Loudness summation for pulsatile electrical stimulation of the cochlea: effects of rate, electrode separation, level, and mode of stimulation.

The aim of these two experiments was to gain systematic data on the amount of loudness summation measured for dual-electrode stimuli with varying temporal and spatial separation of current pulses. Loudness summation is important in the implementation of speech processing strategies for implantees. However, the loudness mapping functions used in current speech processors utilize psychophysical data (thresholds and comfortable loudness levels) derived using single-electrode stimuli, and do not take into account the temporal and spatial patterns of the speech processor output. In the first experiment, the current reduction required to equalize the loudness of a dual-electrode stimulus to that of its component (and equally loud) single-electrode stimuli was measured for three electrode separations (0.75, 2.25, and 7.5 mm), three repetition rates (250, 500, and 1000 Hz), and two loudness levels (comfortably loud, and mid-dynamic range). It was found that electrode separation had little effect on loudness summation, except for interactions with level and rate effects at the smallest separation. More current adjustment (in dB) was required for higher rates and lower levels of stimulation. The second experiment investigated the effects of mode (monopolar versus bipolar) and pulse duration on loudness summation. More current adjustment was required in bipolar mode than in monopolar mode at the lower level only. The main effects in both experiments, and their interactions, are consistent with a loudness model in which the neural excitation density is first obtained by temporal integration of excitation at each cochlear place, then converted to specific loudness via a nonlinear relationship, and finally integrated over cochlear place to obtain the loudness. The two important features which affect the loudness relationships in dual-electrode stimulation in this model are the shape of the excitation density function and the amount by which the neural spike probability per pulse is reduced in areas of overlapping excitation due to refractory effects.