Congenital and prolonged adult-onset deafness cause distinct degradations in neural ITD coding with bilateral cochlear implants

Abstract

Bilateral cochlear implant (CI) users perform poorly on tasks involving interaural time differences (ITD), which are critical for sound localization and speech reception in noise by normal-hearing listeners. ITD perception with bilateral CI is influenced by age at onset of deafness and duration of deafness. We previously showed that ITD coding in the auditory midbrain is degraded in congenitally deaf white cats (DWC) compared to acutely deafened cats (ADC) with normal auditory development (Hancock et al., J. Neurosci, 30:14068). To determine the relative importance of early onset of deafness and prolonged duration of deafness for abnormal ITD coding in DWC, we recorded from single units in the inferior colliculus of cats deafened as adults 6 months prior to experimentation (long-term deafened cats, LTDC) and compared neural ITD coding between the three deafness models. The incidence of ITD-sensitive neurons was similar in both groups with normal auditory development (LTDC and ADC), but significantly diminished in DWC. In contrast, both groups that experienced prolonged deafness (LTDC and DWC) had broad distributions of best ITDs around the midline, unlike the more focused distributions biased toward contralateral-leading ITDs present in both ADC and normal-hearing animals. The lack of contralateral bias in LTDC and DWC results in reduced sensitivity to changes in ITD within the natural range. The finding that early onset of deafness more severely degrades neural ITD coding than prolonged duration of deafness argues for the importance of fitting deaf children with sound processors that provide reliable ITD cues at an early age.

FIG. 1

ITD sensitivity varies widely across IC neurons in long-term, adult-deafened cats. LEFT COLUMN, dot rasters show temporal discharge patterns as a function of ITD. Alternating colors distinguish blocks of responses to different ITDs. Stimulus pulse trains shown on top of each raster plot. Pulse rates were 20 pps (A, B) and 10 pps (C). RIGHT COLUMN, corresponding rate-vs.-ITD curves (mean ± SD). STVR ITD signal to total variance ratio. Examples show good (A), intermediate (B), and poor (C) ITD sensitivity.

FIG. 2

Neural ITD sensitivity is similar for long-term and acutely deafened cats, but degraded in congenitally deaf cats. A Distributions of ITD STVR for IC neurons in the three groups of cats. B Corresponding cumulative distributions. DASHED LINES: cumulative distributions for young (<11 months) and old (>11 months) subdivisions of congenitally deaf cats. STVR signal to total variance ratio.

FIG. 3

Shapes of rate–ITD curves vary across neurons. A Examples from long-term deafened cats showing each of the four shape categories. CIRCLES: data points. BLACK LINE: model fit. B Relative incidence of each shape is similar across the three deafness groups.

FIG. 4

ITD tuning is altered by long-term and congenital deafness. Distributions of ITD tuning metrics for the three groups of animals. A Best ITD. B Halfwidth. C ITD of maximum slope. HORIZONTAL BARS in A, C show the mean ± 1 SD of corresponding distributions.

FIG. 5

Summary of deafness effects on neural ITD coding. Average ITD tuning curve for each deafness group computed from the Gaussian best fits. Assuming ITD is encoded by the total activity in each of two hemispheric channels, behavioral ITD discrimination is predicted by the slopes of these curves. SOLID LINES: include only differences in ITD tuning (Fig. 4). DASHED LINES: scaled by the relative incidence of ITD-sensitive neurons (Table 2) to model effect of degraded ITD sensitivity.

FIG. 6

Effect of stimulus level on ITD sensitivity. TOP ROW: rate–ITD curves from three example neurons. Numbers to the right of each curve indicate stimulus level in decibels re. 1 mA. MIDDLE ROW: corresponding plot of ITD STVR vs. stimulus level. A DWC, 40 pps. B LTDC, 20 pps. C LTDC, 40 pps. D Percentage of neurons with significant ITD sensitivity as a function of level relative to the level producing the maximum ITD STVR. The number of neurons included in each level bin ranges from 12 to 95 (mean = 45).

FIG. 7

Spontaneous firing rates are elevated by prolonged deafness. A Distribution of spontaneous rates for the three deafness groups. INSET: closer view of distribution for rates >1 spike/s. Binwidth = 1 spikes/s. B Neural ITD sensitivity vs. spontaneous rate. STVR signal to total variance ratio (see text).

FIG. 8

Temporal discharge patterns are affected by deafness group. A Period histograms in response to 10-pps pulse train from three IC units in long-term deafened cats illustrate main discharge patterns. BLACK LINES: sum-of-Gaussians fit to data. B Incidence of each discharge pattern compared across deafness groups. C Comparison of ITD sensitivity as a function of discharge pattern and deafness group.

FIG. 9

Illustration of method for computing maximum pulse rate that evoked pulse-locked spiking. Example neuron from long-term deafened cat. A Cross-correlograms plot neural firing rates as a function of time given a stimulus pulse at t = 0. Gray-shaded area shows 95 % confidence bounds estimated from a Monte Carlo simulation. Peaks exceeding the confidence bounds indicate pulse-locked firing. B Normalized peak height as function of pulse rate. Positive values indicate significant pulse locking. Cutoff pulse rate found by interpolating curve to find pulse rate where normalized height drops below significance.

FIG. 10

Pulse-locking limits are lower in congenitally deaf cats than acutely or long-term deafened cats. A Distribution of pulse-locking limits compared across three deafness groups. Bin labeled “<20 pps” represents neurons that were not pulse-locked at any pulse rate. B Corresponding cumulative distributions. C Pulse-locking limit does not predict degree of neural ITD sensitivity.