The auditory fields in the cortex of the guinea pig were investigated with microelectrode mapping techniques. Pure tones of varying frequencies and amplitudes were used as acoustic stimuli. Mainly, multiunit activity was recorded.A large tonotopic area is found in the anterior half of the auditory cortex. This area is named the anterior field (field A). Frequency tuning curves of multiunits in field A are generally narrow. Responses to tone stimuli are strong, and latencies are short. Low best frequencies are represented rostrally, high best frequencies caudally. The tonotopy is continuous and quite regular. Field A is narrow dorsally and becomes gradually broader ventrally. Correspondingly, the isofrequency lines slightly diverge from dorsal to ventral.Caudal to the first field, there is a second, smaller tonotopic area. It lies in the dorsal half of the posterior auditory cortex and is therefore named the dorsocaudal field (field DC). The frequency specificity of the cell clusters in this area is as strong as in field A, but the tonotopy is discontinuous: In the dorsal half of field DC, high best frequencies (16-32 kHz) are represented rostrally; the low frequencies (0.5-2.8 kHz) are represented immediately caudal to the high frequencies, while the intermediate frequencies are missing. Ventrally in field DC, the frequency representation is more complete. Except for this discontinuous map, we did not notice any differences between fields A and DC. A third tonotopic field was found rostral to field A. This field extends over a surface of less than 1 mm2 and was named the small field (field S). It contains a complete representation of the frequency range; high best frequencies are located rostrally, low frequencies caudally. The response latencies are slightly longer in field S than in fields A or DC, and the tuning curves are broader.A broad strip of nontonotopic cortex (auditory belt) surrounds fields A and DC caudally. We subdivided this area into the dorsocaudal and the ventrocaudal belt region. In both areas, tuning curves are often broad, and response latencies are longer than in the tonotopic cortex. In the dorsocaudal belt, most multiunits react with a phasic on-response to pure tones; in the ventrocaudal belt, tonic responses occur more frequently. Another nontonotopic region is located in the anterior auditory cortex, rostral to the tonotopic fields, and was therefore named the rostral belt. Tuning curves in this area are broad, latencies are short, and reponse thresholds are often high.In the discussion, the guinea pig is compared with other mammalian species. Species-specific features in the organization of the tonotopic cortex of the guinea pig are revealed.

We investigated the projection from the medial geniculate body (MG) to the tonotopic fields (the anterior field A, the dorsocaudal field DC, the small field S) and to the nontonotopic ventrocaudal belt in the auditory cortex of the guinea pig. The auditory fields were first delimited in electrophysiological experiments with microelectrode mapping techniques. Then, small quantities of horseradish peroxidase (HRP) and/or fluorescent retrograde tracers were injected into the sites of interest, and the thalamus was checked for labeled cells.The anterior field A receives its main thalamic input from the ventral nucleus of the MG (MGv). The projection is topographically organized. Roughly, the caudal part of the MGv innervates the rostral part of field A and vice versa. After injection of tracer into low or medium best-frequency sites in A, we also found a topographic gradient along the isofrequency contours: the dorsal (ventral) part of a cortical isofrequency strip receives d e r e n t s from the rostral (caudal) portions of the corresponding thalamic isofrequency band. However, it is not so obvious whether such a gradient exists also in the highfrequency part of the projection. A second, weaker projection to field A originates in a magnocellular nucleus that is situated caudomedially in the MG and was therefore named the caudomedial nucleus.The dorsocaudal field DC receives input from the same nuclei as the anterior field, but the location of the labeled cells in the MGv is different. This was demonstrated by injection of different tracers into sites with like best frequencies in fields A and DC, respectively. After injection of HRP into the 1-2-kHz isofrequency strip in field A and injection of Nuclear Yellow (NY) into the 1-2-kHz site in field DC, the labeled cells in the MGv form one continuous array that runs from caudal to rostral over the whole extent of the MGv. The anterior part of this array consists of NY-labeled cells; i.e., it projects to field DC. The caudal part is formed by HRP-labeled cells; i.e., it innervates field A. These findings indicate that there is only one continuous tonotopic map in the MGv. This map is split when projected onto the cortex so that two adjacent tonotopic fields (A and DC) result. The cortical maps are rotated relative to the thalamic map in that rostral portions of the MGv project to caudal parts of the tonotopic cortex and vice versa.The small field S receives its main thalamic input from a region situated in the rostral half of the MG medial to the ventraI nucleus (the rostromedial MG). After injection of tracer into the ventrocaudal belt, labeled cells were found dorsal, lateral, and ventral to the MGv. These cells form a continuous band that surrounds the MGv like a shell. A second population of labeled neurons was found in the caudomedial nucleus of the MG.Similarities and differences between the auditory thalamocortical systems in the guinea pig and other mammalian species are discussed.

The topography of the anatomical projection from isofrequency contours (IFCs) in auditory thalamus to IFCs in primary auditory cortex (field AI) was investigated in the cat. In each experiment, a best-frequency map of AI was obtained with electrophysiological recording techniques. Then, different retrogradely transported tracers (HRP, fluorescent dyes) were introduced into AI. In some experiments, different parts (e.g., dorsal, central, and ventral) of a previously mapped IFC were injected, each part with a different tracer. In other experiments, 2 or 3 rows of tracer injections were made at different dorsoventral levels of AI, over a large frequency range (5-38 kHz); each injection row was oriented orthogonal to the IFCs and contained a different tracer. The main mass of the labeled thalamic cells was found in the ventral nucleus of the medial geniculate body (MGv). The MGv cells projecting to a limited sector (1-2 mm in length in most experiments) of an IFC in AI form one or several densely packed neuron clusters of variable shape. The cells labeled by a given tracer are largely separated in space from cells labeled by a different tracer. Thus, different sectors of a cortical IFC receive input from different portions of the corresponding thalamic IFC. As a general rule, cells labeled from dorsal (ventral) injections are centered rostrally (caudally) in the part of MGv innervating AI. However, the topographic details are variable between individuals, and the rostrocaudal gradient is complicated by numerous irregularities and gradients. Previous studies of the auditory thalamocortical projection in the cat have not recognized the topographic order in the isofrequency dimension. Instead, it was believed that different sectors of a cortical IFC were innervated by coincident thalamic populations.

The auditory thalamus of the guinea pig was investigated with microelectrode mapping techniques. Pure tones of varying frequencies and amplitudes were used as acoustic stimuli, and frequency tuning curves were recorded from 840 multi-units or single cells. The neurons in ventral nucleus of the medial geniculate body (MGv) respond vigorously to pure tones; they have mostly narrow frequency tuning curves and short response latencies (8-12 ms). The MGv is tonotopically organized: High frequencies (16-21 kHz) are located rostrally; the intermediate frequencies (2.8-11 kHz) lie caudomedial of the high frequencies, while the low frequencies (0.5-2.8 kHz) run as a continuous band from rostrolateral to caudomedial. These data confirm a model of tonotopy of the guinea pig MGv which was based on anatomical data from previous tract-tracing experiments. In these experiments, thalamocortical connections were investigated with retrogradely transported tracers (horseradish peroxidase, fluorescent dyes, Redies et al. 1989b). Dorsal, lateral and in part also ventral to MGv, the neuronal responses to pure tones were often less vigorous than in MGv. Many neurons had broad frequency tuning curves, and in nearly all recordings from this region, the response latencies were longer than 12 ms. A tonotopic organization was not apparent here. From the response properties and the location relative to MGv, we concluded that this area corresponds to the shell nucleus of the MG.