| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
1Department of Otolaryngology, Head and Neck Surgery, University Hospital Würzburg, Würzburg, Germany; and 2Department of Otolaryngology, Head and Neck Surgery, University of California, San Francisco, California
Submitted 4 January 2007; accepted in final form 10 September 2007
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
45% of normal and in animals deafened >2.5 yr (long-deafened, LD cats) with severe cochlear pathology (mean SGC density <7% of normal). LD animals were subdivided into unstimulated cats and those that received chronic intracochlear electrical stimulation via a feline cochlear implant. Acutely deafened, implanted adult cats served as controls. Independent of their stimulation history, LD animals had significantly higher EABR and ICC thresholds than SDU and control animals. Moreover, the spread of electrical excitation was significantly broader and the dynamic range significantly reduced in LD animals. Despite the prolonged durations of deafness the fundamental cochleotopic organization was maintained in both the ICX and the ICC of LD animals. There was no difference between SDU and control cats in any of the response properties tested. These findings suggest that long-term auditory deprivation results in a significant and possibly irreversible degradation of response thresholds and spatial selectivity to intracochlear electrical stimulation in the auditory midbrain. |
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
) and with speech discrimination performance (Chatterjee and Shannon 1998; Henry et al. 2000; Throckmorton and Collins 1999; Zwolan et al. 1997).
Particularly poor speech discrimination performance is observed in congenitally and prelingually deaf CI users who are implanted as adults and in individuals who are implanted after long durations of deafness (Busby et al. 1991; Dawson et al. 1992; Ruben 1986). However, it is unclear at present to what extent these observations are due to peripheral pathology or to functional changes in the central auditory system. Although congenitally deafened individuals are likely to have relatively severe cochlear pathology after long periods of deafness, their speech discrimination performance shows some improvement with increasing auditory experience (Busby et al. 1991). These observations suggest that chronic electrical stimulation of the cochlea leads to functional alterations in the central auditory system even after long-term deafness.
Animal studies have shown that hearing loss induced during the early postnatal period results in progressive and more profound anatomical degeneration in the central auditory system (e.g., Lustig et al. 1994; Moore 1990; Nishiyama et al. 2000; Nordeen et al. 1983) and functional degradation or reorganization as compared with changes observed following auditory deprivation later in life (e.g., Hardie et al. 1998; Moore et al. 2002; Raggio and Schreiner 1999; Shepherd and Javel 1997; Shepherd et al. 1999; Silverman and Clopton 1977; Trune 1982). At the University of California San Francisco (UCSF), we have developed an animal model of neonatal or very early acquired profound hearing loss to evaluate the effects of early auditory deprivation and duration of deafness on signal processing in the central auditory system. Prior studies have demonstrated that neonatally induced long-term deafness in cats results in severe peripheral pathology with survival of <10% of spiral ganglion neurons (Rebscher et al. 2001; Vollmer et al. 2000, 2005). In addition, such long-term deafness results in degraded spatial selectivity of sinusoidal electrical stimulation (Rebscher et al. 2001; Vollmer et al. 2000) and in degraded temporal resolution of neurons in the auditory midbrain, specifically in the central nucleus of the inferior colliculus (Vollmer et al. 2005).
The objective of the present study was to assess in more detail the functional consequences of early-acquired bilateral auditory deprivation and duration of deafness on the processing and representation of electrical stimuli in the auditory system. Central neuronal responses to intracochlear electrical stimulation (ICES) were evaluated in neonatally deafened cats after different durations of deafness. Because the nature of the experiments does not allow a parametric series of experiments to study the effects of deafness duration, we selected two groups of animals to represent an intermediate and a prolonged duration of deafness with clearly different extents of spiral ganglion cell (SGC) degeneration. One group of animals (short-deafened unstimulated animals, SDU group) was studied after <1.5 yr of deafness (range: 6–14 mo; mean SGC survival 45% of normal); the second experimental group (long-deafened animals, LD group) was studied after prolonged durations of deafness of >2.5 yr (range: 30–86 mo; mean SGC survival <7% of normal). The deafness histories of the experimental animals are summarized in Table 1.
|
Previous studies have reported conflicting results with respect to changes in IC response thresholds following long-term deafness. Two studies observed increased thresholds in the IC of long-deafened animals (pulses: Shepherd and Javel 1997; sinusoids: Vollmer et al. 2000). In contrast, a third study reported that deafness duration and the degree of peripheral pathology had no effect on IC thresholds to sinusoidal stimulation (Rebscher et al. 2001). To clarify the conflicting results, the present study reports IC thresholds for both pulsatile and sinusoidal stimuli in a larger number of long-deafened animals, including animals reported previously by our lab (Rebscher et al. 2001; Vollmer et al. 2000).
Whether the normal cochleotopic organization of ICES is maintained after long durations of neonatal deafness is likewise disputed in the literature. Previous studies in normal hearing cats (e.g., Brown et al. 1997; Merzenich and Reid 1974; Oliver 1987; Oliver and Morest 1984; Rose et al. 1966), in neonatally deafened, unstimulated cats with deafness durations pooled over 0.5–2.6 yr, and in neonatally deafened, early stimulated cats (age at initial stimulation 7.5–18 wk, stimulation period 4–30 wk) (Snyder et al. 1990, 1991) demonstrated a cochleotopic frequency gradient that was systematically related to IC depth. In contrast, Shepherd and colleagues (1999) described only a rudimentary cochleotopic organization in neonatally deafened animals (duration of deafness: 12 mo) and in one animal that was deafened as a juvenile and studied after a long duration of deafness (7.8 yr). None of these earlier studies specifically explored the effect of long-term auditory deprivation on the cochleotopic organization of the IC in more than one long-deafened animal. To resolve this conflict, a principle focus of the present study is to identify and compare the cochleotopic organization of ICES in both the external (ICX) and the central nucleus (ICC) of the IC in a larger number of short- and long-deafened animals.
We previously reported spatial tuning width data of IC neurons in response to sinusoidal stimulation in several LD animals (Rebscher et al. 2001; Vollmer et al. 2000). The present study extends our previous work by examining a larger cohort of long-deafened animals, including a group examined after undergoing chronic electrical stimulation, by determining the spread of excitation in response to both sinusoidal and pulsatile electrical stimulation, and by examining the dynamic ranges of single neurons and multineuron clusters. The rationale for evaluating the responses to pulses were to examine the spatial selectivity of neurons in response to the pulsatile waveform that was used to evaluate temporal resolution in the same long-deafened animals reported in a previous study (Vollmer et al. 2005) and to investigate waveform-specific differences in thresholds and spatial distributions of responses in the IC for sinusoidal and pulsatile cochlear stimulation. Because pulsatile signals are used in most contemporary CI speech processing strategies, it is also of interest to evaluate spatial selectivity of neuronal responses to pulsatile signals.
A final goal of the current study was to study the effects of chronic electrical stimulation on threshold distributions and spatial selectivity in the adult auditory system following long-term congenital deafness. The LD animals in the present study were divided into two subgroups. The first group was studied acutely after prolonged deafness (long-deafened, unstimulated animals, LDU group), and the second group received a unilateral CI as adults and several weeks to months of electrical stimulation (long-deafened stimulated animals, LDS group) (Table 1). With the exception of one additional animal in the LDU group (K03), the individual animals in the subgroups and the electrical signals used for chronic stimulation of the LDS animals are identical to those reported in our previous study on temporal resolution in LD animals (Vollmer et al. 2005). The earlier study indicated that chronic ICES reversed the degradation in temporal resolution of electrical stimulation observed in unstimulated long-deafened animals. The findings from the prior study demonstrated that auditory experience can profoundly alter the functional status of the long-deafened, adult auditory system despite the extremely severe cochlear pathology observed in these animals and emphasized the potential of brain plasticity in the modulation of temporal resolution and coding of electrical signals in the central auditory system. The evaluation of both temporal and spatial data from the same animals provides a valuable basis to assess the potential and the limitations for modulatory effects elicited by chronic ICES on different signal processing strategies in the long-deafened central auditory system.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
This report includes results obtained from 5 neonatally short-deafened cats that were studied as adults after 
1.5 yr of deafness (range of deafness durations: 6–14 mo) and 12 long-deafened animals with durations of deafness exceeding 2.5 yr (range: 30–86 mo). All of these animals were deafened as newborns by systemic administration of neomycin sulfate (40–70 mg/kg im/sid) beginning on the day after birth and continuing for a total of 16–25 days (e.g., Leake et al. 1991). Neomycin injections were terminated when profound hearing loss (>105 dB) was confirmed by the absence of auditory brain stem responses to clicks (0.2 ms/ph, 20 pps). None of the animals demonstrated any residual hearing at the time of study. The animals with durations of deafness 1.5 yr did not receive any electrical cochlear stimulation until the time of the final acute electrophysiological experiment and will be referred to as short-deafened unstimulated (SDU) animals.
The long-deafened animals were maintained for periods ranging from 2.5 to 7.2 yr prior to study and were divided into two groups. Seven unstimulated cats received a unilateral CI as adults after these prolonged periods of deafness and were studied acutely (long-deafened unstimulated, LDU animals). Two of these LDU animals (K16 and K24) were implanted immediately before study; the others were implanted 
1 wk before the electrophysiological experiment to allow thresholds to stabilize. Five additional long-deafened cats were implanted as adults after prolonged periods of deafness (range: 42–84 mo) and received several weeks to months of chronic electrical stimulation prior to study (long-deafened stimulated, LDS animals). Electrical stimulation of the auditory nerve in these animals was initiated at ages ranging from 3.5 to 7.0 yr with an average age of 5.5 yr. Table 1 summarizes the deafness and chronic stimulation histories of the individual neonatally deafened animals included in the present study.
Fourteen acutely deafened adult cats with normal auditory experience prior to the experiment served as control cats. Controls were deafened 1.5–3.5 wk before study by intravenous administration of kanamycin and ethacrynic acid (Xu et al. 1993). It should be noted that spiral ganglion survival has been shown to be virtually normal 2 wk after deafness is induced by a very similar method (administration of kanamycin and aminooxyacetic acid) (Leake et al. 1987); thus spiral ganglion survival in this control group is assumed to be normal. About half of the control animals were implanted immediately before the experiment, the others were implanted 1–2 wk before study.
Prior to all surgical procedures, the animals were sedated (ketamine: 22–33 mg/kg; acepromazine maleate: 0.1 mg/kg; or inhaled isoflurane), and anesthesia was induced by pentobarbital sodium (7–10 mg/kg) delivered via an intravenous catheter. An areflexic level of anesthesia was maintained by intravenous infusion of pentobarbital sodium in Ringer solution. All procedures followed National Institutes of Health and UCSF/IACUC guidelines for the care and use of laboratory animals. Procedures of deafening, implantation, chronic stimulation, surgical preparation and recording techniques in the physiological experiment have been described in detail in previous reports (e.g., Snyder et al. 1995; Vollmer et al. 1999, 2005).
Cochlear electrodes were fabricated with two bipolar pairs of platinum-iridium ball-shaped electrodes molded into a silicone rubber carrier and were implanted into the left scala tympani under general anesthesia using aseptic surgical procedures. Because of the large round window and expanded scala tympani in the basal region of the cat cochlea, the intracochlear position of the more basal electrode pair (3,4) relative to the modiolus is more variable (Leake et al. 2000). In the present study, the thresholds of the basal electrode pair (3,4) tend to be higher, the spatial tuning curve (STC) widths tend to be broader, and the dynamic ranges tend to be lower than those observed for the more apical electrode pair (1,2). Also, the thresholds, widths and dynamic ranges for electrode pair 3,4 tend to have greater intersubject variability than those observed for electrode pair 1,2. Consequently, with the exception of the analyses of the cochleotopic organization in the IC, only data from stimulation with the apical electrode pair 1,2 are reported in the present study. The stimulating contacts (290 µm in diameter) comprising the apical electrode pair were separated from each other by 1 mm; they were arranged in an offset-radial orientation and were located on average at 49% (electrode 1) and 45% (2) of basilar membrane distance from the cochlear base. In the normal cochlea the stimulating electrodes would have represented frequencies of 5–6.3 kHz. These frequency assignments are calculated based on Greenwood's frequency-position function (Greenwood 1974, 1990) using the revised constants for the cat cochlea suggested by Liberman (1982).
Chronic stimulation (LDS cats)
In the LDS cats, chronic electrical stimulation was applied for 4 h/day, 5 day/wk for a mean duration of 21 ± 11 (SD) wk. EABR thresholds were determined as described previously (Moore et al. 2002), and chronic stimulation levels were set with maximum signal intensity adjusted to 2 dB above EABR threshold for each individual subject. All LDS animals were stimulated with the apical electrode pair 1,2. Due to lead failure, however, subjects CD393 and CH539 were stimulated with electrode pairs 2,3 and 1,4, respectively, during the final weeks of stimulation.
Electrical stimulation was delivered either by an analogue speech processor (SP) that transduced ambient environmental sounds into electrical signals delivered to the implanted electrodes or by computergenerated amplitude modulated pulse trains. For stimulation with the SP, the frequency spectrum of the analogue stimulation was band-pass filtered from 250 Hz to 3 kHz with a roll-off at the shoulder frequencies of 6 dB/octave. The processor output was logarithmically amplitude-compressed to a dynamic range of 6 dB. The maximum peak-to-peak output was set to 6 dB above the animals' individual EABR thresholds. The peak intensity of the pulsatile signals was set to 2 dB above EABR threshold. The computer-generated signal was a continuous train of electrical pulses (200 µs/phase, charge-balanced biphasic square-wave pulses) delivered at a 300 pps carrier rate and sinusoidally amplitude modulated (SAM) at a frequency of 30 Hz with a modulation depth of 100% (300/30 SAM). The choice of these stimuli was based on previous studies showing that chronic electrical stimulation of the developing and long-deafened adult auditory system using these signals resulted in a significant increase in temporal resolution of ICC neurons (Snyder et al. 1995; Vollmer et al. 1999, 2005).
One animal (CH611) received initial stimulation with an unmodulated pulse train at 80 pps that was delivered by a backpack stimulator for a period of 2 wk. This animal did not tolerate stimulation outside its home cage, and the 80 pps stimulation was used until an analogue SP became available as a backpack stimulator for use in the home cage.
All but one LDS cat received additional stimulation during behavioral training sessions (5 day/wk; see Table 1). These training sessions generally lasted about one hour, although the total duration of suprathreshold electrical stimulation during the behavioral sessions was brief (a total of usually <30 s).
Electrophysiological procedures
EABR and IC thresholds (0 dB = 1 µA peak-to-peak) were estimated for each subject, and final acute electrophysiological experiments were conducted using tungsten microelectrodes to record responses of multi-neuronal clusters and single neurons in the IC. Several penetrations through the IC were made, and response thresholds were recorded at intervals of 100 µm. Responses to biphasic electrical pulses (200 µs/phase) and to three cycles of a 100-Hz sinusoid were recorded. To estimate thresholds, the intensities just sufficient to activate the neuron(s) were determined using audiovisual criteria. Thresholds were plotted as a function of IC depth along the tonotopic gradient of the IC to obtain STCs (Fig. 1). As reported previously, despite intra- and interindividual variability in shapes and widths, STCs are typically W-shaped (e.g., Leake et al. 2000; Moore et al. 2002; Rebscher et al. 2001; Snyder et al. 1990; Vollmer et al. 1999, 2000). The highest threshold region between the two locations of minimum thresholds was defined as the border between the two nuclei (see vertical - - - in Fig. 1) and allowed neurons to be assigned to either ICX or ICC. The location of minimum threshold within the IC ("best location") for stimulation with a given electrode/electrode pair is an indirect measure of the cochleotopic organization of the IC. To evaluate the cochleotopic organization after different durations of deafness, the best locations were estimated and compared for stimulation with the apical electrode pair (1,2) and the basal electrode pair (3,4). The best locations were separately determined for the ICX and the ICC.
|
In addition, the difference between minimum ICC threshold and the higher threshold recorded at the border between ICX and ICC was determined to study the range of intensities over which differences in the spatial extent of excitation were elicited as an estimate of the neural dynamic range for electrical stimulation in the ICC (Fig. 1). At suprathreshold intensities that exceeded the dynamic range, continuous regions of both ICX and ICC were activated.
Histology
After completion of the electrophysiological experiment, the cochleae of all short- and long-deafened animals were prepared for histological analyses. The methods for the preparation of cochlear specimens were identical to those described by Leake et al. (1999) and will be described here only briefly. The cochleae were perfused through the scalae with mixed aldehyde fixative, postfixed in osmium tetroxide, decalcified briefly, embedded in LX resin and mounted on glass slides as block surface preparations. Semithin sections (1–2 µm) were cut at 2-mm intervals along the basilar membrane, stained with toluidin-blue and examined in light microscopy to assess the condition of the organ of Corti and the survival of SGCs. SGC volume ratio (a measure of relative cell density) in Rosenthal's canal was determined using a point-counting method (Leake and Hradek 1988; Leake et al. 1999). Earlier studies in normal hearing animals using this method provided normative data for the cat spiral ganglion (Leake and Hradek 1988). These data served as a control reference in the present study and allowed the SGC density of the SDU and LD cats to be expressed as percent of normal.
Data analysis
For descriptive statistics of normally distributed data, the mean ± SD are reported. If the data were not normally distributed, the median (MDN) and the quartile deviation (Q) are reported. The t-test was used for comparisons between two independent groups. If the data did not meet the criteria for normality and equal variance, the nonparametric Mann Whitney-U test was used for comparisons. If the difference between LDU and LDS animals was not significant (P > 0.05), the data were pooled for subsequent statistical analyses and referred to as LD data without further specification. For comparisons among more than two groups, one-way ANOVAs or the nonparametric Kruskal-Wallis ANOVA on ranks were used, followed by pairwise multiple comparisons (Tukey test or Dunn's method, respectively). For paired comparisons between two variables within a group, the paired t-test was used for normally distributed data, and the Wilcoxon signed-rank test was used if the data were not normally distributed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
All SGC densities reported in this study refer to the implanted left cochlea in each animal. Figure 2 illustrates examples of SGC survival in Rosenthal's canal at the 40–45% cochlear sector for control, SDU and LD animals.
|
). Because SGC densities of the control animals were typically between 95 and 100% of normal, these data are summarized in one symbol (
). The mean SGC density in the SDU group is 45% of normal. The degeneration of SGCs in the LD animals is significantly more severe. SGC survival in the LDU group is reduced to 10.2% of normal, whereas in the LDS group the mean SGC density is only 3.1% of normal (Vollmer et al. 2005). The difference in SGC survival between LDU and LDS animals is statistically significant (t-test; t = –3.123, df = 10, P < 0.05)
|
To study the effects of deafness on the tonotopic frequency organization in the IC, we determined the best locations for pulsatile and sinusoidal bipolar stimulation with the apical electrode pair (1,2) that is centered on average at a location that would represent 5.5 kHz in the normal cochlea and a basal electrode pair (3,4) centered at 12 kHz. The best locations in the midbrain were estimated separately for the ICX and the ICC. As demonstrated for the ICC in Fig. 4, the best locations for pulsatile and sinusoidal stimulation were virtually identical for electrode pair 1,2 (A), and the same was true for electrode pair 3,4 (B; paired t-test and Wilcoxon signed-rank test; P > 0.07 for all groups).
|
Figure 5 compares the best locations for sinusoidal stimulation with electrode pairs 1,2 and 3,4 in the ICX (A) and the ICC (B). Generally, the best locations in the ICX (Fig. 5A) for the apical electrode pair 1,2 are at equal or deeper locations than those for the basal electrode pair 3,4. In the LDU (t = 2.862, df = 4) and LDS animals (t = 4.078, df = 6) the mean best locations for pair 1,2 [1,188 ± 460.29 and 1,271.43 ± 467.13 (SD) µm, respectively] were significantly deeper than those for pair 3,4 [1,134 ± 436.84 and 914.29 ± 333.81 (SD) µm, respectively; paired t-test; P < 0.05]. In control and SDU animals, the best locations for pair 1,2 [mean: 1,671.25 ± 788.62 (SD) µm and median: 870 ± 222.5 (Q) µm, respectively] were also deeper than those for pair 3,4 [mean: 1,498.75 ± 805.03 (SD) µm and median: 815 ± 210 (Q) µm, respectively], but these differences did not achieve statistical significance (t-test and Wilcoxon signed-rank test, respectively; P > 0.05).
|
In summary, the cochleotopic organization of the ICC is inverse to that of the ICX, and this organization appears to be maintained independent of the duration of deafness.
Response thresholds
An important issue in the present study is whether progressive duration of deafness and the consequent loss of SGCs lead to changes in neuronal response sensitivity. To address this issue, EABR thresholds to pulses were measured, and for each STC, the minimum neuronal response thresholds in ICX and ICC to pulses and sines were determined.
First, to characterize global trends in the data, minimum neuronal thresholds for sinusoidal stimulation (apical electrode pair 1,2) are pooled for all experimental groups, and the resulting analysis shows that thresholds in the ICX and ICC are strongly correlated (R2 = 0.7642, P < 0.001). In addition, mean thresholds of ICC neurons pooled for all animals [34 ± 6.5 (SD) dB] are only slightly lower than ICX thresholds [35.2 ± 7 (SD) dB], although this small difference in threshold is significant (paired t-test; t = 2.325, df = 47, P < 0.05). In view of the previously reported differences in the physiological response properties to auditory stimulation between ICX and ICC neurons (e.g., higher frequency following ability, lower response latencies in the ICC) (Vollmer et al. 1999) and the somatosensory and efferent cortical inputs known to influence responses in the ICX (e.g., Aitkin 1986), the following results focus on data obtained from the ICC.
A pairwise comparison between minimum response thresholds for sines and for pulses is illustrated in Fig. 6A for all groups of animals. Response thresholds vary over a wide range of stimulus amplitudes. LD animals generally show higher response thresholds than control animals. At low intensities, sinusoidal and pulsatile response thresholds have a tendency to approach equal charge per phase (- - -). At higher threshold levels, sinusoidal thresholds tend to increase more rapidly than pulsatile thresholds, and the linear regression line approaches the equal intensity line (/). Also, at higher response thresholds the scatter of the data decreases. However, minimum sinusoidal thresholds are always lower than minimum pulsatile thresholds. Because of the low number of threshold pairs (n = 2), the SDU animals were excluded from the statistical comparison. For all other groups, the differences between minimum sinusoidal and pulsatile thresholds are significant (LDU, LDS: paired t-test; controls: Wilcoxon signed-rank test; P < 0.001 for all groups).
|
To examine the effect of duration of deafness on response thresholds in greater detail, Fig. 7 displays final EABR thresholds as a function of the number of months of deafness, i.e., the age at which each of these neonatally deafened animals was studied physiologically. EABR thresholds increase progressively with longer durations of deafness. EABR thresholds from SDU animals are well within the variability of thresholds of control animals (grey area: ±SD of mean EABR thresholds for control animals), whereas all EABR thresholds from LD animals with durations of deafness >30 mo (vertical - - -) exceed the variability of the controls. Paired comparisons of EABR thresholds of LDU (K33: 48 dB and K51: 50 dB) and LDS animals (CH611: 49 dB and CD393: 52 dB) with similar durations of deafness (51 and 78 mo vs. 50 and 79 mo, respectively) show similar threshold values for both groups of animals (t-test; P > 0.05). Thus chronic stimulation appears not to affect EABR thresholds. However, given the small number of comparisons (n = 2), this statistical finding must be interpreted with caution.
|
Figure 8, A–C, summarizes the mean EABR thresholds (A), the median minimum ICC thresholds for pulsatile stimulation (B), and the median minimum ICC thresholds for sinusoidal stimulation (C) for control and experimental animals. In Fig. 8, the data for the LDU and LDS animals have been pooled (LD group). This was done because, as previously mentioned, there was no difference in EABR thresholds between two pairs of age-matched LDU and LDS animals (K33 and K51 vs. CH611 and CD393), and further comparisons of these same age-matched pairs showed that there was no difference in ICC thresholds for pulses and sines (t-test; pulses: t = 1.426, df = 16; sines: t = 1.768, df = 11; both P > 0.05). These additional results strengthen the hypothesis that chronic stimulation has no effect on response thresholds in LDS animals.
|
Because minimum ICC thresholds for pulsatile stimulation were available for only two subjects in the SDU group (Fig. 8B), this group was not included in the statistical comparisons. The data indicate that LD animals have significantly higher thresholds for pulses than the control animals (Mann-Whitney ranked-sum test; t = 171.50, P < 0.001).
In addition, the differences in the median ICC thresholds for sines (Fig. 8C) among the control, SDU, and LD animals were significant (Kruskal-Wallis ANOVA on ranks; H = 15.87, df = 2, P < 0.001). In close agreement to the EABR thresholds, an all pairwise multiple comparison procedure (Dunn's method) showed that there was no difference in median minimum ICC thresholds between control [31.5 ± 3.5 (Q) dB] and SDU animals [32 ± 5.25 (Q) dB; P > 0.05]. In contrast, median minimum sinusoidal ICC thresholds from LD animals [39 ± 5.25 (Q) dB] were significantly higher than those from control and SDU animals.
Overall these results confirm that very long durations of deafness lead to significant increases in neuronal response thresholds. These increases were demonstrated both in far-field (EABR) and near-field recordings. Chronic stimulation had no apparent effect on response thresholds. However, because of the limited number of comparisons between age-matched LDU and LDS animals, this finding has to be treated with caution.
Dynamic range
Neural dynamic range is an estimate of the range of intensities between the minimum threshold for electrical excitation in an STC function and the suprathreshold intensity at which major continuous regions of the ICX and ICC are activated. Once the stimulus intensity exceeds the dynamic range, selective activation of neural populations is impossible, and stimulating channels in a multi-channel CI would presumably loose their functional independence. An important goal of the present study was to examine the effects of duration of deafness and electrical stimulation on the dynamic range of CI subjects. Dynamic range was studied for both pulsatile and sinusoidal stimulation.
Due to the difference in phase durations, STCs obtained with 100-Hz sinusoidal stimuli (5 ms/phase) generally have lower thresholds, markedly sharper tuning (Moore et al. 2002) and larger dynamic ranges than those obtained with pulses (biphasic pulses, 200 µs/phase; Fig. 1, A and B). Because STC widths tend to be very broad (Figs. 1B and 11) and dynamic ranges very small in LD animals (Figs. 1B and 9), a clear W shape of the STCs and a border between ICX and ICC were sometimes not identified. As a consequence, data on the dynamic ranges of STCs, particularly for pulsatile stimulation, are more limited for the long-term deafened animals. Figure 9A summarizes the mean dynamic range for pulsatile stimulation in the control and LD animals. Because only one measure of dynamic range for pulsatile stimulation was available for the SDU animals, this group is not included in the statistical analyses. There is no difference in dynamic ranges for pulses between the age-matched LDU and LDS animals (K33, K51 vs. CH611, CD393; t-test; t = 1.735, df = 5, P > 0.05). Thus chronic stimulation does not appear to affect dynamic ranges in LD animals. The data from LDU and LDS animals were, therefore pooled for statistical analysis.
|
|
It is interesting to note that when the dynamic ranges of all, i.e., not age-matched, LDU and LDS animals are compared (data not shown), LDS animals have significantly smaller dynamic ranges for pulses [mean: 6.72 ± 4.27 (SD) dB] than LDU animals [mean: 3.26 ± 2.45 (SD) dB]. Because there is no difference in dynamic range between the age-matched LDU and LDS animals, this result in the larger groups may be due to the longer duration of deafness in the LDS animals.
For stimulation with 100-Hz electrical sinusoids (Fig. 9B), a statistical comparison of dynamic ranges between age-matched LDU and LDS animals was not possible because of insufficient data (n = 6 and n = 1, respectively). Thus we cannot fully exclude the possibility of confounding effects of chronic electric stimulation or duration of deafness on dynamic ranges for sinusoids. However, comparisons of mean dynamic ranges for sinusoidal stimulation between all LDU [9.14 ± 7.621 (SD) dB] and LDS [5.73 ± 4.29 (SD) dB] animals included in the present study did not show a significant difference (t-test; t = 1.04, df = 25, P > 0.05). The data for LDU and LDS animals were therefore pooled for statistical comparison. The differences in mean dynamic range among the control, SDU and LD animals are significant (ANOVA; F = 10.02, df = 2/80, P < 0.001). A subsequent pairwise comparison shows that there is no difference in dynamic ranges for sinusoidal stimulation between control and SDU animals [means: 17.13 ± 10.10 (SD) dB and 18.84 ± 7.09 (SD) dB, respectively; Tukey test; P > 0.05]. In contrast, LD animals have significantly smaller dynamic ranges for sinusoidal stimulation (mean: 8.39 ± 7.09 dB) than both control and SDU animals (P < 0.05).
To better illustrate the relationship between response sensitivity and dynamic range, Fig. 10 displays the quantitative distribution of the ratios of minimum response thresholds (sinusoidal stimulation) divided by the corresponding dynamic ranges separately for the control, SDU and LD animals. LDU and LDS animals are pooled as LD animals because there is no significant difference in the median ratio values between the two groups (Mann-Whitney U; P = 0.924). The ratios of LD animals (C) are clearly shifted toward higher values as compared with those from control (A) and SDU animals (B). The differences in the median threshold/dynamic range ratio values between control, SDU and LD animals are significant (Kruskal-Wallis on ranks; H = 15.781, df = 2, P < 0.001). Subsequent comparisons (Dunn's method) show that the median threshold/dynamic range ratios of control (A) and SDU animals (B) are not different (P > 0.05). In contrast, the median ratio of the LD animals was significantly higher [median: 4.82 ± 2.86 (Q); P < 0.05 compared with both control and SDU animals. These results suggest that long-term deafness leads to a relative increase in threshold and a relative decrease in dynamic range.
|
Measures of STC widths reflect the spread of activation across the ICC at a given intensity above threshold. These measures provide an estimate of the selectivity of activation with respect to the cochleotopic gradient of the central auditory system. Figure 11A summarizes the mean STC widths for stimulation with electrical pulses delivered at 2 dB above minimum threshold for electrode pair 1,2. Data from SDU animals are not included in the statistical analysis because of the small number of available measures (n = 2). The data for LDU and LDS animals are pooled as LD animals because there is no significant difference between the age-matched animals (t-test; P > 0.05). Thus in accordance to the reported findings for thresholds and dynamic ranges, there is no apparent effect of chronic stimulation on STC width for pulses. Further, there is no difference when the data from all LDU and LDS animals are compared (t-test; P > 0.05). For pulsatile electrical stimulation, LD animals exhibit a significantly broader extent of excitation [mean: 1,166.3 ± 404.86 (SD) µm] compared with control animals [mean: 667.14 ± 337.87 (SD) µm; t-test; t = –2.993, df = 32, P < 0.01]. Moreover, as described in the preceding text, the mean STC width for the LD animals may be underestimated because dynamic ranges in these animals were sometimes even smaller than 2 dB for pulsatile stimulation in which cases estimates of STC widths were not possible.
Figure 11B summarizes the median STC widths for 100-Hz sinusoidal stimulation measured at 6 dB above minimum threshold. As mentioned earlier, minimum thresholds for long phase duration sines (5 ms/phase) are typically lower, STCs are typically sharper (Moore et al. 2002) and have larger dynamic ranges than those for pulsatile stimulation (200 µs/phase). These characteristics of the sinusoidal STCs allow a clearer differentiation of the ICX/ICC border and of STC width. For sinusoidal stimulation, a statistical comparison between the age-matched LDU and LDS animals was not possible because of insufficient data (n = 7 and n = 1, respectively). However, when all LDU and LDS animals were compared, there was no significant difference in spread of activation as measured by the STC widths (t-test; P > 0.05). Therefore the data from the two groups were pooled for statistical comparisons (LD group). The differences in median 6 dB widths for sines among the control, SDU and LD animals are significant (Kruskal-Wallis on ranks; H = 8.09, df = 2, P < 0.05). In subsequent comparisons, there is no significant difference between control [median: 870 µm ± 365 (Q) dB] and SDU animals [median: 840 µm ± 261.25 (Q) dB] (Dunn's method; P > 0.05). In contrast, LD animals have significantly broader 6 dB STC widths for sines [median: 1,320 µm ± 580 (Q) dB] than both the control and SDU animals (P < 0.05).
Figure 12 illustrates the relationship between the 6 dB STC widths for sines and the corresponding dynamic ranges of the same STCs for all groups of animals. The slope of the regression line (power function) is negative, and the correlation coefficent of the data is relatively high (R = 0.71). These results indicate a systematic relationship between broader, less selective electrical activation (i.e., larger STC widths) and smaller dynamic ranges. Conversely, relatively selective or narrow STCs have larger neural dynamic ranges. Because dynamic ranges are particularly limited and STC widths are particularly broad in long-deafened animals, the results reflect a marked degradation in spatial selectivity of ICES after long durations of deafness, again suggesting that greater channel interaction would occur in these animals (see summaries in Figs. 9 and 11).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONGRANTSACKNOWLEDGMENTSREFERENCES |
|---|
Cochlear pathology
Peripheral pathology is a likely factor affecting auditory perception of cochlear implant recipients. In an animal model of neonatal deafness, we observed progressive loss of SGCs with increasing durations of deafness. SDU animals had an average SGC density of 45% of normal. After prolonged periods of deafness, the SGC density in LDU animals was severely reduced to an average of 10% of normal and in LDS animals to an average of only 3% of normal. These results reflect previous findings in a partially overlapping set of SDU and LD animals (Rebscher et al. 2001; Vollmer et al. 2000, 2005). It is important to note that the mean SGC survival in the LDS animals was significantly lower than that in the LDU animals. Presumably, this difference is due to the significantly longer duration of deafness in the LDS animals (mean: 69 mo) compared with the LDU animals (mean: 44 mo; t-test; t = –2.663, df = 10, P < 0.05). These results are consistent with other studies that have reported progressive loss of SGCs and radial nerve fibers for many years after aminoglycoside-induced hair cell loss (e.g., Hardie and Shepherd 1999; Leake and Hradek 1988; Lustig et al. 1994; Xu et al. 1993).
A trophic effect of chronic electrical stimulation in promoting SGC survival, as demonstrated in neonatally deafened, early stimulated animals (e.g., Leake et al. 1999), was not observed in the LDS animals, presumably because of the already severe reduction in SGC count after >3.5 yr of deafness. It is interesting to note that the pathology and neural degeneration seen in these cat cochleae after several years of deafness appears to be considerably more severe than the alterations one would expect in the human cochlea after equivalent periods of profound deafness even after congenital deafness (e.g., Linthicum and Anderson 1991; Nadol et al. 2001). The extent and progression of hair cell loss and SGC degeneration are likely dependent on the extent and rapidity of onset of aminoglycoside-induced damage to the cochlea and possibly also on species-specific differences in neuronal survival patterns (Forge and Schacht 2000; Hinojosa et al. 2001; Johnsson et al. 1981; McFadden et al. 2004; Sone et al. 1998). The dosage of aminoglycosides in our experimental design was intentionally higher than would be used clinically to reduce the variability of the results and to provide a reliable model for profound neonatal hearing loss. Obviously when aminoglycosides are used clinically, every effort is made to limit administration to levels that will minimally affect hearing. Moreover, with most etiologies, onset of deafness is often much more gradual than the 2- to 3-wk interval over which these experimental animals were deafened. Thus the degenerative peripheral changes in the LD animals are more severe than those observed in many human cochlear implant users even after many decades of deafness. These animals were specifically intended to model the extreme condition of very severe pathology to examine the functional consequences.
Although there is no clear evidence that the degree of SGC degeneration is a predictor of speech perception (e.g., Fayad and Linthicum 2006; Khan et al. 2005a,b), it is generally assumed that SGC survival plays a critical role for word recognition, especially if the number of SGCs is very low (Clopton et al. 1980; Incesulu and Nadol 1998; Khan et al. 2005b; Nadol 1984; Nadol et al. 1989; Otte et al. 1978). However, many other variables, especially including alterations in the central auditory pathway and its signal processing capacity, likely play important roles in the success of cochlear implantation.
Cochleotopy
The estimation of the cochleotopic organization of the IC was based on the best location for stimulation with different electrode pairs. In contrast to observations in the cortex by Raggio and Schreiner (2003), differences in waveform (pulses vs. sines) did not contribute to differences in the location of highest stimulus sensitivity. Thus at the level of the auditory midbrain, specific properties of the electrical stimuli, including differences in stimulus rise time and stimulus duration, did not influence the most sensitive response location relative to the cochleotopic organization.
Some earlier studies have reported only vague indications that the characteristic frequencies of ICX neurons are tonotopically organized with high frequencies being located more superficially in the ICX and low frequencies being represented deeper in the ICX (e.g., cat: Aitkin et al. 1994; guinea pig: Binns et al. 1992). In contrast, other anatomical and physiological studies suggest solid evidence that all subdivisions of the IC, i.e., including the ICX, exhibit a tonotopic organization (cat: Andersen et al. 1980; rat: Saldaña et al. 1996). When the data from all animal groups were pooled, the present study supports the latter finding. For all groups of animals, including the long-deafened animals, a trend was observed that the best locations for higher frequencies, i.e., more basal stimulating electrodes, were located more superficially in the ICX than those for lower frequencies. However, this trend did not achieve statistical significance in all groups. At least two potential factors may play a role. First, the depth of the ICX is relatively shallow [mean across all groups: 1,646 ± 578 (SD) µm, unpublished observations], and the estimation of response thresholds at intervals of 100 µm may be too coarse to detect small differences in best location. Second, the surgical access to the IC and its exposure over a period of several days as required for the completion of these experiments may result in traumatic or inflammatory changes near the surface of the IC that could affect responses, particularly in the ICX.
The ICC, in contrast, exhibited for all groups of animals a clear cochleotopic frequency gradient that was systematically related to ICC depth. These results are in agreement with data from normal-hearing cats (e.g., Brown et al. 1997; Merzenich and Reid 1974; Oliver 1987; Oliver and Morest 1984; Rose et al. 1966) and neonatally deafened cats that were either unstimulated (deafness durations pooled over 0.5–2.6 yr, including a single long-deafened animal) or implanted and stimulated at a young age (Snyder et al. 1990, 1991). In the present study, cochleotopic organization in the ICC was maintained even after very prolonged periods of deafness. This finding is in contrast to previous studies in the auditory cortex (Dinse et al. 1997; Hartmann et al. 1997; Klinke et al. 1999; Raggio and Schreiner 1999, 2003; Taniguchi et al. 1997) suggesting that sustained lack of auditory input and severe degeneration of peripheral innervation result in a degradation of cortical cochleotopic organization. The present results suggest that the cochleotopicity of the ICC is less affected by long-term auditory deprivation than that of the auditory cortex. This difference may relate to the larger range of sources of subcortical and cortical inputs to the primary auditory cortex that can be altered in their response specificity by long-term deafness.
Thresholds
ICX VERSUS ICC.
Across all animals minimum sinusoidal thresholds in the ICC were slightly (1 dB) but significantly lower than those estimated in the ICX. This difference is in close agreement with ICX/ICC threshold comparisons in our laboratory for a large number of animals with various deafness durations and stimulation histories [number of threshold comparisons = 109; mean difference: 1.69 ± 4.12 (SD) dB; unpublished data]. The reason for the higher minimum threshold values in the ICX may be related to the additional somatosensory (e.g., Aitkin 1986) or descending inputs from both primary and nonprimary cortices (e.g., Anderson et al. 1980; Coleman and Clerici 1987; Faye-Lund 1985; Oliver and Huerta 1992). Such multisensory and multisource cortical inputs may lead to a decreased synchrony of neural responses and thus to increased response thresholds. Furthermore, the lower response sensitivity of ICX neurons may simply reflect an increased risk of mechanical trauma to the ICX during surgical exposure of the auditory midbrain for recording.
PULSES VERSUS SINES.
The present study evaluated the effects of duration of deafness on response sensitivity and spatial selectivity of sinusoidal and pulsatile electrical activation in the IC of adult cats. Comparisons of thresholds derived from animals for which the responses to both pulsatile and sinusoidal stimuli were examined demonstrated some stimulus-specific differences in IC responses. When expressed in peak-to-peak current, minimum response thresholds for sinusoidal stimulation were significantly lower than those for pulsatile stimulation. There was no significant difference in mean threshold difference among the different groups of animals, and the mean threshold difference across all animals was 13.54 ± 6.99 (SD) dB. This difference is in agreement with previous reports about threshold differences for varying phase durations in the auditory nerve (Hartmann et al. 1984; Shepherd and Javel 1997) and the inferior colliculus (Leake et al. 2000; Moore et al. 2002; Snyder et al. 1990), the auditory cortex (Raggio and Schreiner 2003) and reports of psychophysical threshold estimates for different phase durations (Beitel et al. 2000a,b; Moon et al. 1993; Pfingst and Morris 1993; Smith and Finley 1997; Smith et al. 1994).
The differences in thresholds for sinusoidal and pulsatile stimulation can be partially explained by the different stimulus durations (He 1997) and phase durations, i.e., the total charge per phase rather than the peak amplitude. Although the scatter of the data in Fig. 6A was relatively large for low sinusoidal response thresholds, sinusoidal and corresponding pulsatile thresholds tended to approach equal charge per phase. At higher stimulation current levels, a more rapid growth of sinusoidal thresholds relative to the pulsatile thresholds is observed. That is, at these higher intensities, the charge per phase required to reach thresholds is lower for pulsatile stimulation than for sinusoidal stimulation. Thus a simple charge-integrator-model is not valid for higher minimum response thresholds. However, the peak current for sinusoidal thresholds remains below that for pulsatile thresholds. A similar compression of the pulsatile or expansion of the sinusoidal efficacy with increasing response thresholds was also observed in the auditory cortex by Raggio and Schreiner (2003). It is possible that even near threshold the fast rise time of current in pulsatile stimulation leads to a higher temporal synchronization of neural inputs (Heil 1997) and thus to lower response thresholds when expressed relative to charge per phase. The higher degree of synchrony may also contribute to the overall reduction in scatter in the corresponding thresholds at higher intensities. It is also possible that the temporal integration of longer-duration sinusoidal stimuli is more affected by prolonged durations of deafness (that are typically associated with higher thresholds) and peripheral loss of SGCs than that for pulsatile stimulation with shorter phase and shorter stimulus durations (van den Honert and Stypulkowski 1984).
ICC VERSUS EABR THRESHOLDS.
When measured in the same animal, minimum response thresholds in the ICC were lower than EABR thresholds, and the two measures were strongly correlated. The mean threshold difference was on the order of 4.5 dB. These data are consistent with previously published data from other studies in our laboratory (Beitel et al. 2000a,b; Moore et al. 2002; Vollmer et al. 2000
, dynamic ranges for pulsatile stimulation with pair 1,2 (difference between minimum threshold and higher threshold at border between ICX and ICC). BL, best locations in ICX and ICC for sinusoidal stimulation with electrode pair 1,2.
