All procedures were conducted in accordance with guidelines of the National Institute of Health and were approved by the Animal Care and Use Committee of the Massachusetts Eye and Ear Infirmary. Experiments were conducted within a sound-shielded chamber (Ver et al., 1975
). A total of 30 guinea pigs (weight 286–524 g) were used: 29 guinea pigs for injections of KA and 1 guinea pig for a surgical cut of the dorsal and intermediate acoustic striae. The guinea pigs were anesthetized with urethane (0.9 g/kg, i.p.) followed 15 minutes later by fentanyl (0.15 mg/kg, i.m.). Booster injections of fentanyl (one third initial dose) were given approximately every 2 hours to maintain anesthesia. Rectal temperature was controlled to 37°C by heating the air in the chamber and by a heating pad. After tracheotomy, each pinna was removed and small ventilation holes were made in the bullae. The right cochlear nucleus was exposed by a posterior craniotomy and partial aspiration of the overlying cerebellum.
DPOAEs at frequency 2f1
were measured with an Etymotic Research (ER) 10C acoustic system placed in the right ear canal. Two primary tones (f2
= 10 or 12 kHz, with f2
= 1.2) were generated digitally (20-μsec sampling) using a D-A board (AO-6; National Instruments) in a computer operating under LabVIEW control. Total duration of the primary tones was 1 second, and for the second half of this interval a contralateral broad-band noise (approximately 70 dB sound pressure level [SPL], 500-msec duration) was presented by means of a contralaterally placed ER 10C system. There was a 3.5 second silent interval between stimulus presentations. Sound pressure was measured with the low-noise microphone in the ER10C probe. The microphone was calibrated in a coupler with a condenser microphone (Bruel and Kjaer, 0.25 inch). The microphone output was amplified and digitized by an A-D board (A-2000; National Instruments). The digitized microphone output was broken into contiguous samples of 10.24 msec. For each sample, a fast Fourier transform was computed and the amplitude at 2f1
was extracted and plotted vs. post-onset time (e.g., ). Samples were fit with smoothed curves (Stineman, 1980
) for display purposes.
Fig. 2 A,B: Illustration of the distortion product otoacoustic emissions (DPOAEs) -based metric used to assess medial olivocochlear (MOC) reflex strength and the changes seen after kainic acid (KA) injection in two cases. Left: The amplitude of the DPOAE vs. (more ...)
DPOAEs were measured over a matrix of 84, or more, combinations of primary-tone levels. Primary levels were varied in 1-dB steps. L1
(the level of the f1
primary tone) was varied over six or more steps within the range 74 to 84 dB SPL. At each L1
step, there were 14 or more L2
steps that began at 60 dB and increased to L1
). Responses to two to four complete sets (or matrices) of all primary level combinations (Kujawa and Liberman, 2001
; Boyev et al., 2002
) were obtained at the beginning of the experiment before injection of KA. From each response matrix, the largest DPOAE change in amplitude () was chosen and averaged with the largest change from the other response matrices to form the preinjection “MOC reflex strength”. In some experiments, opposite-ear response matrices were obtained by reversing the ears that received the primary tones and the noise bursts (DPOAE measured in the left ear and noise presented to the right [injected side] ear). Test/retest measurements of matrix responses were obtained over 4-hour periods in each of three animals in which the posterior skull was opened, but an injection pipette was not inserted. In all experiments, DPOAE measurements were made for up to 20 minutes after injection pipette placement in the cochlear nucleus, before KA injection, to ensure no change in response strengths as the result of pipette placement alone. Auditory brainstem responses (ABRs) were measured differentially from a vertex electrode screwed into the skull relative to a needle electrode just ventral to the external auditory meatus. Responses to 200 clicks of alternating polarity were amplified 10,000× and averaged. ABR peak amplitudes were measured from baseline to peak for wave P1, or previous trough to peak for peaks P2–P5.
Kainic acid (5 mM in phosphate-buffered saline pH 7.0) was delivered with a micropipette (tip diameter approximately 20 μm) attached to a 5-μl syringe mounted on a micromanipulator. The tip was aimed into the cochlear nucleus by visual guidance and the micromanipulator scale was used to determine the depth of penetration. Pressure injections of KA were made using a hydraulic master/slave remote injection system. KA injections always caused spiking on the ABR recording electrodes and sometimes caused seizures. To avoid seizures and restrict the extent of the lesions, the injections were made in small increments (usually 0.1 μl over a period of 10–115 minutes for a total injection of 0.3–0.6 μl). After the injection, ipsilateral DPOAE response matrices were taken semi-continuously, with opposite-ear response matrices obtained less frequently. Ipsilateral ABR measurements were made at approximately 1 hour intervals. Postinjection survival times are given from the middle of the injection period and were at least 8 hours to allow for development and stabilization of neuronal damage at the injection site (Bird et al., 1978
; Melcher et al., 1996a
At the conclusion of the experiment, animals were perfused intravascularly with 4% paraformaldehyde in phosphate-buffered saline. Serial frozen sections were cut (40 μm) in the transverse plane on a freezing microtome. The sections were mounted on gelatinized slides and counterstained with methylene blue for examination in the light microscope. The lesion extent was assessed relative to the cytoarchitectonically defined subdivisions (AVCN, PVCN, and DCN) as described in previous work in the guinea pig (Hackney and Osen, 1985
; Hackney et al., 1990
), with boundaries verified by comparison with the unlesioned CN. Camera lucida drawings (e.g., Figs. , , ) were made at 40× total magnification to document the lesion area in individual sections. Lesion sites were defined as the regions with almost-complete absence of neurons and the presence of cell fragments, areas that could be clearly delineated. Anatomical boundaries within lesioned areas were determined using the opposite (unlesioned) side as a reference. The lesion areas were surrounded by a highly variable “halo” in which there was abnormally low cell density and abnormal neuronal morphology (e.g., pyknotic nuclei) (Bird et al., 1978
; Melcher et al., 1996a
). Digital grayscale photomicrograph images were acquired by using Metamorph Image (Universal Imaging Corp.) software. The software controlled a Hamamatsu CCD digital camera mounted on a Nikon Eclipse compound microscope with a 4× plan apo objective without optical filters. For , images were imported into Adobe Photoshop and the standard auto contrast function was used to enhance the quality of each image. For comparison purposes, lesions were plotted on an atlas (Figs. , ) with nuclear boundaries on template sections drawn from a representative cochlear nucleus. Counts of the number of neurons in the PVCN were made at 100× total magnification. Separate counts were made for dorsal and ventral halves of the PVCN (halves defined by a line connecting the dorsoventral midpoints measured along the medial and lateral borders of the PVCN, as schematized in , insets). All neurons that contained a nucleolus in the core and cap areas of PVCN were counted, but neurons within the overlying granule-cell areas (Mugnaini et al., 1980
) were not counted.
Fig. 5 Data from four reflex-sparing lesions. Left: Bars show the long-term changes in Ipsi reflex strength on the injected side. Right: Camera-lucida drawings of lesion-containing sections indicate the regions of almost-complete cell loss from the kainic acid (more ...)
Fig. 6 A–D: Data from four reflex-interrupting lesions, all of which involved the PVCN. There was minor involvement of the DCN in one case (C) and more substantial involvement of the AVCN in another case (D). Format and abbreviations are as for (more ...)
Data from four PVCN lesions that spared the reflex. Format and abbreviations are as for .
Fig. 4 A,B: Photomicrographs of rostral posteroventral cochlear nucleus (PVCN) from the lesioned side (A) and control side (B) of Experiment 75. In A, the area of nearly complete cell loss is indicated by a dashed line. The lesion section is from slide 82: schematics (more ...)
Fig. 8 A–G: The six reflex-interrupting cases plotted on an atlas constructed from a template cochlear nucleus. Column G shows overlap of all the cases with areas of cell loss common to all six cases indicated in black. The first four cases are also (more ...)
Fig. 9 A–F: Six reflex-sparing cases plotted on the same atlas as . These six cases were chosen because they were the largest PVCN lesions that spared the reflex. There was no area of cell loss common to all of these cases. Black areas show the (more ...)
Fig. 10 Neuron counts for posteroventral cochlear nucleus (PVCN) lesions that either interrupted (dashed lines) or spared (thin solid lines) the medial olivocochlear reflex. Included are the six cases shown in and the six cases shown in . A,B: (more ...)
To explore the path taken by the axons of the reflex interneurons, in one guinea pig the dorsal acoustic stria and intermediate stria (two of the output pathways of the cochlear nucleus) were cut using a scalpel. Transection was confirmed by postexperiment histology. No KA was injected in this experiment.