This report describes 17 fusiform cells obtained as part of an in vivo intracellular recording and labeling survey of the DCN in barbiturate-anesthetized Mongolian gerbils. Previous reports of the acoustic response properties of identified fusiform cells are limited to one set of studies in anesthetized cats (Rhode et al. 1983; Rhode and Smith 1986; Smith and Rhode 1985) and one study in decerebrate gerbils (Ding et al. 1999). The present study forms a bridge between the previous two in that it combines the anesthetic state of the former and the animal model of the latter. Understanding this relationship is important since DCN physiology differs significantly between decerebrate cats and gerbils (Davis et al. 1996), and between the barbiturate-anesthetized and unanesthetized decerebrate preparations (Evans and Nelson 1973; Fan 2000; Gdowski and Voigt 1997).
Fusiform cell physiology: comparison between cat and gerbil
The acoustic response properties presented here for fusiform cells in the barbiturate-anesthetized gerbil resemble those described in the barbiturate-anesthetized cat (Rhode et al. 1983; Rhode and Smith 1986; Smith and Rhode 1985). Fusiform cells in that preparation were mostly of the pauser or buildup type, although responses were shown to depend on stimulus frequency and intensity. Similar results were obtained in this study.
Another feature of fusiform cell physiology common to the cat and the gerbil is the presence of sustained depolarizations and afterhyperpolarizations of the membrane in response to acoustic stimulation. Fusiform cells in both species exhibit depolarizations about 10 mV in amplitude during the stimulus and long-lasting afterhyperpolarizations of similar magnitudes. Afterhyperpolarizations are often of sufficient magnitude and duration to eliminate spontaneous activity.
Some differences were apparent, but it was not clear whether these were interspecies differences or whether they resulted from methodological differences. Pauser and buildup responses predominated in the cat, but a quarter of the cells in our study (4/16) were classified as chopper units. Furthermore, most of the pausers were subclassified as pauchoppers (6/9), indicating a relatively brief pause followed by a sustained period of regular firing. The responses of DCN neurons are known to be sensitive to interstimulus interval, likely as an effect of the long-lasting afterhyperpolarization (Rhode and Smith 1986). The longer interstimulus interval in the present study (1 s vs. 100 ms) probably allowed greater recovery from the hyperpolarization, resulting in response patterns with more chopper-like character.
The similarity in DCN physiology between the cat and gerbil in the barbiturate-anesthetized preparation contrasts with results from the decerebrate preparation. Specifically, Davis et al. (1996) found that the incidence of type IV units, which show strong on-BF inhibition, was much lower in the gerbil than in the cat (11% vs. 32-45%). Antidromic stimulation studies indicate that at least some fusiform cells in the decerebrate cat are type IV units (Young 1980). In contrast, an intracellular recording and labeling study in the decerebrate gerbil found that 12 of 13 identified fusiform cells were type III units (Ding et al. 1999), which have V-shaped excitatory receptive fields flanked by sideband inhibition. The one fusiform cell in that study with type IV unit properties was actually a type IV-T unit, which exhibits relatively weak on-BF inhibition. Thus an interspecies difference in the balance of excitation and inhibition appears to be masked through the suppression of inhibitory effects by barbiturate anesthesia (Evans and Nelson 1973; Fan 2000; Young and Brownell 1976).
Quantification of discharge patterns
Discharge patterns were quantified using a two-line fit to the ISI plot (Fig. 9). The first bin, representing 5 ms of data, was excluded so that, as described in RESULTS, the value of m1 did a better job of quantifying the rate trend over the first 50-75 ms of the response than did the transient slope mT used in the classification scheme of Gdowski (1995). Also, for the data in this study, the value of m1 appears to be a more useful measure of inhibition than the normalized tone slope, because its values are more evenly distributed over a wider interval (Fig. 10). Normalized tone slope was devised as a useful metric for distinguishing type II units from type III units in the decerebrate cat (Young and Voigt 1982) and appears not to be as well suited for identifying trends within the cell population in this study.
The slope m1 as described here has not been used in the past to analyze PSTH data. Previous studies have used the transient slope as one means of dividing data into discrete categories (Blackburn and Sachs 1989; Gdowski 1995; Parham and Kim 1992; Young et al. 1988). In this study it is known that all of the PSTHs, regardless of shape, arise from the same neuron type. The focus of the present analysis was to quantify regularity histogram shape using a continuous measure within a small, homogeneous neuron population, rather than to identify discrete categories within a large, heterogeneous population. The interpretation of m1 as a measure of inhibition is reasonable in this study because the cells are anatomically similar and presumably receive similar sets of synaptic inputs.
Fusiform cells likely receive input from vertical cells (Saint-Marie et al. 1991; Voigt and Young 1980, 1990; Zhang and Oertel 1994), which discharge in a pattern characterized by transiently decreasing interspike intervals (Rhode 1999 and our own unpublished observations). They are also thought to receive input from a source of wideband inhibition (Nelken and Young 1994; Spirou and Young 1991), possibly arising from stellate cells in the PVCN that discharge with an onset-chopper pattern (Oertel et al. 1990; Smith and Rhode 1989). Since the fusiform cell PSTHs presented here were derived from responses to BF tones, it is more likely that vertical cells are responsible for the observed inhibitory effects, since the onset-choppers of the PVCN typically respond better to broadband noise than to tones (Winter and Palmer 1995). Whatever the source, the results suggest that the strength of the inhibitory input varies significantly across the fusiform cell population.
Responses to pulses of injected current
Individual cells were capable of discharging in pauser, buildup, and chopper patterns, depending on current pulse amplitude and prior hyperpolarization (Fig. 8). These results suggest that the fusiform cells of the adult gerbil express an inactivating K+ conductance similar to that presumed to influence discharge patterns of fusiform cells in vitro (Kanold and Manis 1999; Manis 1990). Furthermore, the effect of these currents can be observed in vivo despite spontaneous synaptic bombardment arising from the parallel fiber network and the peripheral auditory system.
About one-half of the fusiform cells tested in this study showed the voltage-dependent effect as opposed to 90-95% of the fusiform cells in the guinea pig slice (P. B. Manis, personal communication). Input resistance was a limiting factor in producing appreciable latency shifts, and the input resistances in this study were smaller than those in the guinea pig slice (17.9 ± 9.7 M vs. 27.0 ± 16.6 M) (Manis 1990). The presence of spontaneously active input in our preparation may account for this difference, since such input reduces cell input resistance (Destexhe and Paré 1999; Rapp et al. 1992).
Fusiform cells exhibit a variety of response properties
The data in Table 2 show that the fusiform cells in this study exhibited a wide variety of response properties. The resting potentials varied from 76 to 50 mV. The input resistances spanned more than an order of magnitude, from 3 to 37 M. Spontaneous activity ranged from 0 to more than 50 spikes/s, relative noise index from 0 (no noise response) to greater than unity (responds better to noise than to tones), and the normalized slopes of the BF rate-level curves vary from essentially flat to steeply nonmonotonic.
Furthermore, there is no obvious correlation of the different response measures. For example, among the cells with low spontaneous activity, there are those with small (e.g., cells 97002, Fig. 5 and 97073, Fig. 1) and those with large (e.g., cell 97022, Fig. 2) input resistances. Similarly, there are pauser units showing little evidence of inhibition (e.g., cell 98001, Fig. 4) and pausers with pronounced inhibitory responses (e.g., cell 97022, Fig. 2).
What underlies this diversity? A large body of evidence indicates that DCN principal cell responses are shaped by a relatively complex network of acoustically driven inputs. These inputs include excitatory drive from the auditory nerve (Smith and Rhode 1985), narrowband inhibition from vertical cells (Saint-Marie et al. 1991; Voigt and Young 1980, 1990), and wideband inhibition (Nelken and Young 1994; Spirou et al. 1993), possibly originating in the PVCN (Oertel et al. 1990; Smith and Rhode 1989; Zhang and Oertel 1994). In addition, the apical dendrites of fusiform cells are the targets of a second neural network driven by nonauditory inputs (Golding and Oertel 1997; Itoh et al. 1987; Kevetter and Perachio 1989; Mugnaini et al. 1980a; Weedman and Ryugo 1996; Weinberg and Rustioni 1987; Wright and Ryugo 1996). The observed diversity of physiological behavior likely results from cell-to-cell variations in the balance of activity within and across these networks. The techniques used in this study do not allow us to gauge directly this balance of activity, but to the extent that the interaction of a particular cell with these networks is determined by the details of its morphology, the anatomical data may provide some insight into the physiological diversity. A detailed quantitative comparison of fusiform cell physiology and morphology is the subject of the companion paper.
We thank committee members H. S. Colburn, L. H. Carney, M. C. Liberman, and A. M. Berglund for helpful comments and suggestions. We also thank D. Oertel for helpful suggestions and P. Patterson for processing some of the tissue.
K. E. Hancock was supported by a fellowship from The Whitaker Foundation. This work was funded by Grant DC-01099 from the National Institute on Deafness and Other Communication Disorders.
Present address of K. E. Hancock: Eaton-Peabody Laboratory, Massachusetts Eye and Ear Infirmary, 243 Charles St., Boston, MA 02114.
Address for reprint requests: H. F. Voigt, Dept. of Biomedical Engineering, 44 Cummington St., Boston, MA 02215-2407 (E-mail: firstname.lastname@example.org).
Received 27 April 2001; accepted in final form 20 December 2001.