Spatial tuning curves from apical, middle, and basal electrodes in cochlear implant users

Abstract

Forward-masked psychophysical spatial tuning curves (fmSTCs) were measured in 15 cochlear-implant subjects, 10 using monopolar stimulation and 5 using bipolar stimulation. In each subject, fmSTCs were measured at several probe levels on an apical, middle, and basal electrode using a fixed-level probe stimulus and variable-level maskers. Tuning curve slopes and bandwidths did not change significantly with probe level for electrodes located in the apical, middle, or basal region although a few subjects exhibited dramatic changes in tuning at the extremes of the probe level range. Average tuning curve slopes and bandwidths did not vary significantly across electrode regions. Spatial tuning curves were symmetrical and similar in width across the three electrode regions. However, several subjects demonstrated large changes in slope and/or bandwidth across the three electrode regions, indicating poorer tuning in localized regions of the array. Cochlear-implant users exhibited bandwidths that were approximately five times wider than normal-hearing acoustic listeners but were in the same range as acoustic listeners with moderate cochlear hearing loss. No significant correlations were found between spatial tuning parameters and speech recognition; although a weak relation was seen between middle electrode tuning and transmitted information for vowel second formant frequency.

Figure 1

fmSTCs obtained from Advanced Bionics C-I implant users. Each graph in a panel shows fmSTCs obtained at multiple probe levels on one specific electrode. Each row of graphs shows fmSTCs from a single subject: the left graph is for an apical electrode, the middle graph is for a middle electrode, and the right graph is for a basal electrode. The heading in each panel specifies the subject [e.g., C03 in Fig. 1A], the probe electrode pair (rEL04:M, with M for a monopolar reference), and the electrode configuration of the probe and the masker (MPp/MPm, for monopolar probe and monopolar masker) used to generate the fmSTC. The x axis shows electrode number with apical electrodes to the left and basal electrodes to the right. The y axis shows stimulus amplitude in μA. The horizontal position of the open circle in each panel corresponds to the active probe electrode; the vertical level of that symbol indicates the probe level (in μA) used to obtain the lowest level fmSTC. The level of the horizontal bar below the probe symbol indicates the absolute threshold of the probe (THSp). Spatial tuning curves were fitted with apical and basal slopes, shown by the solid lines. Maximum acceptable loudness levels and 3IFC absolute thresholds are shown by the gray symbols. Forward-masked thresholds comprising the fmSTC are represented by the black symbols, with the error bars indicating 1 standard deviation above and below the mean of at least three threshold measurements. The symbol legend entry indicates the level of the probe used to obtain the thresholds for a specific fmSTC (e.g., Lp = 150 μA). Forward-masked masker-level thresholds were determined for all activated electrodes in a given electrode array; however, for clarity, thresholds that reached the MALm on a particular masker electrode are not plotted. Average values of MALm and THSm are shown by the gray squares across the top of the graph and the gray diamonds across the bottom of the graph, respectively.

Figure 2

fmSTCs obtained from Advanced Bionics C-II implant users. Legend as in Fig. 1.

Figure 3

fmSTCs obtained from Nucleus N22 implant users. The heading in each panel specifies the subject code [e.g., N13 in Fig. 3A], the probe electrode pair (rEL07:08), and the electrode configuration of the probe and the masker (BPp/BP + 2m, for bipolar probe and bipolar + 2 masker) used to generate the fmSTC. For some subjects, a BP masker was not sufficiently loud to masker the BP probe, therefore the spatial extent between the active and reference electrodes was extended by one or two electrodes (e.g., BP + 2 for N13). The active electrode of the bipolar probe electrode pair is indicated by an open circle connected to an up arrow that indicates the reference electrode of the pair. The solid line connecting the two indicates the absolute threshold for the probe stimulus. The dashed line across the top of some of the graphs indicates the maximum current amplitude that could be generated by a subject’s device.

Figure 4

Level effects: slope ratios and width ratios of fmSTCs as a function of probe level expressed as %DR in microamps (μA). Ratios are defined in the footnote.

Figure 5

Electrode-region effects: Average slopes of fmSTCs and widths of fmSTCs expressed in mm. Error bars indicate one standard deviation each side of the mean.

Figure 6

Spatial tuning curves plotted in terms of frequency calculated from Greenwood’s (1960) place-frequency map.

Figure 7

Electrode-region effects: Slopes of fmSTCs and widths of fmSTCs expressed in terms of frequency calculated from Greenwood’s (1960) place-frequency map. GW, Greenwood frequencies. Error bars indicate one standard deviation each side of the mean.

Figure 8

Comparisons of Q-factor estimates in electric and acoustic hearing. Black squares and circles show Q-factors (characteristic frequency/bandwidth) obtained from C-I and C-II users with monopolar configurations, respectively. Gray diamond symbols show Q-factors obtained from N22 users with bipolar configurations. The gray upper line shows Q-factor estimates obtained from “equivalent-rectangular-bandwidth” measures (Qerb) by Oxenham and Shera (2003) for acoustic hearing using forward masking procedures. The black lower line is the best-fit power function to the Q-factors from all of the cochlear implant users. The asterisked line shows the predicted Q-factors for acoustic listeners with cochlear hearing loss, calculated from fmPTC data published by Nelson (1991).

Figure 9

Rate of transmitted information (RTI) for feature F2 in vowel identification at an SNR of +20 dB as a function of STC slope (dB/mm) on the middle electrode (left panel) and as a function of weighted average BW (mm) on the middle electrode (right panel).