The effect of Gaussian noise on the threshold, dynamic range, and loudness of analogue cochlear implant stimuli

The deliberate addition of Gaussian noise to cochlear implant signals has previously been proposed to enhance the time coding of signals by the cochlear nerve. Potentially, the addition of an inaudible level of noise could also have secondary benefits: it could lower the threshold to the information-bearing signal, and by desynchronization of nerve discharges, it could increase the level at which the information-bearing signal becomes uncomfortable. Both these effects would lead to an increased dynamic range, which might be expected to enhance speech comprehension and make the choice of cochlear implant compression parameters less critical (as with a wider dynamic range, small changes in the parameters would have less effect on loudness). The hypothesized secondary effects were investigated with eight users of the Clarion cochlear implant; the stimulation was analogue and monopolar. For presentations in noise, noise at 95% of the threshold level was applied simultaneously and independently to all the electrodes. The noise was found in two-alternative forced-choice (2AFC) experiments to decrease the threshold to sinusoidal stimuli (100 Hz, 1 kHz, 5 kHz) by about 2.0 dB and increase the dynamic range by 0.7 dB. Furthermore, in 2AFC loudness balance experiments, noise was found to decrease the loudness of moderate to intense stimuli. This suggests that loudness is partially coded by the degree of phase-locking of cochlear nerve fibers. The overall gain in dynamic range was modest, and more complex noise strategies, for example, using inhibition between the noise sources, may be required to get a clinically useful benefit. © 2006 Association for Research in Otolaryngology.

The sciatic nerve of the toad Xenopus laevis as a physiological model of the human cochlear nerve

The response of single fibres of the human cochlear nerve to electrical stimulation by a cochlear implant has previously been inferred from the response of the cochlear nerve in other mammals. These experiments are hindered by stimulus artefact and the range of stimulus currents used is therefore much less than the perceptual dynamic range (from threshold to discomfort) of human subjects. We have investigated use of the sciatic nerve of the toad Xenopus laevis as a convenient physiological model of the human cochlear nerve. Use of this completely dissected nerve reduces the problems of stimulus artefact whilst maintaining the advantages of a physiological preparation. The validity of the model was assessed by measuring the refractory periods, excitation time-constant, and relative spread of single fibres using microelectrode recording. We have also investigated the response of nerve fibres to sinusoidal stimulation. Based on these measurements, we propose that the sciatic nerve may be a suitable model of the human cochlear nerve if the timescales of stimuli are decreased by a factor of about five to compensate for the slower dynamics of the sciatic nerve and if noise is added to the stimuli to compensate for the lower internal noise of sciatic nerve fibres.