possible (Sando, 1965; Arnesen & Osen, 1978). At the base of the nerve root, low-CF fibers spiral around the outer edge of the nerve bundle along a medial-to-rostral trajectory before bifurcating laterally. Higher-CF fibers enter the nerve root more centrally within the bundle, ultimately bifurcating both at more dorsal and more medial locations than those of lower CFs. Such reconstructions demonstrate that, although the dorsal-ventral axis may be the predominant orientation of fiber distribution with respect to CF, there is also a strong medial-lateral bias that must be considered (Fig. 6.4). The stereotyped primary projection to the CN may be described as forming “C”-shaped laminae that stack vertically upon one another with increasing CFs (Fig. 6.3).

The precise tonotopic projection pattern of auditory nerve fibers provides a solid grounding for observations of orderly frequency tuning in the CN (Bourk et al., 1981; Spirou et al., 1993). Some regions of the nucleus, however, exhibit tuning variations that are explainable on the basis of local perturbations in afferent input. In the first instance, fine-scale frequency organization was described as irregular in the nerve root (Bourk et al., 1981), with high-CF units encountered in locations generally corresponding to low CFs. This circumstance is created by the presence of collaterals arising from the root branch prior to the bifurcation (Fekete et al., 1984). The numbers and lengths of these collaterals can vary widely, but are usually confined to the nerve root region. Fibers with high CFs tend to emit more numerous and extensive collaterals, which can intermingle with low-CF fibers and complicate frequency tuning. Octopus cells, a unique cell type of the PVCN, respond strongly to the onset of sounds or stimulus transients and represent another instance of unconventional tuning (Godfrey et al., 1975; Rhode & Smith, 1986). These cells show sharp tuning near threshold, but exhibit broader tuning 20 dB above threshold (Godfrey et al., 1975). The prominent dendrites of octopus cells intersect a broad swath of descending branch fibers (Osen, 1969), providing a likely substrate for broad tuning. On the other hand, it was noted that descending fibers with CFs greater than 4 kHz emit long collaterals oriented orthogonal to the parent branch, but parallel to octopus cell dendrites (Fekete et al., 1984). The combination of a dendrite intersecting a wide CF swath of auditory nerve terminals and collaterals from a single fiber converging on the same dendritic branch could contribute to narrow tuning near threshold and broad tuning at higher levels.

There has been some question as to whether all frequencies are equally represented in each subdivision of the CN. In the mouse (Muniak et al., 2013) and gerbil (Müller, 1990), reconstructions of tissue volume with respect to CF suggest that proportional volumes of each subdivision are roughly dedicated to equivalent octave ranges (Fig. 6.4). In cats, however, electrode penetrations (Spirou et al., 1993) and fiber reconstructions (Fekete et al., 1984; Ryugo & May, 1993) suggest that greater amounts of tissue in the DCN may be dedicated to mid–high frequencies relative to that in the VCN. Complicating this interpretation, however, is that these data were analyzed with respect to linear distance, rather than tissue volume, which appears not to be uniform (i.e., narrowing cross sections) along the frequency axis. Interestingly, the bat appears to have an expanded representation of the echolocating frequency range (Feng & Vater, 1985). This expansion, however, reflects a similar “fovea” in its cochlear representation (Vater et al., 1985), suggesting that the presence (or absence) of cochlear specializations may be equivalently represented through the projection pattern of auditory nerve fibers.

Hearing involves more than simple analysis of the frequency spectrum. Accordingly, although the predominant organizational feature of auditory nerve projections into the CN is tonotopy, the central projections of type I fibers also exhibit variations with respect to SR, echoing observations in the periphery (Fig. 6.5). Fibers belonging to different SR classes have distinct physiological characteristics, particularly with respect to encoding dynamic range and representing vowel sounds (Sachs & Young, 1979; Evans & Palmer, 1980; May et al., 1996). High-SR fibers with low thresholds may operate best in quiet, whereas low-SR, high-threshold fibers may be optimized for loud and noisy environments. Different groupings of fibers likely contribute to separate roles in acoustic processing and so are expected to differ in their central terminations.