| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Eaton-Peabody Laboratory, Massachusetts Eye and Ear Infirmary, Boston, 02114; 2 Department of Otology and Laryngology, Harvard Medical School, Boston, 02114; 3 Program in Speech and Hearing Bioscience and Technology, Division of Health Science and Technology, Harvard/Massachusetts Institute of Technology, Cambridge, Massachusetts 02139
Submitted 3 July 2003; accepted in final form 12 August 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
). The mammalian OC system consists of two components (Warr and Guinan 1979). The medial (M)OC system originates near the medial superior olive and projects via myelinated fibers primarily to cochlear outer hair cells (OHCs). The lateral (L)OC system originates in and around the lateral superior olive and projects via unmyelinated axons primarily to the dendrites of cochlear afferent neurons in the region beneath inner hair cells (IHCs) (see (Guinan 1996).
En route from the brain to the ear, OC axons coalesce into a bundle that arcs dorsally from the olivary complex and crosses the midline at the floor of the fourth ventricle. Electrical activation of the OC bundle (OCB) at this midline site decreases cochlear sensitivity, as seen in compound action potentials (CAP), responses of single cochlear afferent fibers (Wiederhold and Kiang 1970), intracellular responses from hair cells (Brown and Nuttall 1994), otoacoustic emissions (Mountain 1980), or motion of the basilar membrane (Murugasu and Russell 1996). These classic suppressive effects of OCB stimulation are mediated by MOC fibers contacting OHCs (Guinan 1996) and releasing acetylcholine (ACh). ACh binds to the 
9/10 ACh receptor complex on OHCs (Elgoyhen et al. 2001), leading to Ca2+ entry and activation of K+ channels (Housley and Ashmore 1991). This conductance increase ultimately leads to a reduction in OHC contributions to cochlear amplification. These MOC effects are relatively fast, with an onset time constant of approximately 100 ms (Liberman et al. 1996; Wiederhold and Kiang 1970). Slow MOC effects, with onset time constants of approximately 10 s, have also been documented (Sridhar et al. 1995).
The peripheral effects of the LOC system remain unknown: the unmyelinated axons of this pathway are difficult to stimulate with OCB shocks at the floor of the fourth ventricle (Gifford and Guinan 1987; Ranck 1981). Hints as to LOC effects on the periphery come from 1) studies of chronic versus acute effects of surgical de-efferentation on responses in cochlear afferents (e.g., Liberman 1990), 2) pharmacological studies of the effects on hair cell organs of neurotransmitters immunolocalized in cochlear terminals of the LOC system (e.g., acetylcholine; Felix and Ehrenberger 1992), GABA (Arnold et al. 1998), opioid agonists (Sahley et al. 1991), and calcitonin gene related peptide (CGRP) (Bailey and Sewell 2000), and 3) studies of cochlear responses in mice with targeted deletion of putative LOC transmitters such as urocortin (Vetter et al. 2002) or CGRP (Maison et al. 2003). These studies have suggested that LOC neurons may produce a combination of acute excitatory and inhibitory effects on cochlear neural output and may also play a trophic role in normal cochlear development.
This study takes a new approach to activating the unmyelinated LOC system, i.e., indirect excitation via electrical stimulation in the inferior colliculus (IC). The IC is a nexus of ascending and descending auditory pathways (Huffman and Henson 1990; Nieuwenhuys 1984) and has been shown to project directly to the MOC cells (Thompson and Thompson 1993). Correspondingly, IC stimulation can elicit MOC-like effects on the cochlea (Mulders and Robertson 2000). However, the LOC system may also be a target of IC projections, given that axonal degeneration is seen bilaterally in the lateral superior olive (LSO) following IC lesions (Noort 1969). If IC stimulation can activate the LOC system, it should evoke novel changes in cochlear responses that are fundamentally different from those seen when shocking the OCB.
In this study, electrical stimulation at some IC sites elicited MOC-like effects, both "fast" and "slow." However, at other sites, novel patterns of cochlear response modulation were evoked: long-lasting changes in cochlear neural responses (either enhancement or suppression) without accompanying changes in the OHC-based cochlear responses generated "upstream" from the IHC/afferent synapse (i.e., distortion products and cochlear microphonics). Surgical lesions, placed to differentially interrupt the LOC and MOC systems as well as the autonomic pathways to the cochlea, support the view that these novel effects are LOC mediated. We conclude that IC stimulation can activate the LOC system, and that the LOC activation can elicit both suppression and enhancement of cochlear neural responses.
This work was also submitted in partial fulfillment of the requirements for PhD in the Program for Speech and Hearing Bioscience and Technology at Harvard and MIT.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
Albino guinea pigs (250–650 g) of either sex were anesthetized with droperidol (10 mg/kg, im), fentanyl (0.2 mg/kg, im), and pentobarbital sodium (25 mg/kg, ip). Boosters were given every 2 h or on elevation in heart rate: droperidol and fentanyl (1/3 original dose) and pentobarbital sodium (1/5 original dose). All animal procedures were approved by the IACUC of the Massachusetts Eye and Ear Infirmary.
Surgery
Following the administration of anesthesia, a tracheotomy was performed. The animal was intubated and connected to an artificial respirator. To eliminate effects of middle ear muscles, paralysis was induced with tubocurarine (1.25 mg/kg, im) in most experiments and maintained by re-injection every 2 h. Respiration volume was adjusted to maintain CO2 levels between 4 and 5%, and rectal temperature was maintained at 36–39°C. The bullas were exposed, and the cartilaginous ear canals were cut. Small holes were shaved in the bullas to gain access to the round window (RW) bilaterally. Silver-ball electrodes were centered on the RW membrane. A posterior craniotomy was made extending anteriorly to the tentorium, posteriorly to the foramen magnum, and laterally about 4 mm. The cerebellum was aspirated to expose the floor of the fourth ventricle, and the IC.
Acoustic stimulation and response measurement
Etymotics Research ER10c microphone/dual sound-source assemblies were bilaterally sealed in the ear canal. The ER10C microphone was calibrated in a coupler to a reference microphone (Bruel and Kjaer 1/4"condenser microphone). Digital stimulus generation and response digitization were synchronously and simultaneously sampled at 100,000 Hz using I-O boards (PCI-6052E, National Instruments) in a PC-driven data acquisition system under LabVIEW Control. Acoustic stimuli included tone pips (5-ms duration with 0.5-ms rise-fall times (cos2 shaping), clicks (100-µs duration), and continuous tones to elicit distortion product otoacoustic emissions (DPOAEs).
Responses from the RW electrode, including the tone-pip-evoked CAP, the click-evoked cochlear microphonic (CM), and the RW noise, were amplified 10,000 times by Grass amplifiers and filtered (0.3–30 kHz) before digitization and averaging. Ear canal sound pressure (for measurement of DPOAEs and monitoring of the levels of the primaries) was amplified 100 times by the Etymotics preamp.
To facilitate comparisons of CAP, CM, and RW noise and DPOAE data, response amplitudes were normalized to a 5-min preshock window and expressed as a 
Response in decibels from mean preshock values. Slow baseline drift in the Response was removed by fitting a straight line to 1-min data windows approximately 20 min before and after the shock epoch and subtracting out the fitted line. The slope of this normalization line was always <0.03 dB/min.
Stimulating electrodes and shock trains
Before paralysis, a bipolar stimulating electrode fashioned from two fine platinum wires (0.005 in) with a 1-mm tip separation was placed on the OCB at the floor of the fourth ventricle, and the facial twitch threshold was determined. For OCB stimulation, shock trains consisted of 0.15- or 0.3-ms monophasic pulses delivered at 300–333 Hz), 10 dB above twitch threshold.
Two types of bipolar electrodes were used to stimulate the IC. Early experiments, including most level-function experiments, used a bipolar electrode made of two fine platinum wires (0.005 in) with a tip separation of 1 mm. Later experiments, including all time-function experiments, used a Microprobe 5-kOhm platinum bipolar electrode with a tip exposure of 125 µm and tip separation of 500 µm. Both types of electrode were inserted to a depth of 0.5–1.0 mm. The locations of IC stimulation sites were marked, based on visual inspection, on a diagram of the caudal surface of the IC. The IC shock trains consisted of 0.2–1.0 ms (0.8 ms typical) alternating polarity pulses, delivered at 50–200 Hz (100 Hz typical), at a level of 5–15 dB (10 dB typical) above current levels producing facial twitches at the OCB stimulation site.
In one experiment, electric shocks were delivered to the LSO with a monopolar 15-kOhm tungsten electrode with tip exposure of 125 µm and with a 1-µl Hamilton syringe that was insulated except for 0.5 mm at the tip. In both cases, the reference cathode was in the neck muscles. The tungsten electrode was placed into the LSO stereotaxically while recording gross responses to ipsi, contra, and bilateral stimuli as described by (Guinan et al. 1972). The Hamilton syringe was placed along the same path as the tungsten electrode. LSO shock trains were 40-Hz, 1-ms-duration alternating polarity pulses, delivered at levels of approximately 1 V. Electrode location was verified in Nissl-stained brain stem sections (see Measurement paradigms).
Measurement paradigms
Two paradigms were used. All measures were made simultaneously in the two ears.
For the first (CAP-level function) paradigm, only tone-pip-evoked CAP was measured, as a function of tone-pip level, with and without shocks. One pip was presented every 500 ms; the shock trains were 350 ms in duration, with the last shock ending 10–20 ms before the tone pip. Responses to 16 pips were averaged at each level, and three level functions were measured in succession: without, with, and then again without the shocks.
For the second paradigm (time-function paradigm), a number of measures of cochlear response, including CAP, CM, DPOAEs, and RW noise, were obtained every 13 s, before, during, and after a shock epoch lasting from 1 to 5 min (Fig. 1, A and B). During the first 8 s of each 13-s repeating period, CAP and CM were measured: one of four acoustic stimuli was presented every 125 ms, alternating among 1) 13-kHz tone-pip at 15 dB above CAP threshold, 2) 3-kHz tone-pip at 35 dB above CAP threshold, 3) 13-kHz tone-pip at 50 dB above threshold, and 4) a 0.1-ms click at 50 dB above CAP threshold. CAP threshold was defined as the stimulus level required to evoke a 20-µV response. After one quartet of stimuli was presented, polarities were reversed and another quartet was presented; this 1-s octet (Fig. 1C) was repeated eight times, and separate averages of the RW response to each stimulus computed. After each 8-s ensemble of CAP measurements, a tone pair was presented for 1 s and DPOAEs were measured, followed by 4 s of silence in which RW noise was measured. For DPOAEs, f2 was always at 13 kHz (f2:f1 = 1.2), and f2 level was 10 dB < f1 level. The levels of the primaries were 20 dB greater than that required to elicit a 2f1 – f2 DPOAE of –5 dB sound pressure level (SPL). During shock epochs, the shock train was presented at a 75% duty cycle: a 30-ms pause was inserted every 125 ms, with 25 ms preceding the tone pip or click (Fig. 1C). The same shock pattern was maintained through the 1-s DPOAE measures and the 4-s silent period.
|
To extract from these time functions the magnitude and time constant of long-lasting postshocks effects on response measures, the normalized postshock data for each measure from each trial were fit with an exponential function of the form: A x e(–t/T), where A is the initial postshock magnitude of the change in response in decibels, and T is the time constant characterizing the rate of return toward preshock baseline. A nonlinear algorithm minimizing the mean squared error was used. This algorithm required initial parameters, set by first fitting the data with a straight line.
Pathway lesions
Three types of surgical lesions were performed, in various combinations, to elucidate the pathways mediating the shock-evoked effects under study: 1) transections of the crossed OCBs via a shallow cut at the midline on the floor of the fourth ventricle, and 2) unilateral transections of the entire OCB via a deeper cut positioned laterally at the sulcus limitans on the floor of the fourth ventricle and unilateral surgical removal of the superior cervical ganglion (SCG) to remove the source of cochlear autonomic fibers. OCB transections were performed with a surgical microknife using well-characterized surface landmarks on the floor of the fourth ventricle (Kujawa and Liberman 1997). In selected cases, the bundle transections were verified histological in acetylcholinesterase stained brain stem sections of aldehyde-fixed animals (Osen and Roth 1969). To ablate the SCG, the ganglion was first dissected from underneath the bifurcation of the carotid artery and removed by cutting the sympathetic trunk, the branch joining the facial and vagus nerve through the auricular branch of the vagus, and the sympathetic branches to the base of the skull (Hultcrantz et al. 1982; Spoendlin 1981; Terayama et al. 1966).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONDISCLOSURESACKNOWLEDGMENTSREFERENCES |
|---|
OVERVIEW AND STRATEGY. The CAP is a sensitive measure of cochlear performance and has been commonly used in studies of classic MOC effects. The aim of the first series of experiments was to find brain stem stimulation sites evoking novel changes in CAP amplitude-versus-level functions, i.e., different from MOC effects evoked via OCB stimulation at the floor of the fourth ventricle. The latter are characterized by a reduction in CAP, which is maximal at low sound levels (Desmedt 1962) and shows relatively fast onset and offset time constants of approximately 100 ms (Wiederhold 1970). In addition to these "fast" MOC effects, there is "slow" MOC-mediated CAP suppression, with onset and offset time constants of approximately 10 s (Sridhar et al. 1995). MOC "slow" and "fast" effects can be evoked together by shocking the OCB; both are coupled to enhancements of the cochlear microphonic (CM).
Stimulating electrodes were placed superficially (depth < 1 mm) at 41 different sites across the surface of the IC (Fig. 2G) in 14 experiments. In some cases, stimulating electrodes were also placed on the OCB at the floor of the fourth ventricle. In one experiment, a monopolar stimulating electrode was advanced stereotaxically into the LSO, the site of origin of the LOC system.
|
For IC stimulation sites, shocks often elicited MOC-like effects: i.e., CAP suppression similar to that seen with OCB stimulation. Putative LOC-mediated effects were evoked at some IC and LSO electrode sites. Long-lasting CAP enhancements were sometimes seen, which differed from MOC effects in both the sign of the CAP change and the lack of dependence on stimulus level. Another class of putative LOC-mediated effects was characterized by long-lasting CAP suppression, which differed from MOC slow effects in the lack of level dependence.
THREE CLASSES OF COCHLEAR EFFECTS ARE ELICITED BY IC STIMULATION SITES. Consistent with previous reports (e.g., (Dolan and Nuttall 1988; Mulders and Robertson 2000), shocks at some IC sites (n = 12) elicited CAP suppression similar to those seen with OCB stimulation (Fig. 2, A and D), i.e., with maximal effects at low sound pressure levels. In the case shown in Fig. 2A, only fast effects are visible, i.e., CAP amplitudes have returned to control levels within seconds after the termination of the shock train (gray line in Fig. 2A). This constellation of effects will be referred to as IC-evoked MOC-[fast] effects. Data in Fig. 2A are from the cochlea contralateral to the IC electrodes. MOC[fast] effects were usually smaller in the ipsilateral ear (as described more fully in the second experiment series).
Stimulation at other IC sites (n = 6) enhanced CAP-level functions, and this enhancement persisted for tens of seconds (at least) after shock termination. For the case illustrated (Fig. 2B), CAP amplitudes remain elevated in the "After IC Shock" period, i.e., approximately 5 s after the end of each shock train (Fig. 2B, gray line). These slow enhancements were observed bilaterally and were not seen when shocking the OCB in the same animal (one of the dashed gray lines in Fig. 2D). This constellation of effects will be referred to as CAP[+].
Stimulation at a third set of IC sites (n = 3) caused bilateral suppression of CAP-level functions that differed from MOC-[fast] effects in that 1) suppression was greater (in absolute millivolt response shift) at high sound levels than low sound levels, and 2) suppression was sustained in the After Shocks period (Fig. 2C). This constellation of effects in CAP-level functions will be referred to as CAP[–].
Last, stimulation at many sites, especially those located more rostrally and medially, elicited no significant cochlear effects (Fig. 2G).
In Fig. 2 (middle row), contralateral-ear data are superimposed for all IC stimulation sites at which effects on CAP-level functions were observed. Each electrode site is categorized as producing a MOC[fast], CAP[+], or CAP[–] type response. These data are compared with analogous measures obtained in the same animals when stimulating the OCB (Fig. 2D) and to data obtained in one experiment with a stimulating electrode in the LSO (stars in Fig. 2, E and F). The CAP-level functions are converted to CAP-level functions (see METHODS) and plotted on a logarithmic (dB) scale (Fig. 2, D–F) such that a constant fractional change in CAP magnitude appears as a constant effect in decibels.
IC sites evoking only fast suppression (solid lines in Fig. 2D) also showed CAP functions with strong sound-level dependence, i.e., similar to those seen when stimulating the OCB (dashed lines in Fig. 2D). By contrast, IC sites (solid lines in Fig. 2F) showing suppression both during and after the shock trains showed CAP functions with little level dependence and smaller suppression magnitudes. IC sites eliciting CAP enhancements both during and after the shock trains (Fig. 2E) showed more complex behavior. For example, five of eight showed level-dependent CAP suppression at low levels coupled with level-independent CAP enhancements at high levels. Lesions interrupting only MOC fibers eliminated the low-level suppression in such cases, without changing the high-level enhancement. Further evidence for different sources for the suppression and enhancement in such complex functions is the fact that only the enhancements persisted into the After-Shocks period.
TONOTOPY AND LATERALITY OF IC-EVOKED MOC[FAST] EFFECTS. Data discussed above were obtained with tone pips between 11.0 and 13 kHz. In one experiment (Fig. 3A), the tonotopic organization of descending projections was investigated by measuring CAP suppression at 3, 6, and 13 kHz when shocking the IC in each of three stimulation sites, spanning the dorsalventral axis (Fig. 3A, inset). The ventral stimulation sites maximally suppressed CAP at 13 kHz, while stimulation more centrally in the IC maximally suppressed at 6 kHz, and stimulation dorsally maximally suppressed CAP at 3 kHz (Fig. 3A). Thus the preponderance of "No Effect" sites in central and dorsal IC regions schematized in Fig. 2G may be an artifact of the tone-pip frequency utilized: if lower frequency pips had been studied, many of these sites would likely have shown MOC[fast] effects as well.
|
Data summarized in Fig. 2 were extracted from the cochlea contralateral to the IC electrode. Effects in the ipsilateral cochlea were qualitatively similar. Quantitative differences were evaluated only for MOC[fast] effects, as evoked by electrodes in the ventrolateral corner of the IC, where the density of stimulation sites was highest. As summarized in Fig. 3B, the IC-evoked CAP suppression in the ipsilateral ear was always smaller in magnitude, and the inter-ear difference increased with increasing stimulus frequency.
CAP ENHANCEMENT AND SUPPRESSION CAN BE EVOKED BY LSO SHOCKS. In one experiment, stimulating electrodes were stereotaxically advanced into the LSO. On the first penetration, stimulation produced CAP[–] effects ipsilaterally (Fig. 2F, stars) with no effect contralaterally (data not shown: CAP was < –1 dB between 20–70 dB SPL). On the second penetration, stimulation produced MOC[fast] effects at low SPLs but CAP[+] effects at high sound level in the ipsilateral ear (Fig. 2E, stars). Contralaterally, only MOC[fast] effects were observed (data not shown). Evaluation of Nissl-stained sections revealed that the electrode was centered rostro-caudally on the LSO, at its lateral border. When stimulating electrodes were moved, in the same animal, to the OCB, only MOC[fast] effects were observed (one of the dashed gray lines in Fig. 2D).
Second experimental series: time functions
STRATEGY AND OVERVIEW. To study the time-course of long-lasting IC-evoked effects and to differentiate effects occurring previously postsynaptic to the IHC/afferent synapse, a paradigm was devised in which two neural measures (CAP and RW noise) and two OHC-based measures (CM and DPOAEs) were repeatedly sampled before, during, and after an epoch of shock trains lasting 1–2 min (Fig. 1). Each series of preshock, duringshock, and postshock data, normalized to the preshock baseline (e.g., Fig. 4A), is referred to as a trial. In this series of experiments, 43 IC stimulation sites were studied in 21 animals, with a total of 200 trials.
|
CAPs were measured at two frequencies (3 and 13 kHz) to allow comparison of effects on sound-evoked neural activity in basal and apical halves of the cochlea. These two frequencies were chosen to help differentiate MOC slow effects from putative LOC effects: the former peak for 13-kHz stimulation and are usually smaller for 3-kHz stimulation (Sridhar et al. 1995). At 13 kHz, two sound pressures (15 and 50 dB from CAP threshold) were used to reveal level dependence. At 3 kHz, the tone pips were presented at an intermediate level: 35 dB from CAP threshold. The second indicator of cochlear neural activity was RW noise, measured in the absence of exogenous sound by an electrode on the RW (Dolan et al. 1990), which provides an indirect assessment of levels of spontaneous activity in the auditory nerve.
The two measures of presynaptic, i.e., OHC-based, effects in the time-function paradigm were CM and DPOAEs. The CM was measured in response to moderate-level clicks, as in previous studies of MOC-based effects, to infer changes in OHC conductance (Sridhar et al. 1995). CM was also extracted from the tone-pip evoked responses to both 3 and 13 kHz, although the signal-to-noise ratios were lower than for click-evoked responses. Only data for click-evoked responses will be discussed; however, the pip-evoked results were qualitatively similar. DPOAEs were measured with the higher frequency of the two-tone pair (f2) at 13 kHz, to match the tone-pip frequency used to evoke CAP.
Based on the magnitude and sign of the long-lasting changes in CAP and CM, three categories of IC-evoked effects were defined. Two effect categories were consistent with LOC mediation, given that they involved slow suppression or enhancement of the CAP without significant change in CM: these categories are referred to as CAP[+]CM[0] and CAP[–]CM[0]. The third effect category, CAP[–]CM[+], may represent MOC slow effects, as both show long-lasting CAP suppression mirrored by CM enhancement.
FAST VERSUS SLOW EFFECTS: QUALITATIVE FINDINGS. As expected from the CAP-level results, IC stimulation often produced effects during the shock epochs with fast onset and offset (Fig. 4B), similar to those seen with OCB shocks (Fig. 4A), i.e., suppression of CAP and DPOAEs coupled with enhancement of CM. Given the time sampling of this paradigm (Fig. 1), the speed of effect onset/offset can only be specified as 
13 s. As for IC-evoked MOC[fast] effects seen with CAP-level functions (Fig. 2, A and D), the suppression of CAP to low-level tones is greater than for high-level tones (Fig. 4B: open vs. filled circles). For IC stimulation, the opposing effects on CAP and CM tended to decay more dramatically during the shock epoch than seen with OCB stimulation. Nevertheless, these fast, during-shocks effects on the contralateral CAP and CM disappeared when the crossing MOC pathways were sectioned at the midline; thus they likely represent activation of the MOC system. In search of LOC-mediated effects, we concentrated attention on the postshock (i.e., long-lasting) effects of IC stimulation.
OCB stimulation with the time-function paradigm elicited MOC slow effects like those described previously (Sridhar et al. 1995), i.e., long-lasting CAP suppression coupled with long-lasting enhancement of click-evoked CM. Such MOC [slow] effects were evoked in two of five animals in which the time-function paradigm was performed with OCB shocks. Similar effects were also evoked with the time-function paradigm at some IC sites.
Stimulation with the time-function paradigm at other IC sites (e.g., Fig. 5, A and B) produced long-lasting changes in CAP that were not accompanied by changes in CM. As with the CAP-level paradigm, these long-lasting CAP changes were suppressive for some IC sites and enhancing for others. Furthermore, in contrast to the MOC[fast] during-shocks effects in Fig. 4, the long-lasting CAP for trials in Fig. 5, A and B, was similar in magnitude for low-level and high-level tone pips at 13 kHz, reminiscent of the level-independent long-lasting effects in Fig. 2, B and C. Indeed, we assume these long-lasting CAP effects captured by the time-function paradigm (e.g., Fig. 5, A and B) represent the same phenomena as the CAP[+] and CAP[–] effects captured by the level-function paradigm (e.g., Fig. 2, B and E and C and F). The case is clearest for the long-lasting suppression data in Figs. 2C and 5B, since both were produced by stimulation at the same IC site.
|
QUANTIFYING AND CATEGORIZING LONG-LASTING EFFECTS BASED ON CAP AND CM. To quantify the long-lasting changes in cochlear potentials and to identify all IC sites at which they were evoked, postshock CAP data from the contralateral ear at 3 kHz was fit by a single exponential (solid line in Fig. 5, A and B). The resulting magnitudes and time-constants from trials with sufficient data to allow a reasonable fit (n = 119) are shown in Fig. 5, C and D. The sign of the CAP was never different between the two ears. Note that a single IC site might, in some cases, generate multiple trials. The sign of CAP at one site was always the same across repeated trials.
Trials with postshock CAP enhancement or suppression >0.25 dB (Fig. 5C) were investigated further, including 15 trials with postshock enhancement and 21 with postshock suppression. The offset time-constants for enhancing versus suppressing trials were nonoverlapping (Fig. 5D): those for long-lasting CAP enhancement ranged from 3 to 20 min, whereas those for long-lasting CAP suppression ranged from 0.5 to 3 min.
Long-lasting effects of IC stimulation were further subdivided based on the behavior of the CM: in Fig. 5E, the magnitude and sign of the postshock CAP are plotted versus magnitude and sign of CM for all trials with CAP > 0.25 dB.
All trials with long-lasting CAP enhancement had CM < 0.2 dB and are considered as one category: CAP[+]CM[0]. These 15 trials were from six IC sites (Fig. 6A, 
) from five guinea pigs. One CAP-versus-time function from each of these six sites is illustrated in Fig. 6B: in the postshock period, the behavior of the CAP was reproducible across sites. In contrast, during the shock epoch, the behavior was heterogeneous: At two of the sites, there was a fast-onset CAP suppression that decayed dramatically as the shock trains continued. At other sites, there was minimal or no rapid change in CAP on shock onset, rather a slow, steady rise in CAP amplitude throughout the shock epoch.
|
Some trials with long-lasting CAP suppression also showed minimal change in postshock CM (<0.2 dB). However, in other cases, there were large CM enhancements. Significant CM suppression was never seen. Trials with CM < 0.2 dB (Fig. 5E) were grouped into a second category: CAP[–]CM[0]. The 11 trials in this group were from four IC sites (Fig. 6A, 
) from three guinea pigs. Example CAP-versus-time functions from each of these four sites are shown as black lines in Fig. 6C. For these trials, the postshock behavior of the CAP is homogenous. The CAP during the shock epoch is also similar across sites, but complex in nature. Immediately on shock onset, there is a rapid and strong CAP suppression, which then decays dramatically between the first and second points (as typical for IC stimulation, e.g., Fig. 4B), after which there can be a slower rise in suppression continuing to grow until shock offset. This behavior may arise from the interaction between two different suppressive processes, with different time constants and different origins.
The remaining 10 IC trials (Fig. 5E) exhibited long-lasting CAP suppression coupled with significant CM enhancement, referred to by the shorthand CAP[–]CM[+]. These trials derive from two sites in two animals. Their CAP-versus-time functions (gray lines in Fig. 6C) also show complex behavior during the shock epoch, with an initial rapidly rising and rapidly decaying suppression, followed by a second wave of suppression with a slower onset time constant.
For comparison, two trials, from two animals, representing OCB-evoked MOC[slow] effects (which also comprise CAP suppression and CM enhancement) are included in Fig. 5E. This MOC[slow] effect is qualitatively similar after the shocks to the IC-evoked CAP[–]CM[+] effect; however, during the shocks, the effect magnitude is three times larger than after the shock. In other words, on termination of the shocks, the MOC-[slow] effect abruptly falls to one-quarter its during-shocks magnitude (pairs of filled symbols in Fig. 6D).
CLUSTERING OF EFFECT CATEGORIES ALONG ADDITIONAL RESPONSE DIMENSIONS. In this section, trials from the three categories, defined on the basis of magnitude and sign of the long-lasting changes in CAP and CM, are analyzed and compared along other effect dimensions, including 1) the level and frequency dependence of CAP, 2) the relative size of ipsilateral versus contralateral CAP, 3) effects on other OHC-based response (DPOAEs vs. CM), and 4) effects on other neural-based response (RW noise vs. CAP). Along each of these additional dimensions, elements within each group tend to cluster together, and conversely, elements of different groups tend to be nonoverlapping. Furthermore, the behavior of groups with selective neural changes (i.e., CAP[+]CM[0] and CAP[–]CM[0]) tends to be unlike that seen when evoking MOC fast or slow effects, whereas the CAP[–]CM[+] group tends to be similar to the OCB-evoked MOC[slow] effects.
Level and frequency dependence of CAP. The level and frequency dependence of the CAP from all IC stimulation sites are summarized in Fig. 7, with IC sites showing long-lasting CAP suppression (Fig. 7A) plotted separately from long-lasting enhancement (Fig. 7B).
|
The level dependence is quantified by computing, for each trial, the ratio of the CAP measured with the low-level versus the high-level tone pips at 13 kHz. Recall from analyses of CAP-level functions (Fig. 2) that, for MOC[fast] effects, this level ratio should be »1: i.e., more suppression at low sound levels than at high sound levels. Two trials in Fig. 7 (stars) were evoked by OCB stimulation and thus represent MOC [slow] effects, which show the same large level dependence as MOC[fast] effects. The IC-evoked trials from both CAP[–]CM[+] sites show a large level dependence, and in this way, behave like MOC effects evoked by OCB stimulation. In contrast, both groups with selective neural effects (i.e., CAP[+]CM[0] and CAP[–]CM[0]) show relatively little level dependence. In this way, they act differently from MOC effects, and similar to the long-lasting suppressive and enhancing effects illustrated in Fig. 2, C and F.
The frequency dependence of the CAPs is quantified by computing the ratio of CAPs for the moderate-level 3-kHz tone pip and the high-level 13-kHz tone pip. In general, trials from the CAP[+]CM[0] group show somewhat greater effects at 3 kHz than at 13 kHz (ratios > 1), whereas the CAP[–]CM[0] group tends to show comparable CAP s at the two test frequencies (ratios approximately 1). With respect to the CAP[–]CM[+] group, the frequency dependence differs markedly for the two sites: the large low-frequency bias at site 1 makes these results different from all others evoked in the ventrolateral corner of the IC.
Contralateral versus ipsilateral CAP. As summarized in Fig. 8, trials with IC evoked long-lasting and selective changes in CAP amplitude, i.e., the CAP[+]CM[0] (Fig. 8B) and CAP[–]CM[0] (Fig. 8D) groups, showed little difference between the two ears in the magnitude of the CAP. Indeed, data from CAP[–]CM[0] sites lie very close to the 45° line indicating equilateral effects. Although the CAP[+]CM[0] sites show a slight contralateral bias, the contralateral CAPs are larger than ipsilateral by only a factor of approximately 1.2. In contrast, both CAP[–]CM[+] sites show a large bias in magnitude of the long-lasting CAP toward the contralateral ear. In this respect, the CAP[–]CM[+] effects (Fig. 8D) are similar to the IC-evoked (fast) effects seen during the shock epochs for almost all IC sites (Fig. 8, A and C). The strong contralateral bias for the (fast) during-shocks CAP suppression at all IC sites is consistent with effects mediated by an ipsilateral IC-to-MOC projection (Thompson and Thompson 1993) followed by a 2:1 contralateral bias in the MOC to cochlear projections (Robertson 1985).
|
Another similarity between MOC[fast] effects and the fastonset CAP suppression seen during IC shocks (Fig. 8, A and C) is that, for both phenomena, CAP is highly level dependent: in Fig. 8, A or C, the CAP for low-level tone pips at 13 kHz (
) is always greater than CAP for high-level tone pips at 13kHz (
). This stands in contrast to the lack of level dependence for the (slow) postshocks effects seen in both CAP[–]CM[0] and CAP[+]CM[0] groups, as described above (Fig. 7) and as also seen in the data in Fig. 8, B and D.
OHC-based measures: DPOAEs versus CM. Analysis of the magnitude of long-lasting changes in the DPOAE, in concert with the CM data, provide additional insight into long-lasting shock-evoked effects at the level of OHCs. Figure 9 summarizes the relationships between the CM and DPOAE (magnitudes and signs) for the same set of IC stimulation sites summarized in Figs. 5, 6, 7, 8. Consistent with a purely postsynaptic effect, the CAP[–]CM[0] and CAP[+]CM[0] groups exhibited negligible postshock changes in DPOAE magnitude: all data points are clustered around zero values for both CM and DPOAE (Fig. 9, A and B). In contrast, the CAP[–]CM[+] group exhibited nonzero postshock DPOAEs. As seen for CAP frequency dependence (Fig. 7A), the effects at the two sites in this group were significantly different. Site 2 showed long-lasting DPOAE suppression approximately four times greater than the CM enhancement (Fig. 9A). In contrast, site 1 showed enhancement of both DPOAE and CM of similar magnitude. Once again, the behavior at site 2 was more similar to MOC[slow] effects evoked at the OCB, which also showed long-lasting DPOAE suppression of significantly greater magnitude than the CM enhancement (approximately five to six times).
|
Neural-based measures: RW noise versus CAP. IC sites that produced long-lasting suppression of the CAP generally produced long-lasting suppression of RW noise. Conversely, IC sites which enhanced the CAP usually showed long-lasting enhancement of RW noise. The top panels of Fig. 10 compare RW noise spectra before (open symbols) and approximately 30 s after the shock epoch (filled symbols) for one sample trial from each of the three IC effects categories (Fig. 10, A–C), and for one trial with OCB stimulation evoking MOC[slow] effects (Fig. 10D). As illustrated by these spectra, RW noise shows a peak at roughly 900 Hz, corresponding to the electrical contribution of single auditory nerve spikes, as determined by cross-correlation (Kiang et al. 1976). Thus to estimate shock-induced changes in spontaneous rates in the auditory nerve, the average energy in the 800- to 1,000-Hz band has been extracted from spectra obtained at 13-s intervals (Fig. 1), and the change in amplitude (decibels from average preshock values) is plotted as a function of time, where it can be compared with the CAPs for the same trials (Fig. 10, E–H).
|
As illustrated by the examples chosen for each category, sites evoking selective effects on CAP (CAP[+]CM[0] and CAP[–]CM[0]) exhibited RW noise of the same sign and comparable magnitude to the simultaneously measured CAP. In contrast, RW noise at IC sites evoking CAP[–]CAP[+] effects were a fraction of the CAP (Fig. 10G). Relatively small change in RW noise re CAP was also observed for MOC[slow] effects evoked by OCB stimulation (Fig. 10H).
These trends are illustrated more completely by scatterplots (Fig. 11), where the magnitude of RW noise and CAP are compared for all IC sites. Effects categorized as CAP[–]CM[+] exhibit RW noise suppression less than one-third as large as the CAP suppression in the same trials. Recall that for this group, the multiple trials shown represent data from only two sites: within each site, the ratio of CAP to RW noise remains roughly constant across the multiple trials. For the CAP[–]CM[0] group, the ratio of RW noise to CAP was never smaller than 0.9 and could be as large as 4.25. For the CAP[+]CM[0] group, the ratio of RW noise to CAP was typically close to unity.
|
REPEATABILITY OF EFFECTS WITH MULTIPLE TRIALS AT THE SAME SITE. All categories of long-lasting effects were reproducible when multiple trials were carried out at a single site. As illustrated in Fig. 12, the inter-trial period varied from 5 to 15 min. Thus for the CAP[+]CM[0] effects, which showed offset time constants as long as 20 min (Fig. 5D), the CAP enhancement was sometimes cumulative, i.e., the effects of each shock epoch built on the enhancement of the previous one. In the example shown in Fig. 12A, the inter-trial period was 10 min. In this case, the parallel and comparable enhancement of CAP amplitude and RW noise contrasting with the lack of shockevoked effect on the DPOAE.
|
The offset time constants for both CAP[–] categories were shorter (Fig. 5D) than the minimum interval between trials (5 min). Thus it is unknown if they are cumulative; however, the data in Fig. 12, B–D, show that they are repeatable. Note, for both of the CAP[–]CM[+] sites (Fig. 12, C and D), there is particularly complex behavior of the DPOAE measure: a during-shock suppression followed by a postshock enhancement for site 2 and a during-shocks enhancement and a postshock suppression for site 1. Both these complex behaviors appear repeatedly across the multiple trials at the same site.
Effect of lesions in autonomic or olivocochlear pathways
The effects on cochlear responses elicited by IC shocks could be mediated by a number of pathways including 1) systemic effects on cardiovascular or respiratory system, 2) the autonomic neural projections to the ear, 3) the middle ear muscle pathways, and/or 4) the olivocochlear projections to the ear, either the MOC or the LOC system.
CARDIOVASCULAR OR RESPIRATORY EFFECTS OF IC SHOCKS. Animals were paralyzed and artificially ventilated in almost all experiments; thus changes in respiration rate were not elicited by the electrical shocks. Heart rate was continuously monitored in all experiments with the time-function paradigm. Although IC shocks often affected heart rate, sometimes increasing and sometimes decreasing it, the change in heart rate was not correlated with the change in cochlear responses (data not shown). Some of the largest IC-evoked changes in heart rate were seen with small CAPs, and some of the largest CAP s were seen without significant change in heart rate.
REMOVING AUTONOMIC PROJECTIONS TO THE COCHLEA. The SCG is the major source of autonomic innervation to the inner ear, and the projections are exclusively ipsilateral (Brown 1987; Densert and Flock 1974; Hozawa et al. 1989; Spoendlin 1981). To rule out a role for this feedback system, the SCG was removed unilaterally in three guinea pigs before the IC electrodes were placed. In two cases, IC sites were identified at which long-lasting CAP effects could be repeatedly evoked. In both cases, these repeatable effects were observed bilaterally. In one case (Fig. 13A), effects were of the CAP[+]CM[0] type; in the other case, CAP[–]CM[0] (Fig. 13B. In both cases, the effects were of similar magnitude and time course bilaterally, consistent with laterality previously established for these phenomena in animals with SCG intact bilaterally (Fig. 7). CAP[–]CM[+] effects were not evaluated after SCG ablation.
|
IDENTIFYING CONTRIBUTIONS OF MIDDLE EAR MUSCLES. Middle ear muscles are activated by shocks to the IC (Mulders and Robertson 2000) and also could be activated by shocks near the LSO, since stapedius and tensor tympani motoneurons are located in and near the superior olivary complex (Joseph et al. 1985). To minimize possible contributions of middle ear muscles, animals were paralyzed. A further, independent metric of middle-ear contractions was extractable from the time-function paradigm, since each DPOAE measurement also provided the ear-canal pressure generated by the primary tones. When animals were not curarized, the IC shocks often evoked large changes in the ear-canal sound pressure in response to the primaries, as is expected from the impedance change seen in response to stiffening the ossicular chain. After curarization, the changes in primary level were always insignificant during or after the shock epochs.
CUTTING THE CROSSED VERSUS THE ENTIRE OCB. The projections from MOC and LOC systems to ipsilateral and contralateral cochleas have been described in guinea pig (Robertson 1985). Since approximately two-thirds of the MOC-to-cochlea projections are contralateral, whereas more than 95% of the LOC projections are ipsilateral, cuts to the crossing fibers of the OCB at the midline will interrupt two-thirds of the MOC innervation to both ears while sparing almost all the LOC projections (Fig. 15). Cuts placed more laterally on the floor of the fourth ventricle can interrupt the entire OCB to one ear only (Kujawa and Liberman 1997). Success of crossed OCB lesion by midline incision was verified histologically in three of three cases examined (e.g., Fig. 14A). Verification of the lateral lesion was successful in the one case processed histologically.
|
|
Both midline and lateral OCB incisions were performed with IC-evoked effects on cochlear responses measured before and after such manipulations. Given the unpredictability of finding a particular effect type when placing the IC electrodes, it was not possible to study all permutations of lesion locus and effect category. Nonetheless, the ensemble OC-lesion data, summarized in Figs. 12 and 14, strongly suggest that the fast CAP suppression seen in both CAP-level and time-function paradigms is mediated by the MOC system, whereas the long-lasting enhancement or suppression of CAP, in the absence of changes in CM, is mediated by the LOC system.
Consider first the midline OCB incisions, designed to remove two-thirds of MOC projections without affecting the LOC system. In the case shown in Fig. 14B, the cut selectively eliminates the CAP suppression seen at low sound levels, without affecting the CAP enhancement seen at high sound levels. In another case (Fig. 14C), a midline OCB incision attenuates the fast, during-shocks CAP suppression (not visible due to response scaling) without affecting the magnitude of the long-lasting CAP enhancement in this CAP[+]CM[0] trial. A third example, with midline OCB incision and a CAP[–]CM[0] trial, is shown in Fig. 12B. Here, the selective elimination of MOC projections has little effect on the long-lasting suppression of the CAP and RW noise after the end of the shock epoch. However, during the shock epoch, the MOC lesion transforms a complex pattern of CAP suppression consisting of a rapid-onset, rap
