Asymmetric sensorineural hearing loss induction
Acoustic overstimulation with a continuous 1 kHz tone at 136 dB for 3 hours directed at the right ear induces asymmetric SNHL. The better hearing ear is the left. shows ABR tone specific audiograms for study monkeys at the 6, 12, and 24 weeks recovery time periods. Ranges of thresholds for the right (grey square) and left (grey diamond) ears at each frequency derived from 13 normal young adult monkeys unrelated to this study are ghosted in the background for reference. The mean range of threshold variations from 0.5 to 8 kHz is ~20 dB (right – 22 (5) dB, left – 20 (4) dB; mean (SD)). All better (protected) hearing ears of study monkeys fall into the threshold range of normal monkeys. Overall hearing loss asymmetry in study monkeys is mild, with the right ear generally no poorer than 20 dB relative to the left. Specific patterns of hearing loss are idiosyncratic to individual monkeys.
Figure 2 ABR audiograms for study monkeys after acoustic overstimulation with a 1 kHz tone at 136 dB for 3 hours directed at the right ear. Overall mild asymmetric sensorineural hearing loss. Grey symbols (square, right ear; diamond, left ear) are thresholds for (more ...)
Empirical interaural difference distributions in normal monkeys
AI multiunit extracellular mapping experiments in 4 hemispheres of 3 normal monkeys (sm10, sm82, and sm64L and sm64R) provide control data to establish empirical distributions of interaural characteristic frequency (CF) and threshold differences for determining upper and lower α/2 = 0.025 cutoff values, which are used to identify penetration sites in study monkeys that are significantly different (p < 0.05) from normal monkeys.
shows CFs (range: 0.3–22.9 kHz) determined from contralateral and ipsilateral ear stimulation in normal monkeys to be highly correlated (r = 0.99, p < 0.01). shows the CF difference (contra less ipsi) distribution, with triangles marking the CF empirical significance cutoff values for the upper (+0.63 octaves) and lower (−0.27 octaves) tails (α/2 = 0.025) of the distribution. Asymmetric CF cutoff values may be due to the CF being consistently slightly higher when stimulated by the contralateral ear compared to the ipsilateral ear or random effects secondary to idiosyncratic neuronal sampling specific to this study. CF derived from contralateral stimulation is slightly higher (0.06 (0.23) octaves, difference mean (SD); p < 0.01). The significance of interaural CF asymmetry in normal monkeys is unlikely to be physiologically meaningful because the mean difference is small and well within the error interval of the data extraction procedure. are scatter plots of contralateral and ipsilateral cortical thresholds, respectively, with a behavioral audiogram (Green, 1975
) superimposed and connected by solid lines. For frequencies below 8 kHz, the very lowest cortical thresholds are in close proximity to behavioral thresholds. For frequencies above 8 kHz, there is under-sampling of AI. Neuronal sites with CF > 5 kHz are mostly within the suprameatal plane (Cheung et al., 2001
; Cheung, 2005
), which is inaccessible for visually guided mapping of layer III or IV neurons without removal of the parietal operculum. While the contours of minimum cortical and behavioral thresholds () are qualitatively similar to minimum ABR thresholds (), the latter are quantitatively 10–20 dB higher. Similar results have been reported in monkeys (Kamada et al., 1991
; Lasky et al., 1999
) and humans (Stapells et al., 1990
). shows broad variations in the relationship between contralateral and ipsilateral cortical thresholds (r = 0.67, p < 0.01). The distribution of threshold differences (contra minus ipsi) is shown in , with triangles marking the threshold empirical significance cutoff values for the upper (+18.4 dB) and lower (−24.8 dB) tails (α/2 = 0.025). Asymmetric threshold cutoff values is consistent with dominance of contralateral ear stimulation to drive cortical activity at the lowest thresholds (Benson and Teas, 1976
). Furthermore, suprathreshold stimulation data indicate the ability of the contralateral ear to evoke the strongest cortical responses (Imig and Adrian, 1977
; Semple and Kitzes, 1993
). Accordingly, threshold derived from contralateral stimulation is slightly lower (−3.4 (9.8) dB, difference mean (SD); p < 0.01). The cutoff tail values derived from empirical distributions of CF and threshold differences in normal monkeys serve as metrics to assess significance of differences observed in acoustically overstimulated study monkeys.
Figure 3 Characteristic frequency and threshold data for contralateral versus ipsilateral ear input from 4 cortices in 3 normal squirrel monkeys. A, Relationship between ipsilateral and contralateral CFs. R, right AI; L, left AI. Inset: squirrel monkey brain caricature. (more ...)
Asymmetric hearing loss results in frequency misalignment of interaural maps in AI. displays interaural CF map alignment, variations among monkeys, and shift direction profiles for contralateral versus ipsilateral input at 6, 12, and 24 weeks after acoustic overstimulation. Interaural CF misalignment at 6 and 12 weeks recovery () is mostly shifted to higher contralateral values (CF shift up) that are within ~1 octave of the corresponding ipsilateral CF at 1 kHz. Interaural CF misalignment at 24 weeks recovery () is much less pronounced and mostly shifted to lower contralateral values (CF shift down) that are just outside the lower cutoff mark. Tukey box plots of interaural CF misalignment distributions for all 3 recovery time periods show extended upper and/or lower tails compared to the control group (data from ). CF shift up and CF shift down neurons are accounted for at the tails of study cohort misalignment distributions. At 6 weeks (), monkey sm75 exhibits the largest deviation away from the control group median CF difference (~0 dB) and is the largest contributor to both extended tails of the overall study group (‘All’) distribution. Monkey sm17 contributes principally to the CF shift down tail while monkey sm69 contributes mostly to the CF shift up tail. At 12 weeks, all 3 monkeys contribute evenly to both extended tails. At 24 weeks, all 3 monkeys also contribute evenly to both extended tails, but the CF shift down tail predominates. The proportion of sites exhibiting CF shift, up or down, is similar (~ 30%) for all 3 recovery time periods (, far right bar labeled ‘All’). The main distinguishing feature among study groups is the drop in proportion of sites with CF shift up, from ~ 20% at 6 and 12 weeks to ~10% at 24 weeks. provide a finer grain accounting of CF shift direction by organizing data into single octave bands from 0.5 to 8 kHz. Neuronal units with ipsilateral CF from 0.5 to 4 kHz constitute the bulk of the data. Total sample size for a specific band is indicated by the number inside the parenthesis. In summary, the evolution of interaural frequency maps in the hemisphere contralateral to the poorer hearing ear progresses from gross divergence at 6 and 12 weeks to near convergence at 24 weeks after acoustic overstimulation. While the overall proportion of sites exhibiting CF change remains constant at 30%, the subset of sites with CF shift up declines from 20% to 10%.
Figure 4 Characteristic frequency for contralateral versus ipsilateral ear input in squirrel monkeys with mild asymmetric hearing loss. A, D & G, 6, 12, and 24 weeks recovery time periods. Interaural CF misalignment is evident. There is a reversal of shift (more ...)
Asymmetric hearing loss results in misalignment of interaural intensity preferences in AI. shows AI neuronal threshold distribution profiles for contralateral versus ipsilateral input at 6, 12, and 24 weeks after acoustic overstimulation. Over 40% of contralateral thresholds at 6 and 12 weeks are higher than expected (p < 0.05) relative to corresponding ipsilateral thresholds (), an anticipated consequence of acoustic overstimulation directed at the right ear. When the data are sorted along single octave bands ranging from 0.5 to 8 kHz, the proportion of sites with threshold shift up takes on a low pass filter configuration. The very lowest frequency bands have the highest proportion of sites with threshold shift up, in general accordance with study monkey ABR audiograms (). Remarkably, virtually all contralateral-ipsilateral threshold pairs fall within the 95% confidence interval () at 24 weeks, an unanticipated finding that is predominantly a consequence of unexpected elevations of ipsilateral neuronal thresholds. shows ipsilateral and contralateral cortical threshold Tukey box plots for normal and study monkey groups segmented by single octave bands referenced to the ipsilateral CF. Control data threshold box plots are colored grey to provide reference values for comparisons. Sample size is in parenthesis. At 6 and 12 weeks, contralateral cortical thresholds are elevated while ipsilateral thresholds are not, save the 4 – 8 kHz band at 12 weeks (p < 0.01, single tail t-test). At 24 weeks, both contralateral and ipsilateral cortical thresholds are elevated. When compared to 6 weeks, the 24 weeks mean contralateral threshold is 2.7 dB lower (47.7 (11.1), 6 weeks; 53.5 (12.4), 12 weeks; 45.0 (9.8), 24 weeks; mean (SD); p < 0.01, ANOVA) but the mean ipsilateral threshold is 15.3 dB higher (31.7 (10.8), 6 weeks; 39.2 (9.2), 12 weeks; 47.0 (11.9); mean (SD); p < 0.01, ANOVA). Hence, interaural AI neuronal threshold difference maps undergo temporal evolution. Elevation of ipsilateral cortical threshold appears to be the primary factor effecting reduction in threshold difference at the 24 weeks recovery period. At 6 and 12 weeks after unilaterally directed acoustic overstimulation, contralateral thresholds are higher than ipsilateral thresholds; the direction of interaural threshold difference is almost exclusively shifted to higher values. By 24 weeks, the threshold differences are eliminated because ipsilateral cortical thresholds have unexpectedly elevated at this later time period.
Figure 5 Threshold for contralateral versus ipsilateral ear input in squirrel monkeys with mild asymmetric hearing loss. A, C & E, 6, 12, and 24 weeks recovery time periods. Threshold shift direction is almost exclusively to higher values (threshold shift (more ...)
Figure 6 Cortical threshold Tukey box plots for squirrel monkeys with mild asymmetric hearing loss segmented by single octave ipsilateral CF bands. 6 & 12 Weeks, Contralateral thresholds are elevated while ipsilateral thresholds are indistinguishable from (more ...)
Disaggregate CF and threshold shifts data sets
, , and 3 provide detailed accounting of CF and threshold shifts contributions by individual monkeys for recovery periods 6, 12, and 24 weeks, respectively. The tables are organized by segmenting data into single octave bands referenced to the ipsilateral CF. At 6 weeks, monkey sm75 is the largest contributor to CF shift, up or down, with ~54% of penetration sites showing statistically significant change (p < 0.05, empirical distribution). Monkeys sm17 and sm69 also contribute to CF shift, but to lesser degrees (~20% and ~14%). The overall rate of CF shift at 6 weeks is ~30%. In all 3 monkeys, the rate of CF shift up is at or higher than CF shift down. The overall rate of threshold shift up is ~50%.
6 weeks recovery period disaggregated CF and threshold shifts. N – sample size.
12 weeks recovery period disaggregated CF and threshold shifts. N – sample size.
At 12 weeks, monkeys sm02, sm03, and sm87 have similar rates of aggregate CF shift, up or down, at ~28%. Monkey sm87 departs from sm02 and sm03, and all monkeys at 6 weeks because its rate of CF shift down is slightly higher than CF shift up (15.1% vs. 11.9%). In all other monkeys at 6 and 12 weeks, the rate of CF shift up is at or higher than CF shift down. Threshold shift up remains dominant in all 3 monkeys, but sm87 has a much lower rate (13.5%). The overall rate of threshold shift up is ~42%.
At 24 weeks, monkeys sm15, sm23, and sm27 contribute relatively evenly to the overall aggregate CF shift rate at ~30%. For all monkeys at 24 weeks, there are two important differences when comparing rates of CF and threshold shifts to results at 6 and 12 weeks. First, the rate of CF shift down is at least twice that of CF shift up. Second, the overall rate of threshold shift is below 1%. Those two distinctions demarcate the 24 weeks recovery period as clearly separate from earlier recovery periods.
CF and threshold shifts rates and CF shift magnitude
The rate of CF shift, up or down, is ~30% for all 3 recovery time periods (p > 0.05, chi square). In contrast, the rate of threshold shift is comparable (~45%) for 6 and 12 weeks, but declines to < 1% at 24 weeks (, p < 0.01, chi square).
Figure 7 Summary of CF and threshold shift rates, and CF shift magnitude. A, CF shift rate, up or down, is ~30% for all 3 recovery time periods. Threshold shift rates are indistinguishable (~45%) at 6 and 12 weeks, but they differ from the threshold shift rate (more ...)
The magnitude of CF shift (mean (SD) octaves), computed from absolute values of statistically significant up and down differences (p < 0.05, empirical distributions), is 1.27 (0.76) at 6 weeks, 0.96 (0.54) at 12 weeks, and 0.62 (0.37) at 24 weeks. There is progressive decline in CF divergence or realignment of interaural frequency maps with the passage of time (, p < 0.01, t-test with Bonferroni correction). provides a detailed accounting of CF shift contributions at recovery periods 6, 12, and 24 weeks. The means of CF shift down magnitudes for all 3 recovery periods are remarkably similar at −0.46 octaves. The means of CF shift up magnitudes decrease monotonically from 1.57 to 1.25 to 1.03 octaves over time.
The temporal evolution of interaural CF and threshold shifts in AI contralateral to the poorer hearing ear is most apparent at 24 weeks after overstimulation. Frequency maps come into realignment over time for the following reasons: 1) rate of CF shift up decreases, 2) CF shift up magnitudes decrease, 3) rate of CF shift down increases, and 4) CF shift down mean magnitudes are similar for all 3 recovery periods at −0.46 octaves, whose absolute value is lower than the CF shift up α/2 cutoff mark at 0.63 octaves. Interaural threshold differences are reduced dramatically because cortical thresholds from the ipsilateral, better hearing ear increase until they nearly match thresholds from the contralateral, poorer hearing ear.
Relationship between interaural ABR and CF differences
Absolute ABR interaural asymmetry and maximum CF difference levels are correlated at 6 and 12 weeks but not at 24 weeks. shows the relationship between maximum CF and corresponding ABR (see ) absolute difference levels for individual study monkeys at their respective recovery time periods. The peripheral interaural asymmetry level at a specific ABR frequency for a certain monkey is designated by 2 numbers with different fonts (regular, frequency in kHz; italic, monkey). Linear regression lines are fitted to all data within a particular recovery time period. Inclusion of all ABR frequencies () for regression analysis is not possible in most monkeys because cortical sampling is more sparse for CF > 4 kHz.
Figure 8 Relationship between absolute ABR interaural threshold asymmetry and maximum CF difference levels for all recovery time periods (see text for detail). A & B, Monkeys at 6 and 12 weeks show moderately strong linear correlation (r = 0.52, 6 weeks; (more ...)
At 6 and 12 weeks (), the correlation between absolute interaural ABR asymmetry and absolute maximum CF difference is moderately strong (r = 0.52, 6 weeks; r = 0.47, 12 weeks; both p < 0.05). At 24 weeks (), the linear fit is no longer statistically significant (p > 0.05).
At the earlier recovery time periods, absolute ABR interaural asymmetry levels at specific frequencies have significant power to predict absolute cortical maximum CF difference map divergence levels. Greater asymmetries in peripheral hearing loss tend to be accompanied by greater misalignment of cortical frequencies. At the later 24 weeks recovery time period, interaural cortical frequency maps have realigned and are no longer correlated with levels of peripheral threshold asymmetry.
Relationship between ABR audiogram and minimum cortical thresholds
ABR and cortical minimum thresholds are well aligned at earlier recovery time periods but not at the later interval. shows the relationship between minimum cortical and corresponding ABR thresholds for both ears in individual study monkeys at 6, 12, and 24 weeks. The specific ABR frequency and ear laterality are denoted by symbols (number, frequency in kHz; R, right; L, left) within each panel. Linear regression lines are fitted to data for both ears.
At 6 and 12 weeks, 5 of 6 monkeys (exception: sm03) show linear correlation relationship between minimum cortical and corresponding ABR thresholds that is strong (r > 0.74, all p < 0.05), consistent with an analogous population analysis in normal monkeys (, r = 0.82, p < 0.01). At 24 weeks, all 3 monkeys have non-significant (p > 0.05) linear fits when both ears are combined. Monkey sm03 at 12 weeks appears to be qualitatively more similar to monkeys at 24 weeks.
At earlier recovery periods, central auditory activation levels remain in lock step with changes in peripheral activation sensitivities. At the later 24 weeks recovery time period, however, a single linear regression model is inadequate to capture correlation relationships between minimum central and peripheral activation levels for both ears. Plastic change in cortical neuronal CF and threshold response properties responsible for interaural map realignments has uncoupled previously integrated binaural central and peripheral activation sensitivities. As a consequence, the predictable correlated relationship between minimum central and periphery thresholds observed in normal and earlier recovery period study monkeys is no longer valid at the later 24 weeks recovery period.